)
-(**
- As you can see in the result file, the format is tab separated and contains the fields we specified in our our `customOutputFormat`.
-*)
(**
+As you can see in the result file, the format is tab separated and contains the fields we specified in our our `customOutputFormat`.
+
|# BLASTP 2.2.31+
|# Query: >gi|7525013|ref|NP_051039.1| photosystem II protein D1 (chloroplast) [Arabidopsis thaliana]
diff --git a/docsrc/content/GFF3.fsx b/docsrc/content/GFF3.fsx
index 89ee7cfa..26c5ccec 100644
--- a/docsrc/content/GFF3.fsx
+++ b/docsrc/content/GFF3.fsx
@@ -101,7 +101,7 @@ _Note: The order of key value pairs in field 9 (attributes) may be changed._
///Takes a seq>, a FASTA converter and a destination filepath and writes it into a .gff. Hint: Use converter = id if no FastA sequence is included.
(*** do-not-eval ***)
-let gffExampleWrite = GFF3.write features BioItem.symbol (__SOURCE_DIRECTORY__ + "/data/gffExampleWrite.gff")
+let gffExampleWrite = GFF3.write BioItem.symbol (__SOURCE_DIRECTORY__ + "/data/gffExampleWrite.gff") features
(**
##Example: Sequence of CDS
diff --git a/docsrc/content/release-notes.md b/docsrc/content/release-notes.md
index ead8fd79..bb6fa510 100644
--- a/docsrc/content/release-notes.md
+++ b/docsrc/content/release-notes.md
@@ -1,3 +1,81 @@
+#### 0.1.0 - Thursday, May 30, 2019
+Several bugfixes and additions to multiple sub projects:
+
+* **BioFSharp** (core):
+ * Additional functionality:
+ * [GravyScore](https://github.com/CSBiology/BioFSharp/commit/209c3497e3cdb1db56a0675e0f2a76634a6dbe7a) (Grand Average of Hydropathy) as additional amino acid property
+
+* **BioFSharp.IO**:
+ * Additional functionality:
+ * [GAF Parser](https://github.com/CSBiology/BioFSharp/commit/cbba6a41a9b239e26467f32048aaec3335373faf) for GO Annotation file format: [Gene Association File](http://geneontology.org/docs/go-annotation-file-gaf-format-2.0/)
+ * [Uniprot style Fasta header parser](https://github.com/CSBiology/BioFSharp/commit/f2a16aaa2456b0c431f6d50d0f78a12834671e97)
+ * [FastA to GFF3 converter functions](https://github.com/CSBiology/BioFSharp/commit/2cdd3537398040e1508736bd734b22a67a7c46e7)
+ * [GFF3 Pretty Printer](https://github.com/CSBiology/BioFSharp/commit/eaaa59fbd382721e75fbb9c6558b0ba2ff6afb00)
+
+ * BugFixes:
+ * [Fix OboParser](https://github.com/CSBiology/BioFSharp/commit/0354c9c13e7a4692f2ab61b80ef86ac8f5bd83c3) just taking the last occurence of the alt_id keyword and discards previous ones.
+ * Fix [Fasta](https://github.com/CSBiology/BioFSharp/commit/aff8eff849deb1cca411faf3c640d53f6e410497) and [GFF3](https://github.com/CSBiology/BioFSharp/commit/d0f059ab899c715a37b7f50318292c8a81f18dd9) writers appending to files instead of recreating
+
+* **BioFSharp.BioTools**:
+ * Additional functionality:
+ * [TMHMM biocontainer API wrapper](https://github.com/CSBiology/BioFSharp/commit/f11cb122df29ccaa0809d3c3c951294a1b645e0f) ([TMHMM](http://www.cbs.dtu.dk/services/TMHMM/) 2.0c predicts transmembrane helices in proteins)
+ * [FastP and Hera biocontainer API wrapper](https://github.com/CSBiology/BioFSharp/commit/28b7654d57824bcdfdb8bae16af2f0f706ed60ad)
+ * [IntaRNA biocontainer API wrapper](https://github.com/CSBiology/BioFSharp/commit/a659496179cd754fbea2fe9ef4030544a35eb68b) ([IntaRNA](https://github.com/BackofenLab/IntaRNA) is a tool for prediction of various nucleotide interactions)
+
+ * BugFixes:
+ * Fix Stream entry closed error in [BioContainer.tarOfStream()](https://github.com/CSBiology/BioFSharp/commit/20f8973ea717208627ef5a7ea0b72cbaecb4103c)
+
+* **BioFSharp.ImgP**:
+ * Additional functionality:
+ * [update 3DCWT correlation calculation to loop](https://github.com/CSBiology/BioFSharp/commit/0b4ffe93755d915da64f4231199b0ec54d4d6c4d)
+
+ * Bugfixes:
+ * [fix height adjustment in ricker](https://github.com/CSBiology/BioFSharp/commit/abab82be1ac9fa0c540acfb5f3ccc6bd5143df1f)
+ * [fix Ricker values for discrete time points](https://github.com/CSBiology/BioFSharp/commit/2bb6bb2b67ea43df2d9fe970bd1445e568df53d4)
+
+#### 0.0.16 - Thursday, March 7, 2019
+ * Fix template files for all nuget packages:
+ * use correct paths for binaries
+ * update project descriptions
+ * use type project to infer dependencies
+
+#### 0.0.15 - Thursday, March 7, 2019
+ * BioFSharp.ML - project introduction:
+ * Usage of Microsoft's CNTK library with a biological focus:
+ * This project comes with working CNTK integration:
+ * necessary dependencies to use CNTK and its unmanaged libraries
+ * CNTK loadscript: an easy way to load all dependencies of CNTK. load the script and use the resolveCNTKDependencies() function.
+ * Built-in deep neural network 'd::ppop' ([publication](https://www.frontiersin.org/articles/10.3389/fpls.2018.01559/full))
+ * predict peptide observability in mass spectrometry
+ * Classification: functions to determine peptide feature vectors as input for dppop
+ * Prediction: functions to predict peptide observability using the feature vectors prepared in Classification.
+ * NonPlant and Plant models as embedded resources: the two models used in the original Web API. Additionally load custom models.
+ * Othert additions:
+ * BioFSharp.BioTools:
+ * Integration of new tools as biocontainer APIs:
+ * Hera
+ * FastP
+ * ClustalO
+ * HMMER (experimental):
+ * hmmbuild
+ * hmmalign
+ * hmmsearch
+ * hmmscan
+ * hmmemit
+ * hmmpress
+ * hmmconvert
+
+#### 0.0.14 - Tuesday, February 12, 2019
+ * Addition of blast biocontainer support for makeblastdb, blastp, blastn with full parameter wrapping
+ * Extension of BioContainer functionality:
+ * Add windows/unix path conversions and subpath matching to MounInfo
+ * Add execReturnAsync (returns stdout of docker container) and execAsync (redirects stdout/stderr of container)
+
+#### 0.0.13 - Friday, February 8, 2019
+* Addition of the BioTools project, which uses docker.dotnet to use docker images and stream their output to fsi
+* Low level wrapping of multiple docker.dotnet functions for F#
+* Basic functionality for using docker images from fsi
+
#### 0.0.12 - Friday, December 28, 2018
* Addition of Pretty Printers, SOFT Parser, GEOFTP functions
* Improvement and refactoring of Modification functionality
diff --git a/paket.dependencies b/paket.dependencies
index 8874ee40..34c78cc4 100644
--- a/paket.dependencies
+++ b/paket.dependencies
@@ -1,4 +1,3 @@
-
framework: auto-detect
source https://nuget.org/api/v2
@@ -17,6 +16,11 @@ nuget Microsoft.Xaml
nuget SwaggerProvider
nuget YamlDotNet
+nuget docker.dotnet
+nuget SharpZipLib
+
+nuget CNTK.CPUOnly
+
git https://github.com/CSBiology/FSharpAux.git nuget Packages: /
nuget FSharpAux
nuget FSharpAux.IO
@@ -28,6 +32,7 @@ nuget FSharp.Stats.MSF
clitool dotnet-fake
+
group BioDB
framework:net45
source https://nuget.org/api/v2
diff --git a/paket.lock b/paket.lock
index 70afd277..80c8fb82 100644
--- a/paket.lock
+++ b/paket.lock
@@ -2,7 +2,7 @@ RESTRICTION: || (== net45) (== net47) (== netstandard2.0)
NUGET
remote: https://www.nuget.org/api/v2
Alea (3.0.4)
- Argu (5.1)
+ Argu (5.2)
FSharp.Core (>= 4.0.0.1) - restriction: || (== net45) (== net47) (&& (== netstandard2.0) (>= net45))
FSharp.Core (>= 4.3.2) - restriction: || (&& (== net47) (< net45) (>= netstandard2.0)) (== netstandard2.0)
System.Configuration.ConfigurationManager (>= 4.4) - restriction: || (&& (== net47) (< net45) (>= netstandard2.0)) (== netstandard2.0)
@@ -21,32 +21,30 @@ NUGET
System.ValueTuple (>= 4.4) - restriction: || (&& (== net45) (>= net46)) (&& (== net45) (>= netstandard2.0)) (== net47) (== netstandard2.0)
System.Xml.XmlSerializer (>= 4.3) - restriction: || (&& (== net45) (>= net46)) (&& (== net45) (>= netstandard2.0)) (== net47) (== netstandard2.0)
System.Xml.XPath.XmlDocument (>= 4.3) - restriction: || (&& (== net45) (>= net46)) (&& (== net45) (>= netstandard2.0)) (== net47) (== netstandard2.0)
- CommandLineParser (2.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
- System.Collections (>= 4.0.11-rc2-24027) - restriction: || (&& (== net45) (>= netstandard1.5)) (== net47) (== netstandard2.0)
- System.Console (>= 4.0.0-rc2-24027) - restriction: || (&& (== net45) (>= netstandard1.5)) (== net47) (== netstandard2.0)
- System.Diagnostics.Debug (>= 4.0.11-rc2-24027) - restriction: || (&& (== net45) (>= netstandard1.5)) (== net47) (== netstandard2.0)
- System.Globalization (>= 4.0.11-rc2-24027) - restriction: || (&& (== net45) (>= netstandard1.5)) (== net47) (== netstandard2.0)
- System.IO (>= 4.1.0-rc2-24027) - restriction: || (&& (== net45) (>= netstandard1.5)) (== net47) (== netstandard2.0)
- System.Linq (>= 4.1.0-rc2-24027) - restriction: || (&& (== net45) (>= netstandard1.5)) (== net47) (== netstandard2.0)
- System.Linq.Expressions (>= 4.0.11-rc2-24027) - restriction: || (&& (== net45) (>= netstandard1.5)) (== net47) (== netstandard2.0)
- System.Reflection (>= 4.1.0-rc2-24027) - restriction: || (&& (== net45) (>= netstandard1.5)) (== net47) (== netstandard2.0)
- System.Reflection.Extensions (>= 4.0.1-rc2-24027) - restriction: || (&& (== net45) (>= netstandard1.5)) (== net47) (== netstandard2.0)
- System.Reflection.TypeExtensions (>= 4.1.0-rc2-24027) - restriction: || (&& (== net45) (>= netstandard1.5)) (== net47) (== netstandard2.0)
- System.Resources.ResourceManager (>= 4.0.1-rc2-24027) - restriction: || (&& (== net45) (>= netstandard1.5)) (== net47) (== netstandard2.0)
- System.Runtime (>= 4.1.0-rc2-24027) - restriction: || (&& (== net45) (>= netstandard1.5)) (== net47) (== netstandard2.0)
- System.Runtime.Extensions (>= 4.1.0-rc2-24027) - restriction: || (&& (== net45) (>= netstandard1.5)) (== net47) (== netstandard2.0)
- dotnet-fake (5.10.1) - clitool: true
- Expecto (8.6)
- Argu (>= 5.1) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netstandard2.0)) (== net47) (== netstandard2.0)
- Mono.Cecil (>= 0.10) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netstandard2.0)) (== net47) (== netstandard2.0)
+ CNTK.CPUOnly (2.6)
+ CNTK.Deps.MKL (2.6)
+ CNTK.Deps.OpenCV.Zip (2.6)
+ CNTK.Deps.MKL (2.6)
+ CNTK.Deps.OpenCV.Zip (2.6)
+ CommandLineParser (2.4.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
+ Docker.DotNet (3.125.2)
+ Newtonsoft.Json (>= 9.0.1)
+ System.Buffers (>= 4.3) - restriction: || (&& (== net45) (>= net46)) (== net47) (&& (== netstandard2.0) (>= net46))
+ System.Buffers (>= 4.4) - restriction: || (&& (== net47) (< net45) (>= netstandard2.0)) (== netstandard2.0)
+ System.Runtime (>= 4.3) - restriction: || (&& (== net45) (>= net46)) (== net47) (&& (== netstandard2.0) (>= net46))
+ System.ValueTuple (>= 4.4)
+ dotnet-fake (5.12) - clitool: true
+ Expecto (8.8)
+ Argu (>= 5.2) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netstandard2.0)) (== net47) (== netstandard2.0)
+ Mono.Cecil (>= 0.10.1) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netstandard2.0)) (== net47) (== netstandard2.0)
System.Diagnostics.FileVersionInfo (>= 4.3) - restriction: || (&& (== net45) (>= netstandard2.0)) (&& (== net47) (< net461) (>= netstandard2.0)) (== netstandard2.0)
- Expecto.BenchmarkDotNet (8.6)
+ Expecto.BenchmarkDotNet (8.8)
BenchmarkDotNet (>= 0.10.14) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
FSharp.Core (>= 4.3.4) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
- Expecto.FsCheck (8.6)
- Expecto (>= 8.6) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netstandard2.0)) (== net47) (== netstandard2.0)
- FsCheck (>= 2.10.4) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netstandard2.0)) (== net47) (== netstandard2.0)
- Expecto.VisualStudio.TestAdapter (10.0.0) - version_in_path: true
+ Expecto.FsCheck (8.8)
+ Expecto (>= 8.8) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netstandard2.0)) (== net47) (== netstandard2.0)
+ FsCheck (>= 2.13) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netstandard2.0)) (== net47) (== netstandard2.0)
+ Expecto.VisualStudio.TestAdapter (10.0.1) - version_in_path: true
Expecto (>= 8.0 < 9.0) - restriction: || (&& (== net45) (>= net461)) (== net47) (&& (== netstandard2.0) (>= net461))
FSharp.Core (>= 4.0 < 5.0) - restriction: || (&& (== net45) (>= net461)) (== net47) (&& (== netstandard2.0) (>= net461))
Microsoft.TestPlatform.ObjectModel (>= 15.0 < 16.0) - restriction: || (&& (== net45) (>= net461)) (== net47) (&& (== netstandard2.0) (>= net461))
@@ -54,11 +52,11 @@ NUGET
Newtonsoft.Json (>= 10.0 < 11.0) - restriction: || (&& (== net45) (>= net461)) (== net47) (&& (== netstandard2.0) (>= net461))
FsCheck (2.13) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netstandard2.0)) (== net47) (== netstandard2.0)
FSharp.Core (>= 4.2.3) - restriction: || (&& (== net45) (>= net452)) (&& (== net45) (>= netstandard1.6)) (== net47) (== netstandard2.0)
- FSharp.Core (4.5.4)
+ FSharp.Core (4.6.1)
FSharp.Data.TypeProviders (5.0.0.6)
FSharp.Core (>= 3.1.2.5)
FSharp.Plotly (1.1.21)
- Microsoft.CodeAnalysis.Analyzers (2.6.2) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
+ Microsoft.CodeAnalysis.Analyzers (2.6.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
Microsoft.CodeAnalysis.Common (2.10) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
Microsoft.CodeAnalysis.Analyzers (>= 2.6.1) - restriction: || (&& (== net45) (>= netstandard1.3)) (== net47) (== netstandard2.0)
System.AppContext (>= 4.3) - restriction: || (&& (== net45) (>= netstandard1.3)) (== net47) (== netstandard2.0)
@@ -104,7 +102,7 @@ NUGET
Microsoft.CodeAnalysis.Common (2.10)
Microsoft.DotNet.PlatformAbstractions (2.1) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
System.Runtime.InteropServices.RuntimeInformation (>= 4.0)
- Microsoft.NETCore.Platforms (2.1.2) - restriction: || (&& (== net45) (>= netstandard2.0)) (&& (== net47) (< net46) (>= netstandard2.0)) (&& (== net47) (< netstandard1.3) (>= netstandard2.0)) (== netstandard2.0)
+ Microsoft.NETCore.Platforms (2.2) - restriction: || (&& (== net45) (>= netstandard2.0)) (&& (== net47) (< net46) (>= netstandard2.0)) (&& (== net47) (< netstandard1.3) (>= netstandard2.0)) (== netstandard2.0)
Microsoft.NETCore.Targets (2.1) - restriction: || (&& (== net45) (>= netstandard2.0) (< portable-net45+win8+wp8+wpa81)) (&& (== net47) (< net45) (>= netstandard2.0)) (&& (== net47) (>= netstandard2.0) (< portable-net45+win8+wp8+wpa81)) (== netstandard2.0)
Microsoft.TestPlatform.ObjectModel (15.9.0) - version_in_path: true, restriction: || (&& (== net45) (>= net461)) (== net47) (&& (== netstandard2.0) (>= net461))
System.Reflection.Metadata (>= 1.3) - restriction: || (&& (== net45) (>= net451)) (&& (== net45) (>= netstandard1.5)) (== net47) (== netstandard2.0)
@@ -113,13 +111,14 @@ NUGET
System.Security.AccessControl (>= 4.5) - restriction: || (&& (== net45) (>= monoandroid)) (&& (== net45) (>= monotouch)) (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (&& (== net45) (>= netstandard2.0)) (&& (== net45) (>= xamarinios)) (&& (== net45) (>= xamarinmac)) (&& (== net45) (>= xamarintvos)) (&& (== net45) (>= xamarinwatchos)) (== net47) (== netstandard2.0)
System.Security.Principal.Windows (>= 4.5) - restriction: || (&& (== net45) (>= monoandroid)) (&& (== net45) (>= monotouch)) (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (&& (== net45) (>= netstandard2.0)) (&& (== net45) (>= xamarinios)) (&& (== net45) (>= xamarinmac)) (&& (== net45) (>= xamarintvos)) (&& (== net45) (>= xamarinwatchos)) (== net47) (== netstandard2.0)
Microsoft.Xaml (4.0.0.1)
- Mono.Cecil (0.10.1) - version_in_path: true, restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netstandard2.0)) (== net47) (== netstandard2.0)
+ Mono.Cecil (0.10.3) - version_in_path: true, restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netstandard2.0)) (== net47) (== netstandard2.0)
Newtonsoft.Json (10.0.3) - version_in_path: true
+ SharpZipLib (1.1)
SwaggerProvider (0.8.2)
FSharp.Core (>= 4.0)
Newtonsoft.Json (>= 10.0)
System.AppContext (4.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
- System.Buffers (4.5) - restriction: || (&& (== net47) (>= monoandroid) (>= netstandard2.0)) (&& (== net47) (>= monotouch) (>= netstandard2.0)) (&& (== net47) (< net45) (>= netstandard2.0)) (&& (== net47) (< netstandard1.1) (>= netstandard2.0)) (&& (== net47) (>= netstandard2.0) (>= wpa81)) (&& (== net47) (>= netstandard2.0) (>= xamarinmac)) (&& (== net47) (>= netstandard2.0) (>= xamarintvos)) (&& (== net47) (>= netstandard2.0) (>= xamarinwatchos)) (&& (== net47) (>= xamarinios)) (== netstandard2.0)
+ System.Buffers (4.5) - restriction: || (&& (== net45) (>= net46)) (== net47) (== netstandard2.0)
System.Collections (4.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
System.Collections.Concurrent (4.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
System.Collections.Immutable (1.5) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
@@ -166,10 +165,10 @@ NUGET
System.Linq (4.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
System.Linq.Expressions (4.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
System.Management (4.5) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
- System.Memory (4.5.1) - restriction: || (&& (== net47) (< net45) (>= netstandard2.0)) (== netstandard2.0)
+ System.Memory (4.5.2) - restriction: || (&& (== net47) (< net45) (>= netstandard2.0)) (== netstandard2.0)
System.Buffers (>= 4.4)
System.Numerics.Vectors (>= 4.4) - restriction: || (&& (== net45) (>= net461)) (== net47) (== netstandard2.0)
- System.Runtime.CompilerServices.Unsafe (>= 4.5)
+ System.Runtime.CompilerServices.Unsafe (>= 4.5.2)
System.Numerics.Vectors (4.5) - restriction: || (&& (== net47) (< net45) (>= netstandard2.0)) (== netstandard2.0)
System.Reflection (4.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp1.1) (>= netstandard2.0)) (&& (== net45) (>= netcoreapp2.0)) (&& (== net45) (>= netstandard2.0) (< portable-net45+win8+wpa81)) (== net47) (== netstandard2.0)
Microsoft.NETCore.Platforms (>= 1.1) - restriction: || (&& (== net45) (< portable-net45+win8+wp8+wpa81)) (&& (== net47) (< net45)) (&& (== net47) (< portable-net45+win8+wp8+wpa81)) (== netstandard2.0)
@@ -178,18 +177,14 @@ NUGET
System.Reflection.Primitives (>= 4.3) - restriction: || (&& (== net45) (< portable-net45+win8+wp8+wpa81)) (&& (== net47) (< net45)) (&& (== net47) (< portable-net45+win8+wp8+wpa81)) (== netstandard2.0)
System.Runtime (>= 4.3) - restriction: || (&& (== net45) (< portable-net45+win8+wp8+wpa81)) (&& (== net47) (< net45)) (&& (== net47) (< portable-net45+win8+wp8+wpa81)) (== netstandard2.0)
System.Reflection.Emit.Lightweight (4.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
- System.Reflection.Extensions (4.3) - restriction: || (&& (== net45) (>= net461) (>= netstandard1.5)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
System.Reflection.Metadata (1.6) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (&& (== net45) (>= netstandard2.0)) (== net47) (== netstandard2.0)
System.Collections.Immutable (>= 1.5)
System.Reflection.Primitives (4.3) - restriction: || (&& (== net45) (>= netcoreapp1.1) (>= netstandard2.0)) (&& (== net45) (>= netstandard2.0) (< portable-net45+win8+wpa81)) (&& (== net47) (< net45) (>= netstandard2.0)) (&& (== net47) (>= netcoreapp1.1) (>= netstandard2.0)) (&& (== net47) (>= netstandard2.0) (< portable-net45+win8+wpa81)) (== netstandard2.0)
Microsoft.NETCore.Platforms (>= 1.1) - restriction: || (&& (== net45) (< portable-net45+win8+wp8+wpa81)) (&& (== net47) (< net45)) (&& (== net47) (< portable-net45+win8+wp8+wpa81)) (== netstandard2.0)
Microsoft.NETCore.Targets (>= 1.1) - restriction: || (&& (== net45) (< portable-net45+win8+wp8+wpa81)) (&& (== net47) (< net45)) (&& (== net47) (< portable-net45+win8+wp8+wpa81)) (== netstandard2.0)
System.Runtime (>= 4.3) - restriction: || (&& (== net45) (< portable-net45+win8+wp8+wpa81)) (&& (== net47) (< net45)) (&& (== net47) (< portable-net45+win8+wp8+wpa81)) (== netstandard2.0)
- System.Reflection.TypeExtensions (4.5.1) - restriction: || (&& (== net45) (>= net461) (>= netstandard1.5)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
System.Resources.ResourceManager (4.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
- System.Runtime (4.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (&& (== net45) (>= netstandard2.0)) (== net47) (== netstandard2.0)
- Microsoft.NETCore.Platforms (>= 1.1) - restriction: || (&& (== net45) (< portable-net45+win8+wp8+wpa81)) (&& (== net47) (< net45)) (&& (== net47) (< portable-net45+win8+wp8+wpa81)) (== netstandard2.0)
- Microsoft.NETCore.Targets (>= 1.1) - restriction: || (&& (== net45) (< portable-net45+win8+wp8+wpa81)) (&& (== net47) (< net45)) (&& (== net47) (< portable-net45+win8+wp8+wpa81)) (== netstandard2.0)
+ System.Runtime (4.3) - restriction: || (&& (== net45) (>= net46)) (&& (== net45) (>= netstandard2.0)) (== net47) (== netstandard2.0)
System.Runtime.CompilerServices.Unsafe (4.5.2) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (== netstandard2.0)
System.Runtime.Extensions (4.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (&& (== net45) (>= netstandard2.0)) (== net47) (== netstandard2.0)
Microsoft.NETCore.Platforms (>= 1.1) - restriction: || (&& (== net45) (< portable-net45+win8+wp8+wpa81)) (&& (== net47) (< net45)) (&& (== net47) (< portable-net45+win8+wp8+wpa81)) (== netstandard2.0)
@@ -229,19 +224,19 @@ NUGET
Microsoft.NETCore.Platforms (>= 1.1) - restriction: || (&& (== net45) (< portable-net45+win8+wp8+wpa81)) (&& (== net47) (< net45)) (&& (== net47) (< portable-net45+win8+wp8+wpa81)) (== netstandard2.0)
Microsoft.NETCore.Targets (>= 1.1) - restriction: || (&& (== net45) (< portable-net45+win8+wp8+wpa81)) (&& (== net47) (< net45)) (&& (== net47) (< portable-net45+win8+wp8+wpa81)) (== netstandard2.0)
System.Runtime (>= 4.3) - restriction: || (&& (== net45) (< portable-net45+win8+wp8+wpa81)) (&& (== net47) (< net45)) (&& (== net47) (< portable-net45+win8+wp8+wpa81)) (== netstandard2.0)
- System.Text.Encoding.CodePages (4.5) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
- System.Runtime.CompilerServices.Unsafe (>= 4.5) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (&& (== net45) (>= netstandard2.0)) (== net47) (== netstandard2.0)
+ System.Text.Encoding.CodePages (4.5.1) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
+ System.Runtime.CompilerServices.Unsafe (>= 4.5.2) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (&& (== net45) (>= netstandard2.0)) (== net47) (== netstandard2.0)
System.Text.Encoding.Extensions (4.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
System.Threading (4.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
System.Threading.Tasks (4.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (&& (== net45) (>= netstandard2.0) (< portable-net45+win8+wp8+wpa81)) (== net47) (== netstandard2.0)
Microsoft.NETCore.Platforms (>= 1.1) - restriction: || (&& (== net45) (< portable-net45+win8+wp8+wpa81)) (&& (== net47) (< net45)) (&& (== net47) (< portable-net45+win8+wp8+wpa81)) (== netstandard2.0)
Microsoft.NETCore.Targets (>= 1.1) - restriction: || (&& (== net45) (< portable-net45+win8+wp8+wpa81)) (&& (== net47) (< net45)) (&& (== net47) (< portable-net45+win8+wp8+wpa81)) (== netstandard2.0)
System.Runtime (>= 4.3) - restriction: || (&& (== net45) (< portable-net45+win8+wp8+wpa81)) (&& (== net47) (< net45)) (&& (== net47) (< portable-net45+win8+wp8+wpa81)) (== netstandard2.0)
- System.Threading.Tasks.Extensions (4.5.1) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
- System.Runtime.CompilerServices.Unsafe (>= 4.5)
+ System.Threading.Tasks.Extensions (4.5.2) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
+ System.Runtime.CompilerServices.Unsafe (>= 4.5.2)
System.Threading.Tasks.Parallel (4.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
System.Threading.Thread (4.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
- System.ValueTuple (4.5) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
+ System.ValueTuple (4.5)
System.Xml.ReaderWriter (4.3.1) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
System.Xml.XDocument (4.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
System.Xml.XmlDocument (4.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
@@ -252,19 +247,19 @@ NUGET
System.Xml.XPath.XmlDocument (4.3) - restriction: || (&& (== net45) (>= net461)) (&& (== net45) (>= netcoreapp2.0)) (== net47) (&& (== netstandard2.0) (>= net461)) (&& (== netstandard2.0) (>= netcoreapp2.0))
System.Xml.XmlDocument (>= 4.3) - restriction: || (&& (== net45) (>= net46)) (&& (== net45) (>= netstandard1.3)) (== net47) (== netstandard2.0)
System.Xml.XPath (>= 4.3) - restriction: || (&& (== net45) (>= net46)) (&& (== net45) (>= netstandard1.3)) (== net47) (== netstandard2.0)
- YamlDotNet (5.2.1)
+ YamlDotNet (5.3)
remote: paket-files/github.com/CSBiology/FSharp.Stats
- FSharp.Stats (0.0.1)
- FSharp.Stats.MSF (0.0.1)
+ FSharp.Stats (0.0.13)
+ FSharp.Stats.MSF (0.0.13)
remote: paket-files/github.com/CSBiology/FSharpAux
- FSharpAux (0.0.12)
- FSharpAux.IO (0.0.12)
+ FSharpAux (0.0.13)
+ FSharpAux.IO (0.0.13)
GIT
remote: https://github.com/CSBiology/FSharpAux.git
- (8b77b4f32aaa1f09c5cf436a61f6c446c27ba28e)
+ (9b35d33d818bdea8da54b75a060c309b3ed509bf)
path: /
remote: https://github.com/CSBiology/FSharp.Stats.git
- (4bf70fc0ddfb288cb7f166382b4c3d41b14bf415)
+ (e2ab6cf7746aadf50e01aa98a6666fdac5aefbc1)
path: /
GROUP BioDB
RESTRICTION: == net45
diff --git a/paket.references b/paket.references
index 999041f6..2492401e 100644
--- a/paket.references
+++ b/paket.references
@@ -1 +1,2 @@
-dotnet-fake
\ No newline at end of file
+dotnet-fake
+docker.dotnet
\ No newline at end of file
diff --git a/src/BioFSharp.BioDB/AssemblyInfo.fs b/src/BioFSharp.BioDB/AssemblyInfo.fs
index 4d0a8f68..0357f92e 100644
--- a/src/BioFSharp.BioDB/AssemblyInfo.fs
+++ b/src/BioFSharp.BioDB/AssemblyInfo.fs
@@ -5,8 +5,8 @@ open System.Reflection
[]
[]
[")>]
-[]
-[]
+[]
+[]
[]
do ()
@@ -14,6 +14,6 @@ module internal AssemblyVersionInformation =
let [] AssemblyTitle = "BioFSharp.BioDB"
let [] AssemblyProduct = "BioFSharp"
let [] AssemblyDescription = "An open source bioinformatics toolbox written in F#. "
- let [] AssemblyVersion = "0.0.12"
- let [] AssemblyFileVersion = "0.0.12"
+ let [] AssemblyVersion = "0.1.0"
+ let [] AssemblyFileVersion = "0.1.0"
let [] AssemblyConfiguration = "Release"
diff --git a/src/BioFSharp.BioDB/BioFSharp.BioDB.fsproj b/src/BioFSharp.BioDB/BioFSharp.BioDB.fsproj
index 096015ae..01fc5fc3 100644
--- a/src/BioFSharp.BioDB/BioFSharp.BioDB.fsproj
+++ b/src/BioFSharp.BioDB/BioFSharp.BioDB.fsproj
@@ -1,4 +1,4 @@
-
+
net45
BioFSharp.BioDB
@@ -20,6 +20,7 @@
+
diff --git a/src/BioFSharp.BioDB/paket.template b/src/BioFSharp.BioDB/paket.template
index ccfea2ea..7b61152d 100644
--- a/src/BioFSharp.BioDB/paket.template
+++ b/src/BioFSharp.BioDB/paket.template
@@ -1,9 +1,10 @@
type project
+id BioFSharp.BioDB
title
BioFSharp.BioDB
owners
Timo Mühlhaus
-authors
+authors
Timo Mühlhaus
projectUrl
https://github.com/CSBiology/BioFSharp
@@ -13,11 +14,17 @@ licenseUrl
https://github.com/CSBiology/BioFSharp/blob/master/LICENSE.txt
requireLicenseAcceptance
false
+language
+ F#
copyright
- Copyright 2018
+ Copyright 2019
tags
- F# FSharp bioinformatics
+ bioinformatics F# fsharp database-access
summary
- F# FSharp bioinformatics
+ APIs for querying common biological databases
description
- F# FSharp bioinformatics
+ APIs for querying common biological databases
+include-referenced-projects
+ true
+files
+ ../../bin/BioFSharp.BioDB ==> lib
diff --git a/src/BioFSharp.BioTools/AssemblyInfo.fs b/src/BioFSharp.BioTools/AssemblyInfo.fs
new file mode 100644
index 00000000..851e6769
--- /dev/null
+++ b/src/BioFSharp.BioTools/AssemblyInfo.fs
@@ -0,0 +1,19 @@
+// Auto-Generated by FAKE; do not edit
+namespace System
+open System.Reflection
+
+[]
+[]
+[")>]
+[]
+[]
+[]
+do ()
+
+module internal AssemblyVersionInformation =
+ let [] AssemblyTitle = "BioFSharp.BioTools"
+ let [] AssemblyProduct = "BioFSharp"
+ let [] AssemblyDescription = "An open source bioinformatics toolbox written in F#. "
+ let [] AssemblyVersion = "0.1.0"
+ let [] AssemblyFileVersion = "0.1.0"
+ let [] AssemblyConfiguration = "Release"
diff --git a/src/BioFSharp.BioTools/BioContainer.fs b/src/BioFSharp.BioTools/BioContainer.fs
new file mode 100644
index 00000000..78ad1c0f
--- /dev/null
+++ b/src/BioFSharp.BioTools/BioContainer.fs
@@ -0,0 +1,350 @@
+namespace BioFSharp.BioTools
+
+open System
+open System.Threading
+open Docker.DotNet
+open Docker.DotNet.Models
+
+/// BioContainer helper
+module BioContainer =
+
+ open Docker
+ open System.IO
+
+ //[]
+ type MountInfo =
+ | NoMount
+ | HostDir of string
+
+ override this.ToString() =
+ match this with
+ | NoMount -> "NoMount"
+ | HostDir _ -> sprintf "%s:%s" (MountInfo.getHostDir this) (MountInfo.getContainerPath this)
+
+ ///get the full mounted unix path used in the container
+ static member getContainerPath (hd:MountInfo) =
+ match hd with
+ | NoMount -> failwithf "No mount directory set."
+ | HostDir hostdirectory ->
+ if hostdirectory.Contains(" ") then
+ failwithf "paths mounted to docker cannot contain spaces.\r\nThe path %s contains spaces." hostdirectory
+ else
+ sprintf "/data/%s" ((Path.GetFullPath(hostdirectory).Replace(":","")) |> BioContainerIO.toUnixDirectorySeparator )
+
+ ///get the path of the windows host directory used to mount in the container
+ static member getHostDir (hd:MountInfo) =
+ match hd with
+ | NoMount -> failwithf "No mount directory set."
+ | HostDir hostdirectory -> Path.GetFullPath (hostdirectory)
+
+ ///get the container full mounted unix path of a file in a subfolder of the mounted host directory
+ static member containerPathOf (m:MountInfo) (filePath:string) =
+ let winDir = MountInfo.getHostDir m
+ let containerBase = MountInfo.getContainerPath m
+
+ //spaces not supported in unix paths
+ if filePath.Contains(" ") then
+ failwithf "paths mounted to docker cannot contain spaces.\r\nThe path %s contains spaces." filePath
+ else
+ //the given path is relative
+ if filePath.StartsWith(".") then
+ let fullFilePath =
+ //get absolute combined path
+ Path.Combine(containerBase,filePath)
+ |> Path.GetFullPath
+ |> BioContainerIO.toUnixDirectorySeparator
+
+ //check that combined path does not go above base (eg base/../../)
+ if (fullFilePath.StartsWith(containerBase)) then
+ fullFilePath |> BioContainerIO.toUnixDirectorySeparator
+ else
+ failwithf ("the relative path \r\n%s\r\n escapes the scope of the container base path \r\n%s\r\n. the combined path is:\r\n%s\r\n") filePath containerBase fullFilePath
+
+ else
+ //Path is not relative. Use Path functions to resolve ../ and check if absolute path is a subpath of the windows base path
+ // TO-DO: make subpath matchin non-case-sensitive because that works on the windows side
+ let fullFilePath = filePath |> Path.GetFullPath
+ if fullFilePath.StartsWith(winDir) then
+ //if absolute windows path is correct, replace it with the containerbase
+ fullFilePath.Replace(winDir,containerBase)
+ |> fun x -> x.Replace(":","")
+ |> BioContainerIO.toUnixDirectorySeparator
+ else
+ failwithf "The given path \r\n%s\r\n is not a subpath of the mounted host directory \r\n%s\r\n. If you want to use relative paths start them with ./" fullFilePath winDir
+
+
+
+
+
+
+
+
+
+ type BcContext = {
+ Id : Guid
+ Connection : DockerClient
+ ImageName : string
+ ContainerId : string
+ Mount : MountInfo
+ }
+
+
+ /// Connect to docker engine (docker deamon)
+ let connect str =
+ (new DockerClientConfiguration(new Uri(str)) ).CreateClient()
+
+
+ /// Connect to default local docker engine (docker deamon: "npipe://./pipe/docker_engine")
+ let connectLocalDefault () =
+ // TODO: Use System.Runtime.InteropServices.RuntimeInformation.IsOSPlatform(OSPlatform.Linux)
+ connect "npipe://./pipe/docker_engine"
+
+
+
+ /// Runs a container of a specified image and keeps it running
+ let initBcContextAsync (connection:DockerClient) (image: DockerId) =
+ if not (Docker.Image.exists connection image) then failwithf "Image %s does not exists! Please pull the image first." (string image )
+ async {
+ let! container =
+ let param = Docker.Container.ContainerParams.InitCreateContainerParameters(User="root",Image=string image,OpenStdin=true)
+ Docker.Container.createContainerWithAsync connection param
+
+ let! isRunning =
+ let param =
+ Docker.Container.ContainerParams.InitContainerStartParameters()
+
+ Docker.Container.startContainerWithAsync connection param container.ID
+
+ return {Id=Guid.NewGuid();Connection=connection;ImageName=string image;ContainerId=container.ID;Mount=MountInfo.NoMount}
+ }
+
+
+ /// Runs a container of a specified image and keeps it running on the local default docker engine
+ let initBcContextLocalDefaultAsync (image: DockerId) =
+ let client = connectLocalDefault ()
+ initBcContextAsync client image
+
+
+ /// Runs a container of a specified image and keeps it running. Bind mounts the host directory under /data/ (without ':' and lower letter according to BioContainer standards).
+ let initBcContextWithMountAsync (connection:DockerClient) (image: DockerId) (hostdirectory:string) =
+ if not (Docker.Image.exists connection image) then failwithf "Image %s does not exists! Please pull the image first." (string image )
+ let hd = MountInfo.HostDir hostdirectory
+ async {
+ let! container = // volume bind
+
+ let hostdirectory' = MountInfo.getHostDir hd
+ let target = MountInfo.getContainerPath hd //sprintf "/data/%s" (hostdirectory'.ToLower().Replace(":",""))
+ let mount = Docker.Container.ContainerParams.InitMount(Type="bind",Source=hostdirectory',Target=target,ReadOnly=false)
+ let hc = Docker.Container.ContainerParams.InitHostConfig(Mounts=[mount])
+ let param = Docker.Container.ContainerParams.InitCreateContainerParameters(User="root",HostConfig=hc,Image=string image,OpenStdin=true)
+ Docker.Container.createContainerWithAsync connection param
+
+ let! isRunning =
+ let param =
+ Docker.Container.ContainerParams.InitContainerStartParameters()
+
+ Docker.Container.startContainerWithAsync connection param container.ID
+
+ return {Id=Guid.NewGuid();Connection=connection;ImageName=string image;ContainerId=container.ID;Mount=hd}
+ }
+
+ /// Executes a command in the biocontainer context and returns the either the standard output of the container or the standard error of the container if stdout is empty
+ let execReturnAsync (bc:BcContext) cmd =
+ async {
+
+ let! execContainer =
+ let param =
+ Docker.Container.ContainerParams.InitContainerExecCreateParameters(
+ AttachStderr=true,
+ AttachStdout=true,
+ AttachStdin=false,
+ Cmd=cmd,
+ Detach=false
+ )
+
+ Docker.Container.execCreateContainerAsync bc.Connection param (bc.ContainerId)
+
+ let! stream =
+ let param =
+ Docker.Container.ContainerParams.InitContainerExecStartParameters(
+ AttachStderr=true,
+ AttachStdout=true,
+ AttachStdin=false,
+ Cmd=cmd
+ )
+ Docker.Container.startContainerWithExecConfigAsync bc.Connection param execContainer.ID
+
+
+ let stdOutputStream = new System.IO.MemoryStream()
+ let stdErrStream = new System.IO.MemoryStream()
+ let streamTask =
+ stream.CopyOutputToAsync(null,stdOutputStream,stdErrStream,CancellationToken.None)
+
+ do! streamTask |> Async.AwaitTask
+
+
+ let result =
+ if stdOutputStream.Length < 1L then
+ stdErrStream.Position <- 0L
+ BioContainerIO.readFrom stdErrStream
+ else
+ stdOutputStream.Position <- 0L
+ BioContainerIO.readFrom stdOutputStream
+
+ if stdErrStream.Length > 0L then
+ stdErrStream.Position <- 0L
+ System.Console.Error.Write(BioContainerIO.readFrom stdErrStream)
+
+ return result
+
+ }
+
+ ///Executes a command in the biocontainer context. Passes stdout and stderr of the container to stoud/stderr.
+ let execAsync (bc:BcContext) cmd =
+ async {
+
+ let! execContainer =
+ let param =
+ Docker.Container.ContainerParams.InitContainerExecCreateParameters(
+ AttachStderr=true,
+ AttachStdout=true,
+ AttachStdin=false,
+ Cmd=cmd,
+ Detach=false
+ )
+
+ Docker.Container.execCreateContainerAsync bc.Connection param (bc.ContainerId)
+
+ let! stream =
+ let param =
+ Docker.Container.ContainerParams.InitContainerExecStartParameters(
+ AttachStderr=true,
+ AttachStdout=true,
+ AttachStdin=false,
+ Cmd=cmd
+ )
+ Docker.Container.startContainerWithExecConfigAsync bc.Connection param execContainer.ID
+
+
+ let stdOutputStream = new System.IO.MemoryStream()
+ let stdErrStream = new System.IO.MemoryStream()
+ let streamTask =
+ stream.CopyOutputToAsync(null,stdOutputStream,stdErrStream,CancellationToken.None)
+
+ do! streamTask |> Async.AwaitTask
+
+
+
+ if stdErrStream.Length > 0L then
+ stdErrStream.Position <- 0L
+ System.Console.Error.Write(BioContainerIO.readFrom stdErrStream)
+
+ if stdOutputStream.Length > 0L then
+ stdOutputStream.Position <- 0L
+ System.Console.Write(BioContainerIO.readFrom stdOutputStream)
+
+ return ()
+
+ }
+
+
+ /// Disposes the biocontainer context (stops and removes the underlying container)
+ let disposeAsync (bc:BcContext) =
+ let param = Docker.Container.ContainerParams.InitContainerRemoveParameters(Force=true)
+ Docker.Container.removeContainerWithAsync bc.Connection param (Docker.DockerId.ContainerId bc.ContainerId)
+
+
+ /// Copies file from a container (only single file is supported)
+ let getFileAsync (bc:BcContext) (filePath) =
+ async {
+ let param = Docker.Container.ContainerParams.InitGetArchiveFromContainerParameters(Path=filePath)
+ let! res = Docker.Container.getArchiveFromContainerAsync bc.Connection param false bc.ContainerId
+ return BioContainerIO.tarToStream res.Stream
+ }
+
+ /// Puts a stream in a container (only single file is supported)
+ let putStreamAsync (bc:BcContext) (sourceStream:System.IO.Stream) targetFileName =
+ async {
+ let targetPath = BioContainerIO.directoryName targetFileName
+ let targetName = BioContainerIO.fileName targetFileName
+
+ // ! Set the target filename as tar-entry name to make renameing possible
+ let stream = BioContainerIO.tarOfStream targetName sourceStream
+
+ let param = Docker.Container.ContainerParams.InitContainerPathStatParameters(AllowOverwriteDirWithFile=true, Path=targetPath)
+ do!
+ Docker.Container.extractArchiveToContainerAsync bc.Connection param (bc.ContainerId ) stream
+
+ sourceStream.Close()
+ }
+
+ /// Copies file into a container (only single file is supported)
+ let putFileAsync (bc:BcContext) (sourceFileName:string) targetFileName =
+ async {
+ let fileStream = new System.IO.FileStream(sourceFileName,System.IO.FileMode.Open)
+ do!
+ putStreamAsync bc fileStream targetFileName
+ }
+
+
+ //let runCmdAsync (connection:DockerClient) (dockerid: DockerId) cmd =
+ // // Function creates and deletes new container all the time
+ // // !maybe use Containers.StartWithConfigContainerExecAsync (Docker.DotNet.Models.ContainerExecStartParameters()) in the future
+
+ // if not (Docker.Image.exists connection dockerid) then failwithf "Image %s does not exists! Please pull the image first." (dockerid.ToString())
+ // async {
+ // let! container =
+ // let param = Docker.Container.ContainerParams.InitCreateContainerParameters(Image=dockerid.ToString(),Cmd=cmd)
+ // Docker.Container.createContainerWithAsync connection param
+
+ // let! isRunning =
+ // let param = Docker.Container.ContainerParams.InitContainerStartParameters()
+ // Docker.Container.startContainerWithAsync connection param container.ID
+
+ // let! wait =
+ // Docker.Container.waitContainerAsync connection container.ID
+
+ // let! logs =
+ // let param = Docker.Container.ContainerParams.InitContainerLogsParameters(ShowStdout=true)
+ // Docker.Container.getContainerLogsAsync connection param container.ID
+
+ // do! Docker.Container.removeContainerAsync connection (DockerId.ContainerId container.ID)
+
+ // return logs
+ // }
+
+
+
+
+ ///// Run = create + start (or only start if available)
+ //let tryRunCmd (connection:DockerClient) (dockerid: DockerId) cmd =
+ // if Docker.Image.exists connection dockerid then
+ // let res = Docker.Container.createContainerByImage connection (dockerid.ToString())
+ // if Docker.Container.startContainer connection res.ID then
+ // Some
+ // else
+ // None
+
+ // else
+ // None
+
+ ///// Run = create + start (or only start if available)
+ //let tryRun (connection:DockerClient) (dockerid: DockerId) =
+ // if Docker.Image.exists connection dockerid then
+
+ // if Docker.Container.existsByImage connection dockerid then
+
+ // else
+ // if Docker.Container.startContainer connection (dockerid.ToString()) then
+ // Some DockerId
+ // else
+ // None
+ // else
+ // None
+
+
+
+
+
+
+
diff --git a/src/BioFSharp.BioTools/BioContainerIO.fs b/src/BioFSharp.BioTools/BioContainerIO.fs
new file mode 100644
index 00000000..0be7c35f
--- /dev/null
+++ b/src/BioFSharp.BioTools/BioContainerIO.fs
@@ -0,0 +1,78 @@
+namespace BioFSharp.BioTools
+
+open System
+open System.Threading
+open Docker.DotNet
+open Docker.DotNet.Models
+open System.IO
+
+/// BioContainer helper
+module BioContainerIO =
+
+ open ICSharpCode.SharpZipLib.GZip
+ open ICSharpCode.SharpZipLib.Tar
+ open System.IO
+
+ let toUnixDirectorySeparator (filename:string) =
+ let dirSep = "/"
+ filename.Replace("\\", dirSep)
+ .TrimEnd(Path.DirectorySeparatorChar)
+
+ let directoryName (filename:string) =
+ let dirSep = Path.DirectorySeparatorChar
+ let tmp = filename.Split([|dirSep;'/'|])
+ tmp
+ |> Seq.take (tmp.Length-1)
+ |> String.concat ("/")
+ |> fun s -> s + "/"
+
+ let fileName (filename:string) =
+ let dirSep = Path.DirectorySeparatorChar
+ let tmp = filename.Split([|dirSep;'/'|])
+ let last = tmp.Length-1
+ if last > 0 then tmp.[last] else ""
+
+ let readFrom (stream:System.IO.Stream) =
+ let length = (stream.Length) |> int
+ let tmp : array = Array.zeroCreate length
+ stream.Read(tmp,0,length) |> ignore
+
+ System.Text.Encoding.UTF8.GetString(tmp,0,length)
+
+ // Returns the first file as stream
+ let tarToStream inStream =
+ let tarIn = new TarInputStream(inStream)
+ if tarIn.GetNextEntry().IsDirectory then
+ tarIn.GetNextEntry() |> ignore
+ tarIn :> Stream
+ else
+ tarIn :> Stream
+
+
+ /// Returns a tar-archive MemoryStream (only one entry supported)
+ let tarOfStream (tarEntryName:string) (inputStream:Stream) =
+ let outStream = new MemoryStream()
+ let tarOutputStream = new TarOutputStream(outStream)
+
+ let fileSize = inputStream.Length
+ let entry = TarEntry.CreateTarEntry(tarEntryName)
+ // Must set size, otherwise TarOutputStream will fail when output exceeds.
+ entry.Size <- fileSize
+ // Add the entry to the tar stream, before writing the data.
+ tarOutputStream.PutNextEntry(entry)
+ // this is copied from TarArchive.WriteEntryCore
+ let localBuffer : byte [] = Array.zeroCreate (32 * 1024)
+ let rec loop () =
+ let numRead = inputStream.Read(localBuffer, 0, localBuffer.Length)
+ if (numRead <= 0) then
+ tarOutputStream.CloseEntry()
+ else
+ tarOutputStream.Write(localBuffer, 0, numRead)
+ loop ()
+
+ loop ()
+ tarOutputStream.IsStreamOwner <- false
+ tarOutputStream.Close()
+ outStream.Position <- 0L
+ outStream
+
diff --git a/src/BioFSharp.BioTools/BioFSharp.BioTools.fsproj b/src/BioFSharp.BioTools/BioFSharp.BioTools.fsproj
new file mode 100644
index 00000000..6fa10fb9
--- /dev/null
+++ b/src/BioFSharp.BioTools/BioFSharp.BioTools.fsproj
@@ -0,0 +1,40 @@
+
+
+
+ net45;net47;netstandard2.0
+ BioFSharp.BioTools
+ BioFSharp.BioTools
+ BioFSharp.BioTools
+ Library
+ true
+
+ true
+
+ true
+
+ $(AllowedOutputExtensionsInPackageBuildOutputFolder);.pdb
+ Debug;Release;Mono
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
diff --git a/src/BioFSharp.BioTools/BioFSharp.BioTools.fsx b/src/BioFSharp.BioTools/BioFSharp.BioTools.fsx
new file mode 100644
index 00000000..ffea26d0
--- /dev/null
+++ b/src/BioFSharp.BioTools/BioFSharp.BioTools.fsx
@@ -0,0 +1,495 @@
+#r "netstandard"
+#r "../../packages/Newtonsoft.Json.10.0.3/lib/netstandard1.3/Newtonsoft.Json.dll"
+#r "../../packages/System.Buffers/lib/netstandard2.0/System.Buffers.dll"
+#r "../../packages/Docker.DotNet/lib/netstandard2.0/Docker.DotNet.dll"
+
+#r "../../packages/SharpZipLib/lib/netstandard2.0/ICSharpCode.SharpZipLib.dll"
+
+#r "../../packages/SharpZipLib/lib/netstandard2.0/ICSharpCode.SharpZipLib.dll"
+#r "../../packages/FSharpAux.IO/lib/netstandard2.0/FSharpAux.dll"
+#r "../../packages/FSharpAux.IO/lib/netstandard2.0/FSharpAux.IO.dll"
+
+#load "Docker.fs"
+#load "BioContainerIO.fs"
+#load "BioContainer.fs"
+#load "TargetP.fs"
+#load "Blast.fs"
+#load "ClustalO.fs"
+#load "HMMER.fs"
+
+open System.Threading
+open System.Threading
+open Docker.DotNet
+open System.Threading
+open System.Buffers
+open System.Threading.Tasks
+
+open BioFSharp.BioTools
+open System.Collections.Generic
+open Docker.DotNet.Models
+open System.IO
+
+
+open ICSharpCode.SharpZipLib.GZip
+open ICSharpCode.SharpZipLib.Tar
+open Newtonsoft.Json.Serialization
+open System
+
+
+
+let client = Docker.connect "npipe://./pipe/docker_engine"
+
+
+
+let ubuntu = Docker.ImageName "ubuntu"
+
+let bcContextUbuntu =
+ BioContainer.initBcContextWithMountAsync client ubuntu "C:/tmp"
+ |> Async.RunSynchronously
+
+
+
+BioContainer.disposeAsync bcContextUbuntu
+|> Async.Start
+
+
+let bcContext =
+ BioContainer.initBcContextLocalDefaultAsync TargetP.ImageTagetP
+ |> Async.RunSynchronously
+
+
+let stream = new FileStream("C:/tmp/twelve.fsa",FileMode.Open)
+
+let res = TargetP.run bcContext (TargetP.NonPlant) stream
+
+
+
+BioContainer.disposeAsync bcContext
+|> Async.Start
+
+
+
+res |> Seq.head |> fun tp -> tp.Mtp
+
+
+
+
+
+//// https://github.com/Microsoft/Docker.DotNet/issues/223 -> write
+//// https://github.com/Microsoft/Docker.DotNet/issues/212 -> read
+
+////-i, --interactive=false Keep STDIN open even if not attached
+////-t, --tty=false Allocate a pseudo-TTY
+
+//let exe =
+// async {
+// //let! container =
+// // let param = Docker.Container.ContainerParams.InitCreateContainerParameters(Image=dockerid.ToString(),Cmd=cmd,OpenStdin=true)
+// // Docker.Container.createContainerWithAsync connection param
+
+// //let! isRunning =
+// // let param =
+// // Docker.Container.ContainerParams.InitContainerStartParameters()
+
+// // Docker.Container.startContainerWithAsync connection param container.ID
+
+// let! execContainer =
+// let param =
+// Docker.Container.ContainerParams.InitContainerExecCreateParameters(
+// AttachStderr=true,
+// AttachStdout=true,
+// AttachStdin=false,
+// Cmd=cmd',
+// Detach=false
+// //Tty=false
+// )
+
+// Docker.Container.execCreateContainerAsync connection param (cont)
+// return tmp
+// }
+
+// |> Async.RunSynchronously
+
+
+//////docker stop $(docker ps -a -q)
+
+////Docker.Container.removeContainerAsync connection (Docker.DockerId.ContainerId (container.ID))
+////|> Async.RunSynchronously
+
+
+//let ms =
+// async {
+
+// let! execContainer =
+// let param =
+// Docker.Container.ContainerParams.InitContainerExecCreateParameters(
+// AttachStderr=true,
+// AttachStdout=true,
+// AttachStdin=false,
+// Cmd=cmd,
+// Detach=false
+// )
+
+// Docker.Container.execCreateContainerAsync connection param (cont)
+
+// let! stream =
+// let param =
+// Docker.Container.ContainerParams.InitContainerExecStartParameters(
+// AttachStderr=true,
+// AttachStdout=true,
+// AttachStdin=false,
+// Cmd=cmd
+// )
+// Docker.Container.startContainerWithExecConfigAsync connection param cont // startContainerExecAsync connection exe.ID //
+
+// printfn "Start Exec"
+
+// //let stopParam = new ContainerStopParameters()
+// //let! st = connection.Containers.StopContainerAsync(cont,stopParam) |> Async.AwaitTask
+
+// //printfn "Stop: %b" st
+
+// let stdOutputStream = new System.IO.MemoryStream()
+// let streamTask =
+// stream.CopyOutputToAsync(null,stdOutputStream,null,CancellationToken.None)
+
+
+// do! streamTask |> Async.AwaitTask
+
+// printfn "Streamed"
+
+// //let! wait =
+// // Docker.Container.waitContainerAsync connection container.ID
+
+// let result =
+// stdOutputStream.Position <- 0L
+// readFrom stdOutputStream
+
+// //do! Docker.Container.removeContainerAsync connection (Docker.DockerId.ContainerId container.ID)
+
+// return result
+
+// }
+// |> Async.RunSynchronously
+
+
+
+//let ms =
+// async {
+// let! container =
+// let param =
+// Docker.Container.ContainerParams.InitCreateContainerParameters(
+// ArgsEscaped=false,
+// AttachStderr=true,
+// AttachStdout=true,
+// AttachStdin=false,
+// Image=string dockerid,
+// Cmd=cmd
+// )
+
+// Docker.Container.createContainerWithAsync connection param
+
+// //let! isRunning =
+// // let param = Docker.Container.ContainerParams.InitContainerStartParameters()
+// // Docker.Container.startContainerWithAsync connection param container.ID
+
+// let! stream =
+// let param = Docker.Container.ContainerParams.InitContainerAttachParameters (Stdout=true,Stderr=true,Stdin=false,Stream=true)
+// connection.Containers.AttachContainerAsync(container.ID,false,param)
+// |> Async.AwaitTask
+
+// let stdOutputStream = new System.IO.MemoryStream()
+// let streamTask =
+// stream.CopyOutputToAsync(null,stdOutputStream,null,CancellationToken.None)
+
+// let! isRunning =
+// let param =
+// Docker.Container.ContainerParams.InitContainerExecStartParameters(
+// AttachStderr=true,
+// AttachStdout=true,
+// AttachStdin=false,
+// Cmd=cmd
+// )
+// Docker.Container.startContainerWithExecConfigAsync connection param container.ID
+
+// do! streamTask |> Async.AwaitTask
+
+// let! wait =
+// Docker.Container.waitContainerAsync connection container.ID
+
+// let result =
+// stdOutputStream.Position <- 0L
+// readFrom stdOutputStream
+
+// do! Docker.Container.removeContainerAsync connection (Docker.DockerId.ContainerId container.ID)
+
+// return result
+
+// }
+// |> Async.RunSynchronously
+
+
+
+////let tmp =
+//// BioContainer.runCmdAsync client (Docker.DockerId.ImageName "ubuntu") ["echo"; "hello world"]
+//// |> Async.RunSynchronously
+//// |> readFrom
+
+
+
+//Docker.Image.exists client (Docker.DockerId.ImageName "targetp_image")
+
+
+//Docker.Image.listImages client
+//|> Seq.map (fun i -> i.ID )
+//|> Seq.toArray
+
+
+//Docker.Container.existsByImage client (Docker.DockerId.ImageName "targetp_image")
+
+
+////ancestor=([:], or )
+
+//let filters =
+// Docker.Container.ContainerParams.InitContainerListParameters(All=true,Filters=Docker.Filters.InitContainerFilters(Ancestor=Docker.DockerId.ImageName "ubuntu"))
+
+
+//Docker.Container.listContainersWithAsync client filters
+//|> Async.RunSynchronously
+//|> Seq.map (fun x -> x.Command,x.Image,x.Labels)
+//|> Seq.toArray
+
+////client.Containers.StartWithConfigContainerExecAsync
+//let p = Docker.DotNet.Models.ContainerExecStartParameters()
+
+
+//let ap = Docker.DotNet.Models.ContainerAttachParameters()
+
+
+
+
+//Docker.Container.existsByImage client (Docker.DockerId.ImageName "targetp_image")
+
+
+//let idtp = "61fbfbc30382e83dd585c99583c036ef8c5ced4eb10e1b274f199da6b6969588"
+
+////let pipe = System.Uri("npipe://./pipe/docker_engine")
+
+////let config = new DockerClientConfiguration(pipe)
+////let client = config.CreateClient()
+
+////let createByImage (client:DockerClient) imageName =
+//// async {
+//// let param = Models.CreateContainerParameters()
+//// param.Image <- imageName
+//// param.Cmd <- System.Collections.Generic.List(["echo"; "hello world"])
+//// let! container =
+//// client.Containers.CreateContainerAsync (param,CancellationToken.None)
+//// |> Async.AwaitTask
+//// return container.ID
+//// }
+
+
+////let result =
+//// async {
+//// let paramLog = Models.ContainerLogsParameters() // (Stdout = System.Nullable(true),Stdin = System.Nullable(true))
+//// paramLog.ShowStdout <- System.Nullable(true)
+//// let paramRun = Models.ContainerStartParameters ()
+
+//// //let id =
+//// // "4243adc7f3832ea35bdaad79aabe86f8e1c54f5c3a799cc72e060a8402bc24cb"
+
+//// let! id = createByImage client "ubuntu"
+
+//// let! isRunnig =
+//// client.Containers.StartContainerAsync(id,paramRun,CancellationToken.None)
+//// |> Async.AwaitTask
+
+//// let! wait =
+//// client.Containers.WaitContainerAsync(id,CancellationToken.None)
+//// |> Async.AwaitTask
+
+//// let! logs =
+//// client.Containers.GetContainerLogsAsync (id,paramLog,CancellationToken.None)
+//// |> Async.AwaitTask
+
+
+
+//// return logs
+//// }
+//// |> Async.RunSynchronously
+
+
+////let tmp : array = Array.zeroCreate 1024
+////result.Read(tmp,0,1024)
+
+////System.Text.Encoding.UTF8.GetString(tmp,0,1024)
+
+// Include CsbScaffold
+//#load "../../.env/CsbScaffold.fsx"
+//#r @"C:\Users\Kevin\source\repos\CSBiology\BioFSharp\bin\BioFSharp.BioTools\net47\BioFSharp.BioTools.dll"
+
+//open BioFSharp.BioTools
+
+//open BioFSharp.IO
+//open BioFSharp.IO.BlastNCBI
+//open BioFSharp.IO.BlastNCBI.Parameters
+
+//let typeOfDatabase = Parameters.MakeDbParams.DbType Parameters.Protein
+
+
+//BlastWrapper(@"C:\Users\Kevin\source\repos\CSBiology\BioFSharp\lib\ncbi-blast\bin").makeblastdb @"C:\Users\Kevin\Source\Repos\CsbScaffold\Docker\data\Chlamy_Cp.fastA" ([typeOfDatabase;] |> seq)
+
+//let outputFormat=
+
+// [
+// OutputCustom.Query_SeqId;
+// OutputCustom.Subject_SeqId;
+// OutputCustom.Query_Length;
+// OutputCustom.Subject_Length;
+// OutputCustom.AlignmentLength;
+// OutputCustom.MismatchCount;
+// OutputCustom.IdentityCount;
+// OutputCustom.PositiveScoringMatchCount;
+// OutputCustom.Evalue;
+// OutputCustom.Bitscore;
+// ]
+// |> List.toSeq
+
+//let outputType = OutputType.TabularWithComments
+
+//let customOutputFormat = OutputTypeCustom(outputType , outputFormat)
+
+
+
+//BlastNCBI.BlastWrapper(@"C:\Users\Kevin\source\repos\CSBiology\BioFSharp\lib\ncbi-blast\bin")
+// .blastP
+// @"C:\Users\Kevin\Source\Repos\CsbScaffold\Docker\data\Chlamy_Cp.fastA"
+// @"C:\Users\Kevin\Source\Repos\CsbScaffold\Docker\data\testQuery.fastA"
+// @"C:\Users\Kevin\Source\Repos\CsbScaffold\Docker\data/Output.txt"
+// ([customOutputFormat;] |> seq)
+
+
+
+open FSharpAux
+open FSharpAux.IO
+open FSharpAux.IO.SchemaReader.Attribute
+open System.IO
+open BioFSharp.BioTools.BioContainer
+open BioFSharp.BioTools.BioContainerIO
+open Blast
+
+let client = Docker.connect "npipe://./pipe/docker_engine"
+
+let ImageBlast = Docker.DockerId.ImageId "blast"
+
+let blastContext =
+ BioContainer.initBcContextWithMountAsync client ImageBlast @"C:\Users\Kevin\source\repos\CsbScaffold\Docker\data"
+ |> Async.RunSynchronously
+
+let paramz =
+ [
+ MakeDbParams.DbType Protein
+ MakeDbParams.Input @"C:\Users\Kevin\source\repos\CsbScaffold\Docker\data\Chlamy_Cp.fastA"
+ MakeDbParams.Output@"C:\Users\Kevin\source\repos\CsbScaffold\Docker\data\Chlamy_Cp.fastA"
+ ]
+
+let outputFormat=
+
+ [
+ OutputCustom.Query_SeqId;
+ OutputCustom.Subject_SeqId;
+ OutputCustom.Query_Length;
+ OutputCustom.Subject_Length;
+ OutputCustom.AlignmentLength;
+ OutputCustom.MismatchCount;
+ OutputCustom.IdentityCount;
+ OutputCustom.PositiveScoringMatchCount;
+ OutputCustom.Evalue;
+ OutputCustom.Bitscore;
+ ]
+
+let blastPParamz = [
+ BlastParams.SearchDB @"C:\Users\Kevin\source\repos\CsbScaffold\Docker\data\Chlamy_Cp.fastA"
+ BlastParams.Query @"C:\Users\Kevin\source\repos\CsbScaffold\Docker\data\testQuery.fastA"
+ BlastParams.Output @"C:\Users\Kevin\source\repos\CsbScaffold\Docker\data\Output.txt"
+ OutputTypeCustom
+ (
+ OutputType.TabularWithComments,
+ [
+ OutputCustom.Query_SeqId;
+ OutputCustom.Subject_SeqId;
+ OutputCustom.Query_Length;
+ OutputCustom.Subject_Length;
+ OutputCustom.AlignmentLength;
+ OutputCustom.MismatchCount;
+ OutputCustom.IdentityCount;
+ OutputCustom.PositiveScoringMatchCount;
+ OutputCustom.Evalue;
+ OutputCustom.Bitscore;
+ ]
+ )
+]
+
+runMakeBlastDBAsync blastContext paramz
+|> Async.RunSynchronously
+
+runMakeBlastDB blastContext paramz
+
+runBlastPAsync blastContext blastPParamz
+|> Async.RunSynchronously
+
+
+BioContainer.execAsync blastContext ["makeblastdb"; "-dbtype"; "prot" ;"-in"; "/data/C/Users/Kevin/Source/Repos/CsbScaffold/Docker/data/Chlamy_Cp.fastA"; "-out"; "/data/C/Users/Kevin/Source/Repos/CsbScaffold/Docker/data/Chlamy_Cp.fastA"]
+|> Async.RunSynchronously
+
+BioContainer.disposeAsync blastContext
+|> Async.RunSynchronously
+
+
+open ClustalO
+
+let clustalImage = Docker.ImageName "clustal-omega"
+
+let clustalContext =
+ BioContainer.initBcContextWithMountAsync client clustalImage @"C:\Users\Kevin\source\repos\CsbScaffold\Docker\data"
+ |> Async.RunSynchronously
+
+// ClustalO tests
+let clustalOParamz = [
+ ClustalOParams.Input
+ (
+ FileInput.SequenceFile @"C:\Users\Kevin\source\repos\CsbScaffold\Docker\data\Chlamy_Cp.fastA",
+ [
+ InputCustom.Format FileFormat.FastA
+ ]
+ )
+ ClustalOParams.Output
+ (
+ @"C:\Users\Kevin\source\repos\CsbScaffold\Docker\data\Chlamy_Cp.aln",
+ []
+ )
+ ClustalOParams.Miscellaneous
+ [
+ MiscellaneousCustom.Force
+ ]
+]
+
+runClustalO clustalContext clustalOParamz
+
+open HMMER
+open HMMER.HMMbuild
+
+let HMMERImage = Docker.ImageName "hmmer"
+
+let hmmerContext =
+ BioContainer.initBcContextWithMountAsync client HMMERImage @"C:\Users\Kevin\source\repos\CsbScaffold\Docker\data"
+ |> Async.RunSynchronously
+
+let hmmbuildParamz =
+ [
+ InputMSAFile @"C:\Users\Kevin\source\repos\CsbScaffold\Docker\data\hmmer_testfiles\globins4.sto"
+ OutputHMMFile @"C:\Users\Kevin\source\repos\CsbScaffold\Docker\data\hmmer_testfiles\testOutput.hmm"
+ ]
+
+runHMMbuild hmmerContext hmmbuildParamz
\ No newline at end of file
diff --git a/src/BioFSharp.BioTools/Blast.fs b/src/BioFSharp.BioTools/Blast.fs
new file mode 100644
index 00000000..1cd46b23
--- /dev/null
+++ b/src/BioFSharp.BioTools/Blast.fs
@@ -0,0 +1,245 @@
+namespace BioFSharp.BioTools
+
+module Blast =
+
+ open FSharpAux
+ open FSharpAux.IO
+ open FSharpAux.IO.SchemaReader.Attribute
+ open BioContainer
+ open BioContainerIO
+
+ type DbType =
+ | Protein
+ | Nucleotide
+
+ static member make = function
+ | Protein -> "prot"
+ | Nucleotide -> "nucl"
+
+
+ type MakeDbParams =
+ | Input of string
+ | Output of string
+ | DbType of DbType
+ | MaskData of string
+ | ParseSeqIds
+
+ static member makeCmdWith (m: MountInfo) = function
+ | Input (path) -> ["-in" ;(MountInfo.containerPathOf m path)]
+ | Output (path) -> ["-out" ;(MountInfo.containerPathOf m path)]
+ | DbType (dbt) -> ["-dbtype"; (DbType.make dbt)]
+ | MaskData (path) -> ["-mask_data"; (sprintf "%s.asnb") (MountInfo.containerPathOf m path)]
+ | ParseSeqIds -> ["-parse_seqids"]
+
+
+ type OutputType =
+ | Pairwise
+ | Query_anchored
+ | Query_anchored_NoIdentities
+ | Query_anchored_Flat
+ | Query_anchored_Flat_NoIdentities
+ | XML
+ | Tabular
+ | TabularWithComments
+ | TextASN1
+ | BinaryASN1
+ | CSV
+ | BLAST_ArchiveFormat
+ | JSON_Seqalign
+ | JSON_Blast
+ | XML2_Blast
+
+ static member make = function
+ | Pairwise -> 0
+ | Query_anchored -> 1
+ | Query_anchored_NoIdentities -> 2
+ | Query_anchored_Flat -> 3
+ | Query_anchored_Flat_NoIdentities -> 4
+ | XML -> 5
+ | Tabular -> 6
+ | TabularWithComments -> 7
+ | TextASN1 -> 8
+ | BinaryASN1 -> 9
+ | CSV -> 10
+ | BLAST_ArchiveFormat -> 11
+ | JSON_Seqalign -> 12
+ | JSON_Blast -> 13
+ | XML2_Blast -> 14
+
+ //When not provided, the default value is:
+ //'qseqid sseqid pident length mismatch gapopen qstart qend sstart send
+ //evalue bitscore', which is equivalent to the keyword 'std'
+ type OutputCustom =
+ | Query_SeqId
+ | Query_GI
+ | Query_Accesion
+ | Query_Accesion_Version
+ | Query_Length
+ | Subject_SeqId
+ | Subject_All_SeqIds
+ | Subject_GI
+ | Subject_All_GIs
+ | Subject_Accession
+ | Subject_Accession_Version
+ | Subject_All_Accession
+ | Subject_Length
+ | Query_StartOfAlignment
+ | Query_EndOfAlignment
+ | Subject_StartOfAlignment
+ | Subject_EndOfAlignment
+ | Query_AlignedPartOf
+ | Subject_AlignedPartOf
+ | Evalue
+ | Bitscore
+ | RawScore
+ | AlignmentLength
+ | Identity
+ | IdentityCount
+ | MismatchCount
+ | PositiveScoringMatchCount
+ | GapOpeningCount
+ | GapCount
+ | PositiveScoringMatch
+ //means Query and subject frames separated by a '/'
+ | Frames
+ | Query_Frames
+ | Subject_Frames
+ //means Blast traceback operations (BTOP)
+ | BTOP
+ | Subject_TaxonomyIDs
+ | Subject_Scientific_Names
+ | Subject_Common_Names
+ | Subject_Blast_Names
+ | Subject_Super_Kingdoms
+ | Subject_Title
+ | Subject_All_Titles
+ | Subject_Strand
+ | Query_CoveragePerSubject
+ | Query_CoveragePerHSP
+
+ static member make = function
+ | Query_SeqId -> "qseqid"
+ | Query_GI -> "qgi"
+ | Query_Accesion -> "qacc"
+ | Query_Accesion_Version -> "qaccver"
+ | Query_Length -> "qlen"
+ | Subject_SeqId -> "sseqid"
+ | Subject_All_SeqIds -> "sallseqid"
+ | Subject_GI -> "sgi"
+ | Subject_All_GIs -> "sallgi"
+ | Subject_Accession -> "sacc"
+ | Subject_Accession_Version -> "saccver"
+ | Subject_All_Accession -> "sallacc"
+ | Subject_Length -> "slen"
+ | Query_StartOfAlignment -> "qstart"
+ | Query_EndOfAlignment -> "qend"
+ | Subject_StartOfAlignment -> "sstart"
+ | Subject_EndOfAlignment -> "send"
+ | Query_AlignedPartOf -> "qseq"
+ | Subject_AlignedPartOf -> "sseq"
+ | Evalue -> "evalue"
+ | Bitscore -> "bitscore"
+ | RawScore -> "score"
+ | AlignmentLength -> "length"
+ | Identity -> "pident"
+ | IdentityCount -> "nident"
+ | MismatchCount -> "mismatch"
+ | PositiveScoringMatchCount -> "positive"
+ | GapOpeningCount -> "gapopen"
+ | GapCount -> "gaps"
+ | PositiveScoringMatch -> "ppos"
+ //means Query and subject frames separated by a '/'
+ | Frames -> "frames"
+ | Query_Frames -> "qframe"
+ | Subject_Frames -> "sframe"
+ //means Blast traceback operations (BTOP)
+ | BTOP -> "btop"
+ | Subject_TaxonomyIDs -> "staxids"
+ | Subject_Scientific_Names -> "sscinames"
+ | Subject_Common_Names -> "scomnames"
+ | Subject_Blast_Names -> "sblastnames"
+ | Subject_Super_Kingdoms -> "sskingdoms"
+ | Subject_Title -> "stitle"
+ | Subject_All_Titles -> "salltitles"
+ | Subject_Strand -> "sstrand"
+ | Query_CoveragePerSubject -> "qcovs"
+ | Query_CoveragePerHSP -> "qcovhsp"
+
+
+ type BlastParams =
+ | SearchDB of string
+ | Query of string
+ | Output of string
+ | OutputType of OutputType
+ | OutputTypeCustom of OutputType * seq
+ | Num_threads of int
+ | Max_Hits of int
+
+ static member makeCmdWith (m: MountInfo) = function
+ | SearchDB (path) -> ["-db" ; (MountInfo.containerPathOf m path)]
+ | Query (path) -> ["-query" ; (MountInfo.containerPathOf m path)]
+ | Output (path) -> ["-out" ; (MountInfo.containerPathOf m path)]
+ | OutputType(format) -> ["-outfmt"; string (format |> OutputType.make)]
+ | OutputTypeCustom(t,p) -> let tmp =
+ p
+ |> Seq.map OutputCustom.make
+ |> String.concat " "
+ match t with
+ | OutputType.Tabular -> ["-outfmt"; sprintf "%s %s" (string (t |> OutputType.make)) tmp]
+ | OutputType.TabularWithComments -> ["-outfmt"; sprintf "%s %s" (string (t |> OutputType.make)) tmp]
+ | OutputType.CSV -> ["-outfmt"; sprintf "%s %s" (string (t |> OutputType.make)) tmp]
+ | _ -> failwithf "Output format %A does not support custom columns." t
+ | Num_threads(i) -> ["-num_threads"; string i]
+ | Max_Hits (i) -> ["-max_target_seqs"; string i]
+
+
+ let runMakeBlastDBAsync (bcContext:BioContainer.BcContext) (opt:MakeDbParams list) =
+
+ let cmds = (opt |> List.map (MakeDbParams.makeCmdWith bcContext.Mount))
+ let tp = "makeblastdb"::(cmds |> List.concat)
+
+ printfn "Starting process makeblastdb\r\nparameters:"
+ cmds |> List.iter (fun op -> printfn "\t%s" (String.concat " " op))
+
+ async {
+ let! res = BioContainer.execAsync bcContext tp
+ return res
+ }
+
+ let runMakeBlastDB (bcContext:BioContainer.BcContext) (opt:MakeDbParams list) =
+
+ runMakeBlastDBAsync bcContext opt
+ |> Async.RunSynchronously
+
+ let runBlastPAsync (bcContext:BioContainer.BcContext) (opt:BlastParams list) =
+ let cmds = (opt |> List.map (BlastParams.makeCmdWith bcContext.Mount))
+ let tp = "blastp"::(cmds |> List.concat)
+
+ printfn "Starting process blastp\r\nparameters:"
+ cmds |> List.iter (fun op -> printfn "\t%s" (String.concat " " op))
+
+ async {
+ let! res = BioContainer.execAsync bcContext tp
+ return res
+ }
+
+ let runBlastP (bcContext:BioContainer.BcContext) (opt:BlastParams list) =
+ runBlastPAsync bcContext opt
+ |> Async.RunSynchronously
+
+ let runBlastNAsync (bcContext:BioContainer.BcContext) (opt:BlastParams list) =
+ let cmds = (opt |> List.map (BlastParams.makeCmdWith bcContext.Mount))
+ let tp = "blastn"::(cmds |> List.concat)
+
+ printfn "Starting process blastn\r\nparameters:"
+ cmds |> List.iter (fun op -> printfn "\t%s" (String.concat " " op))
+
+ async {
+ let! res = BioContainer.execAsync bcContext tp
+ return res
+
+ }
+
+ let runBlastN (bcContext:BioContainer.BcContext) (opt:BlastParams list) =
+ runBlastNAsync bcContext opt
+ |> Async.RunSynchronously
\ No newline at end of file
diff --git a/src/BioFSharp.BioTools/ClustalO.fs b/src/BioFSharp.BioTools/ClustalO.fs
new file mode 100644
index 00000000..423229d0
--- /dev/null
+++ b/src/BioFSharp.BioTools/ClustalO.fs
@@ -0,0 +1,401 @@
+namespace BioFSharp.BioTools
+
+module ClustalO =
+
+ open FSharpAux
+ open BioContainer
+
+// SEQUENCE INPUT
+// -i, --in, --infile={,-}
+// Multiple sequence input file (- for stdin)
+//
+// --hmm-in=
+// HMM input files
+//
+// --dealign
+// Dealign input sequences
+//
+// --profile1, --p1=
+// Pre-aligned multiple sequence file (aligned columns will be kept fixed)
+//
+// --profile2, --p2=
+// Pre-aligned multiple sequence file (aligned columns will be kept fixed)
+//
+// --is-profile
+// disable check if profile, force profile (default no)
+//
+// -t, --seqtype={Protein, RNA, DNA}
+// Force a sequence type (default: auto)
+//
+// --infmt={a2m=fa[sta],clu[stal],msf,phy[lip],selex,st[ockholm],vie[nna]}
+// Forced sequence input file format (default: auto)
+
+
+ type FileFormat =
+ ///FastA file format
+ | FastA
+ ///Clustal file format
+ | Clustal
+ ///MSF file format
+ | MSF
+ ///Phylip file format
+ | Phylip
+ ///Selex file format
+ | Selex
+ ///Stockholm file format
+ | Stockholm
+ ///Vienna file format
+ | Vienna
+ static member make = function
+ | FastA -> "fa"
+ | Clustal -> "clu"
+ | MSF -> "msf"
+ | Phylip -> "phy"
+ | Selex -> "selex"
+ | Stockholm -> "st"
+ | Vienna -> "vie"
+
+ ///Types of sequences
+ type SeqType =
+ | Protein
+ | DNA
+ | RNA
+
+ static member make = function
+ | Protein -> "--seqtype=Protein"
+ | RNA -> "--seqtype=RNA"
+ | DNA -> "--seqtype=DNA"
+
+ type InputCustom =
+ ///Forced sequence input file format (default: auto)
+ | Format of FileFormat
+ ///Dealign input sequences
+ | Dealign
+ ///Disable check if profile, force profile (default no)
+ | IsProfile
+ ///Force a sequence type (default: auto)
+ | SeqType of SeqType
+
+ static member make = function
+ | Format f -> sprintf "--infmt=%s" (FileFormat.make f)
+ | Dealign -> "--dealign "
+ | IsProfile -> "--is-profile "
+ | SeqType s -> SeqType.make s
+
+ ///Specify the type of input and assign file path
+ type FileInput =
+ ///Use this option to make a multiple alignment from a set of sequences. A sequence file must contain more than one sequence (at least two sequences).
+ | SequenceFile of string
+ ///Use this option to align two alignments (profiles) together.
+ | TwoProfiles of string * string
+ /// Use this option to add new sequences to an existing alignment.
+ | SequenceFileAndProfile of string * string
+ /// Use this option to make a new multiple alignment of sequences from the input file and use the HMM as a guide (EPA).
+ | SequenceFileAndHMM of string * string
+
+ static member make = function
+ | SequenceFile path -> ["-i"; path]
+ | TwoProfiles (path1,path2) -> [sprintf "--p1=%s" path1 ; sprintf"--p2=%s" path2]
+ | SequenceFileAndProfile (path1,path2) -> ["-i"; path1; sprintf "--p1=%s " path2]
+ | SequenceFileAndHMM (path1,path2) -> ["-i"; path1; sprintf "--hmm-in=%s" path2]
+
+ static member makeWith (m: MountInfo) =
+ let cPath p = (MountInfo.containerPathOf m p)
+ function
+ | SequenceFile path -> ["-i"; (MountInfo.containerPathOf m path)]
+ | TwoProfiles (path1,path2) -> [sprintf "--p1=%s" (cPath path1) ; sprintf "--p2=%s" (cPath path2)]
+ | SequenceFileAndProfile (path1,path2) -> ["-i"; (cPath path1); sprintf "--p1=%s " (cPath path2)]
+ | SequenceFileAndHMM (path1,path2) -> ["-i"; (cPath path1); sprintf "--hmm-in=%s" (cPath path2)]
+
+// ALIGNMENT OUTPUT
+// -o, --out, --outfile={file,-}
+// Multiple sequence alignment output file (default: stdout)
+//
+// --outfmt={a2m=fa[sta],clu[stal],msf,phy[lip],selex,st[ockholm],vie[nna]}
+// MSA output file format (default: fasta)
+//
+// --residuenumber, --resno
+// in Clustal format print residue numbers (default no)
+//
+// --wrap=
+// number of residues before line-wrap in output
+//
+// --output-order={input-order,tree-order}
+// MSA output order like in input/guide-tree
+
+ ///Optional modifiers for input
+ type OutputCustom =
+ /// MSA output file format (default: fasta)
+ | Format of FileFormat
+ /// in Clustal format print residue numbers (default no)
+ | ResidueNumber
+ /// number of residues before line-wrap in output
+ | Wrap of int
+ /// Aligned sequences are ordered according to guide tree instead of input order
+ | OutputOrderAsTree
+
+ static member make = function
+ | Format f -> sprintf "--outfmt=%s" (FileFormat.make f)
+ | ResidueNumber -> "--residuenumber"
+ | Wrap i -> sprintf "--wrap=%i" i
+ | OutputOrderAsTree -> "--output-order=tree-order"
+
+
+
+
+ ///Collection of parameters for specifying clustalo alignment
+// CLUSTERING
+// --distmat-in=
+// Pairwise distance matrix input file (skips distance computation)
+//
+// --distmat-out=
+// Pairwise distance matrix output file
+//
+// --guidetree-in=
+// Guide tree input file
+// (skips distance computation and guide tree clustering step)
+//
+// --guidetree-out=
+// Guide tree output file
+//
+// --full
+// Use full distance matrix for guide-tree calculation (slow; mBed is default)
+//
+// --full-iter
+// Use full distance matrix for guide-tree calculation during iteration (mBed is default)
+//
+// --cluster-size=
+// soft maximum of sequences in sub-clusters
+//
+// --clustering-out=
+// Clustering output file
+//
+// --use-kimura
+// use Kimura distance correction for aligned sequences (default no)
+//
+// --percent-id
+// convert distances into percent identities (default no)
+//
+
+ ///Optional modifiers to specify clustering
+ type ClusteringCustom =
+ ///Pairwise distance matrix input file (skips distance computation)
+ | DistanceMatrixInput of string
+ ///Pairwise distance matrix output file
+ | DistanceMatrixOutput of string
+ ///Guide tree input file (skips distance computation and guide tree clustering step)
+ | GuideTreeInput of string
+ ///Guide tree output file
+ | GuideTreeOutput of string
+ ///Use full distance matrix for guide-tree calculation (slow; mBed is default)
+ | Full
+ ///Use full distance matrix for guide-tree calculation during iteration (mBed is default)
+ | FullIter
+ /// Soft maximum of sequences in sub-clusters
+ | ClusterSize of int
+ /// Clustering output file
+ | ClusteringOut of string
+ /// Use Kimura distance correction for aligned sequences (default no)
+ | UseKimura
+ /// convert distances into percent identities (default no)
+ | PercentID
+
+ static member make = function
+ | DistanceMatrixInput path -> [sprintf "--distmat-in=%s" path ]
+ | DistanceMatrixOutput path -> [sprintf "--distmat-out=%s" path ]
+ | GuideTreeInput path -> [sprintf "--guidetree-in=%s" path ]
+ | GuideTreeOutput path -> [sprintf "--guidetree-out=%s" path ]
+ | Full -> ["--full" ]
+ | FullIter -> ["--full-iter" ]
+ | ClusterSize i -> [sprintf "--cluster-size=%i" i ]
+ | ClusteringOut path -> [sprintf "--clustering-out=%s" path ]
+ | UseKimura -> ["--use-kimura" ]
+ | PercentID -> ["--percent-id" ]
+
+ static member makeWith (m: MountInfo) =
+ let cPath p = (MountInfo.containerPathOf m p)
+ function
+ | DistanceMatrixInput path -> [sprintf "--distmat-in=%s" (cPath path) ]
+ | DistanceMatrixOutput path -> [sprintf "--distmat-out=%s" (cPath path) ]
+ | GuideTreeInput path -> [sprintf "--guidetree-in=%s" (cPath path) ]
+ | GuideTreeOutput path -> [sprintf "--guidetree-out=%s" (cPath path) ]
+ | Full -> ["--full"]
+ | FullIter -> ["--full-iter"]
+ | ClusterSize i -> [sprintf "--cluster-size=%i" i]
+ | ClusteringOut path -> [sprintf "--clustering-out=%s" (cPath path) ]
+ | UseKimura -> ["--use-kimura"]
+ | PercentID -> ["--percent-id"]
+
+//ITERATION:
+//
+// --iterations, --iter= Number of (combined guide tree/HMM) iterations
+//
+// --max-guidetree-iterations= Maximum guide tree iterations
+//
+// --max-hmm-iterations= Maximum number of HMM iterations
+
+ ///Specify maximum number of iterations for given step
+ type IterationCustom =
+ /// Number of (combined guide tree/HMM) iterations
+ | Iterations of int
+ /// Maximum guide tree iterations
+ | MaxGuideTreeIterations of int
+ /// Maximum number of HMM iterations
+ | MaxHMMIterations of int
+
+ static member make = function
+ | Iterations i -> [sprintf "--iter=%i" i]
+ | MaxGuideTreeIterations i -> [sprintf "--max-guidetree-iterations=%i" i]
+ | MaxHMMIterations i -> [sprintf "--max-hmm-iterations=%i" i]
+
+
+//LIMITS (will exit early, if exceeded):
+//
+// --maxnumseq= Maximum allowed number of sequences
+//
+// --maxseqlen= Maximum allowed sequence length
+ /// Will exit early, if exceeded
+ type LimitsCustom =
+ /// Maximum allowed number of sequences
+ | MaxSeqNumber of int
+ /// Maximum allowed sequence length
+ | MaxSeqLength of int
+
+ static member make = function
+ | MaxSeqNumber i -> [sprintf "--maxnumseq=%i" i]
+ | MaxSeqLength i -> [sprintf "--maxseqlen=%i" i]
+
+
+//MISCELLANEOUS:
+//
+// --auto Set options automatically (might overwrite some of your options)
+//
+// --threads= Number of processors to use
+//
+// -l, --log= Log all non-essential output to this file
+//
+// -h, --help Print help and exit
+//
+// -v, --verbose Verbose output (increases if given multiple times)
+//
+// --version Print version information and exit
+//
+// --long-version Print long version information and exit
+//
+// --force Force file overwriting
+
+ ///Optional, miscallaneous modifiers
+ type MiscellaneousCustom =
+ /// Set options automatically (might overwrite some of your options)
+ | Auto
+ /// Number of processors to use
+ | Threads of int
+ /// Log all non-essential output to this file
+ | Log of string
+ /// Print help and exit
+ //| Help
+ /// Verbose output (ranging from 0 [nonverbose,standard] to 3 [very verbose,everything above 3 is set to 3])
+ | VerboseLevel of int
+ /// Print version information and exit
+ | Version
+ /// Print long version information and exit
+ | LongVersion
+ /// Force file overwriting
+ | Force
+
+ static member make = function
+ | Auto -> ["--auto"]
+ | Threads i -> [sprintf "--threads=%i" i]
+ | Log s -> [sprintf "--log=%s" s]
+ //| Help -> "--help "
+ | VerboseLevel i->
+ if i > 0 && i < 4 then
+ [for n = 0 to i-1 do yield "-v"]
+ elif i > 3 then
+ ["-v"; "-v"; "-v"]
+ else
+ []
+ | Version -> ["--version"]
+ | LongVersion -> ["--long-version"]
+ | Force -> ["--force"]
+
+ static member makeWith (m: MountInfo) =
+ let cPath p = (MountInfo.containerPathOf m p)
+ function
+ | Auto -> ["--auto"]
+ | Threads i -> [sprintf "--threads=%i" i]
+ | Log path -> [sprintf "--log=%s" (cPath path)]
+ //| Help -> "--help "
+ | VerboseLevel i->
+ if i > 0 then
+ [for n = 0 to i-1 do yield "-v"]
+ else
+ []
+ | Version -> ["--version"]
+ | LongVersion -> ["--long-version"]
+ | Force -> ["--force"]
+
+ type ClustalOParams =
+ /// Specify input parameters
+ | Input of FileInput * InputCustom list
+ /// Specify output parameters
+ | Output of string * OutputCustom list
+ /// Specify clustering parameters
+ | Clustering of ClusteringCustom list
+ /// Specify iteration parameters
+ | Iteration of IterationCustom list
+ /// Specify limits parameters
+ | Limits of LimitsCustom list
+ /// Specify miscallaneous parameters
+ | Miscellaneous of MiscellaneousCustom list
+
+ static member makeCmd = function
+ | Input (i,p) -> let tmp =
+ p
+ |> List.map InputCustom.make
+ (FileInput.make i)@tmp
+ | Output (o,p) -> let tmp =
+ p
+ |> List.map OutputCustom.make
+ ["-o"; o]@tmp
+ | Clustering cl -> cl |> List.map ClusteringCustom.make |> List.concat
+ | Iteration it -> it |> List.map IterationCustom.make |> List.concat
+ | Limits l -> l |> List.map LimitsCustom.make |> List.concat
+ | Miscellaneous misc-> misc |> List.map MiscellaneousCustom.make|> List.concat
+
+
+ static member makeCmdWith (m: MountInfo) =
+ let cPath p = (MountInfo.containerPathOf m p)
+ function
+ | Input (i,p) -> let tmp =
+ p
+ |> List.map InputCustom.make
+ (FileInput.makeWith m i)@tmp
+ | Output (o,p) -> let tmp =
+ p
+ |> List.map OutputCustom.make
+ ["-o"; (cPath o)]@tmp
+ | Clustering cl -> cl |> List.map (ClusteringCustom.makeWith m) |> List.concat
+ | Iteration it -> it |> List.map IterationCustom.make |> List.concat
+ | Limits l -> l |> List.map LimitsCustom.make |> List.concat
+ | Miscellaneous misc-> misc |> List.map (MiscellaneousCustom.makeWith m)|> List.concat
+
+
+ let runClustalOAsync (bcContext:BioContainer.BcContext) (opt:ClustalOParams list) =
+
+ let cmds = (opt |> List.map (ClustalOParams.makeCmdWith bcContext.Mount))
+ let tp = "clustalo"::(cmds |> List.concat)
+
+ printfn "Starting process clustalo\r\nparameters:"
+ cmds |> List.iter (fun op -> printfn "\t%s" (String.concat " " op))
+
+ async {
+ let! res = BioContainer.execAsync bcContext tp
+ return res
+ }
+
+ let runClustalO (bcContext:BioContainer.BcContext) (opt:ClustalOParams list) =
+ runClustalOAsync bcContext opt
+ |> Async.RunSynchronously
+
+
diff --git a/src/BioFSharp.BioTools/Docker.fs b/src/BioFSharp.BioTools/Docker.fs
new file mode 100644
index 00000000..818341ac
--- /dev/null
+++ b/src/BioFSharp.BioTools/Docker.fs
@@ -0,0 +1,945 @@
+namespace BioFSharp.BioTools
+
+open System
+open System.Threading
+open System.Collections.Generic
+open Docker.DotNet
+open Docker.DotNet.Models
+
+
+// https://docs.docker.com/engine/api/v1.24/
+
+/// Docker helper
+module Docker =
+
+
+ /// Dockerfile > (Build) > Image > (Create/Run) > Container <- start/stop
+ type DockerId =
+ | ImageId of string
+ | ImageName of string
+ | ContainerId of string
+ | ContainerName of string
+ | Tag of string*string
+
+ override this.ToString() =
+ match this with
+ | ImageId s -> s
+ | ImageName s -> s
+ | ContainerId s -> s
+ | ContainerName s -> s
+ | Tag (s,t) -> sprintf "%s:%s" s t
+
+ /// Dockerfile > (Build) > Image > (Create/Run) > Container <- start/stop
+ type ContainerStatus =
+ | Created
+ | Restarting
+ | Running
+ | Paused
+ | Exited
+ | Dead
+
+
+ override this.ToString() =
+ match this with
+ | Created -> "created"
+ | Restarting -> "restarting"
+ | Running -> "running"
+ | Paused -> "paused"
+ | Exited -> "exited"
+ | Dead -> "dead"
+
+
+ /// Provides a set of static methods for creating Image parameter.
+ type Filters =
+
+ static member private DicOfSeq (s:('k * 'v) seq) = new Dictionary<'k,'v>(s |> Map.ofSeq) :> IDictionary<'k,'v>
+
+ /// Creates ImagesListParameters for pre-filtering list function
+ static member InitContainerFilters
+ (
+ ?Ancestor:DockerId,
+ ?Status:ContainerStatus,
+ // Containers that exited with status code
+ ?Exited:int,
+ ?Label,
+ ?Isolation,
+ ?Before:DockerId,
+ ?Since:DockerId,
+ ?Volume,
+ ?Network
+ ) =
+
+ let filter = System.Collections.Generic.Dictionary()
+ Ancestor |> Option.iter (fun nv -> filter.Add( "ancestor", [nv.ToString(),true] |> Filters.DicOfSeq) )
+ Status |> Option.iter (fun nv -> filter.Add( "status", [nv.ToString(),true] |> Filters.DicOfSeq) )
+ Exited |> Option.iter (fun nv -> filter.Add( "exited", [nv.ToString(),true] |> Filters.DicOfSeq) )
+ Label |> Option.iter (fun nv -> filter.Add( "label", [nv.ToString(),true] |> Filters.DicOfSeq) )
+ Isolation |> Option.iter (fun nv -> filter.Add( "isolation", [nv.ToString(),true] |> Filters.DicOfSeq) )
+ Before |> Option.iter (fun nv -> filter.Add( "before", [nv.ToString(),true] |> Filters.DicOfSeq) )
+ Since |> Option.iter (fun nv -> filter.Add( "since", [nv.ToString(),true] |> Filters.DicOfSeq) )
+ Volume |> Option.iter (fun nv -> filter.Add( "volume", [nv.ToString(),true] |> Filters.DicOfSeq) )
+ Network |> Option.iter (fun nv -> filter.Add( "network", [nv.ToString(),true] |> Filters.DicOfSeq) )
+
+ filter
+
+
+ /// Connect to docker engine (docker deamon)
+ let connect str =
+ (new DockerClientConfiguration(new Uri(str)) ).CreateClient()
+
+ module Image =
+
+
+ /// Provides a set of static methods for creating Image parameter.
+ type ImagesParams =
+
+ /// Creates ImagesListParameters for pre-filtering list function
+ static member InitImagesListParameters
+ (
+ ?All,
+ ?Filters,
+ ?MatchName
+ ) =
+
+ let param = new ImagesListParameters()
+ All |> Option.iter (fun nv -> param.set_All (Nullable(nv)))
+ Filters |> Option.iter param.set_Filters
+ MatchName |> Option.iter param.set_MatchName
+
+ param
+
+ /// Creates ImagesCreateParameters
+ static member InitImagesCreateParameters
+ (
+ ?FromImage,
+ ?FromSrc,
+ ?Repo,
+ ?Tag
+ ) =
+
+ let param = new ImagesCreateParameters()
+ FromImage |> Option.iter param.set_FromImage
+ FromSrc |> Option.iter param.set_FromSrc
+ Repo |> Option.iter param.set_Repo
+ Tag |> Option.iter param.set_Tag
+
+ param
+
+ ///// Creates ImageBuildParameters
+ //static member InitImageBuildParameters
+ // (
+ // ?FromImage,
+ // ?FromSrc,
+ // ?Repo,
+ // ?Tag
+ // ) =
+
+ // let param = new ImageBuildParameters()
+
+ // FromImage |> Option.iter param.set_FromImage
+ // FromSrc |> Option.iter param.set_FromSrc
+ // Repo |> Option.iter param.set_Repo
+ // Tag |> Option.iter param.set_Tag
+
+ // param
+
+ /// Lists available images with ImagesListParameters for filtering (async)
+ let listImagesWithAsync (connection:DockerClient) (param:ImagesListParameters) =
+ // ImagesListParameters are only for filtering
+ async {
+ let! tmp =
+ connection.Images. ListImagesAsync(param)
+ |> Async.AwaitTask
+ return (tmp |> Seq.map id)
+ }
+
+
+ /// Lists all available images (async)
+ let listImagesAsync (connection:DockerClient) =
+ listImagesWithAsync connection (ImagesParams.InitImagesListParameters())
+
+
+ /// Lists all available images
+ let listImages (connection:DockerClient) =
+ listImagesAsync connection
+ |> Async.RunSynchronously
+
+
+ /// Returns true if an image matches the dockerid (async)
+ let existsAsync (connection:DockerClient) (dockerid:DockerId) =
+ let param = ImagesParams.InitImagesListParameters(MatchName=dockerid.ToString())
+ async {
+ let! tmp = listImagesWithAsync connection param
+
+ return (Seq.length tmp > 0)
+ }
+
+
+ /// Returns true if an image matches the name
+ let exists (connection:DockerClient) (dockerid:DockerId) =
+ existsAsync connection dockerid
+ |> Async.RunSynchronously
+
+ /// Lists available images with ImagesListParameters for filtering (async)
+ let buildImageFromDockerfileAsync (connection:DockerClient) (param) (stream) =
+ // ImagesListParameters are only for filtering
+ async {
+ let! stream =
+ connection.Images.BuildImageFromDockerfileAsync(stream,param)
+ |> Async.AwaitTask
+ return stream
+ }
+
+ //#####################################################
+ //#####################################################
+ //#####################################################
+
+
+ module Container =
+
+
+ /// Provides a set of static methods for creating Image parameter.
+ type ContainerParams =
+
+ /// Creates ContainersListParameters for pre-filtering list function
+ static member InitContainerListParameters
+ (
+ ?All,
+ ?Before,
+ ?Filters,
+ ?Limit,
+ ?Since,
+ ?Size
+ ) =
+
+ let param = new ContainersListParameters()
+ All |> Option.iter (fun nv -> param.set_All (Nullable(nv)))
+ Before |> Option.iter param.set_Before
+ Filters |> Option.iter param.set_Filters
+ Limit |> Option.iter (fun nv -> param.set_Limit (Nullable(nv)))
+ Since |> Option.iter param.set_Since
+ Size |> Option.iter (fun nv -> param.set_Size (Nullable(nv)))
+
+ param
+
+ /// Creates ContainersListParameters for pre-filtering list function
+ static member InitContainerListParameters
+ (
+ ?WaitBeforeKillSeconds
+ ) =
+
+ let param = new ContainerStopParameters()
+
+ WaitBeforeKillSeconds |> Option.iter (fun nv -> param.set_WaitBeforeKillSeconds (Nullable(nv)))
+
+ param
+
+ /// Creates ContainerAttachParameters
+ static member InitContainerAttachParameters
+ (
+ ?DetachKeys,
+ ?Logs,
+ ?Stderr,
+ ?Stdin,
+ ?Stdout,
+ ?Stream
+ ) =
+
+ let param = new ContainerAttachParameters()
+ DetachKeys |> Option.iter param.set_DetachKeys
+ Logs |> Option.iter param.set_Logs
+ Stderr |> Option.iter (fun v -> param.set_Stderr (Nullable(v) ) )
+ Stdin |> Option.iter (fun v -> param.set_Stdin (Nullable(v) ) )
+ Stdout |> Option.iter (fun v -> param.set_Stdout (Nullable(v) ) )
+ Stream |> Option.iter (fun v -> param.set_Stream (Nullable(v) ) )
+
+ param
+
+ /// Creates ContainerExecStartParameters
+ static member InitContainerExecStartParameters
+ (
+
+ ?AttachStderr,
+ ?AttachStdin,
+ ?AttachStdout,
+ ?Cmd:seq,
+ ?Detach,
+ ?DetachKeys,
+ ?Env:seq,
+ ?Privileged,
+ ?Tty,
+ ?User
+ ) =
+
+ let param = new ContainerExecStartParameters()
+
+ AttachStderr |> Option.iter param.set_AttachStderr
+ AttachStdin |> Option.iter param.set_AttachStdin
+ AttachStdout |> Option.iter param.set_AttachStdout
+ Cmd |> Option.iter (fun v -> param.set_Cmd (Collections.Generic.List(v)) )
+ Detach |> Option.iter param.set_Detach
+ DetachKeys |> Option.iter param.set_DetachKeys
+ Env |> Option.iter (fun v -> param.set_Env (Collections.Generic.List(v)) )
+ Privileged |> Option.iter param.set_Privileged
+ Tty |> Option.iter param.set_Tty
+ User |> Option.iter param.set_User
+
+ param
+
+ /// Creates ContainerExecStartParameters
+ static member InitContainerExecCreateParameters
+ (
+
+ ?AttachStderr,
+ ?AttachStdin,
+ ?AttachStdout,
+ ?Cmd:seq,
+ ?Detach,
+ ?DetachKeys,
+ ?Env:seq,
+ ?Privileged,
+ ?Tty,
+ ?User
+ ) =
+
+ let param = new ContainerExecCreateParameters()
+
+ AttachStderr |> Option.iter param.set_AttachStderr
+ AttachStdin |> Option.iter param.set_AttachStdin
+ AttachStdout |> Option.iter param.set_AttachStdout
+ Cmd |> Option.iter (fun v -> param.set_Cmd (Collections.Generic.List(v)) )
+ Detach |> Option.iter param.set_Detach
+ DetachKeys |> Option.iter param.set_DetachKeys
+ Env |> Option.iter (fun v -> param.set_Env (Collections.Generic.List(v)) )
+ Privileged |> Option.iter param.set_Privileged
+ Tty |> Option.iter param.set_Tty
+ User |> Option.iter param.set_User
+
+ param
+
+ /// Creates BindOptions
+ static member InitBindOptions
+ (
+ ?Propagation
+ ) =
+
+ let param = new BindOptions()
+ Propagation |> Option.iter param.set_Propagation
+
+ param
+
+ /// Creates VolumeOptions
+ static member InitVolumeOptions
+ (
+ ?DriverConfig,
+ ?Labels,
+ ?NoCopy
+ ) =
+
+ let param = new VolumeOptions()
+ DriverConfig |> Option.iter param.set_DriverConfig
+ Labels |> Option.iter param.set_Labels
+ NoCopy |> Option.iter param.set_NoCopy
+
+ param
+
+ /// Creates TmpfsOptions
+ static member InitTmpfsOptions
+ (
+ ?Mode,
+ ?SizeBytes
+ ) =
+
+ let param = new TmpfsOptions()
+ Mode |> Option.iter param.set_Mode
+ SizeBytes |> Option.iter param.set_SizeBytes
+
+ param
+
+ /// Creates Mount
+ static member InitMount
+ (
+ ?Type ,
+ ?Source ,
+ ?Target ,
+ ?ReadOnly ,
+ ?BindOptions ,
+ ?VolumeOptions ,
+ ?TmpfsOptions
+ ) =
+
+ let param = new Mount()
+
+ Type |> Option.iter param.set_Type
+ Source |> Option.iter param.set_Source
+ Target |> Option.iter param.set_Target
+ ReadOnly |> Option.iter param.set_ReadOnly
+ BindOptions |> Option.iter param.set_BindOptions
+ VolumeOptions |> Option.iter param.set_VolumeOptions
+ TmpfsOptions |> Option.iter param.set_TmpfsOptions
+
+ param
+
+ /// Creates HostConfig
+ static member InitHostConfig
+ (
+ ?ContainerIDFile ,
+ ?LogConfig ,
+ ?NetworkMode ,
+ ?PortBindings ,
+ ?RestartPolicy ,
+ ?AutoRemove ,
+ ?VolumeDriver ,
+ ?VolumesFrom : seq,
+ ?CapAdd ,
+ ?CapDrop : seq,
+ ?DNS : seq,
+ ?DNSOptions : seq,
+ ?DNSSearch : seq,
+ ?ExtraHosts : seq,
+ ?GroupAdd : seq,
+ ?IpcMode ,
+ ?Cgroup ,
+ ?Links : seq,
+ ?OomScoreAdj ,
+ ?PidMode ,
+ ?Privileged ,
+ ?PublishAllPorts ,
+ ?ReadonlyRootfs ,
+ ?SecurityOpt : seq,
+ ?StorageOpt ,
+ ?Tmpfs ,
+ ?UTSMode ,
+ ?UsernsMode ,
+ ?ShmSize ,
+ ?Sysctls ,
+ ?Runtime ,
+ ?ConsoleSize ,
+ ?Isolation ,
+ ?CPUShares ,
+ ?Memory ,
+ ?NanoCPUs ,
+ ?CgroupParent ,
+ ?BlkioWeight ,
+ ?BlkioWeightDevice : seq<_>,
+ ?BlkioDeviceReadBps : seq<_>,
+ ?BlkioDeviceWriteBps : seq<_>,
+ ?BlkioDeviceReadIOps : seq<_>,
+ ?BlkioDeviceWriteIOps: seq<_>,
+ ?CPUPeriod ,
+ ?CPUQuota ,
+ ?CPURealtimePeriod ,
+ ?CPURealtimeRuntime ,
+ ?CpusetCpus ,
+ ?CpusetMems ,
+ ?Devices : seq<_>,
+ ?DiskQuota ,
+ ?KernelMemory ,
+ ?MemoryReservation ,
+ ?MemorySwap ,
+ ?MemorySwappiness ,
+ ?OomKillDisable ,
+ ?PidsLimit ,
+ ?Ulimits : seq,
+ ?CPUCount ,
+ ?CPUPercent ,
+ ?IOMaximumIOps ,
+ ?IOMaximumBandwidth ,
+ ?Mounts : seq,
+ ?Init ,
+ ?InitPath
+ ) =
+
+ let param = new HostConfig()
+
+ ContainerIDFile |> Option.iter param.set_ContainerIDFile
+ LogConfig |> Option.iter param.set_LogConfig
+ NetworkMode |> Option.iter param.set_NetworkMode
+ PortBindings |> Option.iter param.set_PortBindings
+ RestartPolicy |> Option.iter param.set_RestartPolicy
+ AutoRemove |> Option.iter param.set_AutoRemove
+ VolumeDriver |> Option.iter param.set_VolumeDriver
+ VolumesFrom |> Option.iter (fun v -> param.set_VolumesFrom (Collections.Generic.List(v)) )
+ CapAdd |> Option.iter param.set_CapAdd
+ CapDrop |> Option.iter (fun v -> param.set_CapDrop (Collections.Generic.List(v)) )
+ DNS |> Option.iter (fun v -> param.set_DNS (Collections.Generic.List(v)) )
+ DNSOptions |> Option.iter (fun v -> param.set_DNSOptions (Collections.Generic.List(v)) )
+ DNSSearch |> Option.iter (fun v -> param.set_DNSSearch (Collections.Generic.List(v)) )
+ ExtraHosts |> Option.iter (fun v -> param.set_ExtraHosts (Collections.Generic.List(v)) )
+ GroupAdd |> Option.iter (fun v -> param.set_GroupAdd (Collections.Generic.List(v)) )
+ IpcMode |> Option.iter param.set_IpcMode
+ Cgroup |> Option.iter param.set_Cgroup
+ Links |> Option.iter (fun v -> param.set_Links (Collections.Generic.List(v)) )
+ OomScoreAdj |> Option.iter param.set_OomScoreAdj
+ PidMode |> Option.iter param.set_PidMode
+ Privileged |> Option.iter param.set_Privileged
+ PublishAllPorts |> Option.iter param.set_PublishAllPorts
+ ReadonlyRootfs |> Option.iter param.set_ReadonlyRootfs
+ SecurityOpt |> Option.iter (fun v -> param.set_SecurityOpt (Collections.Generic.List(v)) )
+ StorageOpt |> Option.iter param.set_StorageOpt
+ Tmpfs |> Option.iter param.set_Tmpfs
+ UTSMode |> Option.iter param.set_UTSMode
+ UsernsMode |> Option.iter param.set_UsernsMode
+ ShmSize |> Option.iter param.set_ShmSize
+ Sysctls |> Option.iter param.set_Sysctls
+ Runtime |> Option.iter param.set_Runtime
+ ConsoleSize |> Option.iter param.set_ConsoleSize
+ Isolation |> Option.iter param.set_Isolation
+ CPUShares |> Option.iter param.set_CPUShares
+ Memory |> Option.iter param.set_Memory
+ NanoCPUs |> Option.iter param.set_NanoCPUs
+ CgroupParent |> Option.iter param.set_CgroupParent
+ BlkioWeight |> Option.iter param.set_BlkioWeight
+ BlkioWeightDevice |> Option.iter (fun v -> param.set_BlkioWeightDevice (Collections.Generic.List(v)) )
+ BlkioDeviceReadBps |> Option.iter (fun v -> param.set_BlkioDeviceReadBps (Collections.Generic.List(v)) )
+ BlkioDeviceWriteBps |> Option.iter (fun v -> param.set_BlkioDeviceWriteBps (Collections.Generic.List(v)) )
+ BlkioDeviceReadIOps |> Option.iter (fun v -> param.set_BlkioDeviceReadIOps (Collections.Generic.List(v)) )
+ BlkioDeviceWriteIOps |> Option.iter (fun v -> param.set_BlkioDeviceWriteIOps (Collections.Generic.List(v)) )
+ CPUPeriod |> Option.iter param.set_CPUPeriod
+ CPUQuota |> Option.iter param.set_CPUQuota
+ CPURealtimePeriod |> Option.iter param.set_CPURealtimePeriod
+ CPURealtimeRuntime |> Option.iter param.set_CPURealtimeRuntime
+ CpusetCpus |> Option.iter param.set_CpusetCpus
+ CpusetMems |> Option.iter param.set_CpusetMems
+ Devices |> Option.iter (fun v -> param.set_Devices (Collections.Generic.List(v)) )
+ DiskQuota |> Option.iter param.set_DiskQuota
+ KernelMemory |> Option.iter param.set_KernelMemory
+ MemoryReservation |> Option.iter param.set_MemoryReservation
+ MemorySwap |> Option.iter param.set_MemorySwap
+ MemorySwappiness |> Option.iter (fun v -> param.set_MemorySwappiness (Nullable(v) ) )
+ OomKillDisable |> Option.iter (fun v -> param.set_OomKillDisable (Nullable(v) ) )
+ PidsLimit |> Option.iter param.set_PidsLimit
+ Ulimits |> Option.iter (fun v -> param.set_Ulimits (Collections.Generic.List(v)) )
+ CPUCount |> Option.iter param.set_CPUCount
+ CPUPercent |> Option.iter param.set_CPUPercent
+ IOMaximumIOps |> Option.iter param.set_IOMaximumIOps
+ IOMaximumBandwidth |> Option.iter param.set_IOMaximumBandwidth
+ Mounts |> Option.iter (fun v -> param.set_Mounts (Collections.Generic.List(v)) )
+ Init |> Option.iter (fun v -> param.set_Init (Nullable(v) ) )
+ InitPath |> Option.iter param.set_InitPath
+
+ param
+
+ /// Creates CreateContainerParameters
+ static member InitCreateContainerParameters
+ (
+ ?Hostname,
+ ?HostConfig,
+ ?Domainname,
+ ?User,
+ ?AttachStdin,
+ ?AttachStdout,
+ ?AttachStderr,
+ ?ExposedPorts,
+ ?Tty,
+ ?OpenStdin,
+ ?StdinOnce,
+ ?Env : seq,
+ ?Cmd : seq,
+ ?Healthcheck,
+ ?ArgsEscaped,
+ ?Image,
+ ?Volumes,
+ ?WorkingDir,
+ ?Entrypoint : seq,
+ ?NetworkDisabled,
+ ?MacAddress,
+ ?OnBuild : seq,
+ ?Labels,
+ ?StopSignal,
+ ?StopTimeout,
+ ?Shell : seq
+ ) =
+
+ let param = new CreateContainerParameters()
+
+ Hostname |> Option.iter param.set_Hostname
+ HostConfig |> Option.iter param.set_HostConfig
+ Domainname |> Option.iter param.set_Domainname
+ User |> Option.iter param.set_User
+ AttachStdin |> Option.iter param.set_AttachStdin
+ AttachStdout |> Option.iter param.set_AttachStdout
+ AttachStderr |> Option.iter param.set_AttachStderr
+ ExposedPorts |> Option.iter param.set_ExposedPorts
+ Tty |> Option.iter param.set_Tty
+ OpenStdin |> Option.iter param.set_OpenStdin
+ StdinOnce |> Option.iter param.set_StdinOnce
+ Env |> Option.iter (fun v -> param.set_Env (Collections.Generic.List(v)) )
+ Cmd |> Option.iter (fun v -> param.set_Cmd (Collections.Generic.List(v)) )
+ Healthcheck |> Option.iter param.set_Healthcheck
+ ArgsEscaped |> Option.iter param.set_ArgsEscaped
+ Image |> Option.iter param.set_Image
+ Volumes |> Option.iter param.set_Volumes
+ WorkingDir |> Option.iter param.set_WorkingDir
+ Entrypoint |> Option.iter (fun v -> param.set_Entrypoint (Collections.Generic.List(v)) )
+ NetworkDisabled |> Option.iter param.set_NetworkDisabled
+ MacAddress |> Option.iter param.set_MacAddress
+ OnBuild |> Option.iter (fun v -> param.set_OnBuild (Collections.Generic.List(v)) )
+ Labels |> Option.iter param.set_Labels
+ StopSignal |> Option.iter param.set_StopSignal
+ StopTimeout |> Option.iter (fun v -> param.set_StopTimeout (Nullable(v) ) )
+ Shell |> Option.iter (fun v -> param.set_Shell (Collections.Generic.List(v)) )
+
+ param
+
+
+ /// Creates ContainerStartParameters
+ static member InitContainerStartParameters
+ (
+ ?DetachKeys
+ ) =
+
+ let param = new ContainerStartParameters()
+ DetachKeys |> Option.iter param.set_DetachKeys
+
+ param
+
+
+ /// Creates ContainerLogsParameters
+ static member InitContainerLogsParameters
+ (
+ ?Follow,
+ ?ShowStderr,
+ ?ShowStdout,
+ ?Since,
+ ?Tail,
+ ?Timestamps
+ ) =
+
+ let param = new ContainerLogsParameters()
+ Follow |> Option.iter (fun v -> param.set_Follow (Nullable(v) ) )
+ ShowStderr |> Option.iter (fun v -> param.set_ShowStderr (Nullable(v) ) )
+ ShowStdout |> Option.iter (fun v -> param.set_ShowStdout (Nullable(v) ) )
+ Since |> Option.iter param.set_Since
+ Tail |> Option.iter param.set_Tail
+ Timestamps |> Option.iter (fun v -> param.set_Timestamps (Nullable(v) ) )
+
+ param
+
+
+ /// Creates ContainerRemoveParameters
+ static member InitContainerRemoveParameters
+ (
+ ?Force,
+ ?RemoveLinks,
+ ?RemoveVolumes
+ ) =
+
+ let param = new ContainerRemoveParameters ()
+ Force |> Option.iter (fun v -> param.set_Force (Nullable(v) ) )
+ RemoveLinks |> Option.iter (fun v -> param.set_RemoveLinks (Nullable(v) ) )
+ RemoveVolumes |> Option.iter (fun v -> param.set_RemoveVolumes (Nullable(v) ) )
+
+ param
+
+
+ /// Creates ContainerPathStatParameters
+ static member InitContainerPathStatParameters
+ (
+ ?AllowOverwriteDirWithFile,
+ // Target path (example: /tmp/)
+ ?Path
+ ) =
+
+ let param = new ContainerPathStatParameters()
+ AllowOverwriteDirWithFile |> Option.iter (fun v -> param.set_AllowOverwriteDirWithFile (Nullable(v) ) )
+ Path |> Option.iter param.set_Path
+
+ param
+
+
+ /// Creates ContainerPathStatParameters
+ static member InitGetArchiveFromContainerParameters
+ (
+ ?Path
+ ) =
+
+ let param = new GetArchiveFromContainerParameters()
+ Path |> Option.iter param.set_Path
+
+ param
+
+
+ /// Creates docker container with CreateContainerParameters for config (async)
+ let createContainerWithAsync (connection:DockerClient) (param:CreateContainerParameters) =
+ async {
+ let! tmp =
+ connection.Containers.CreateContainerAsync(param)
+ |> Async.AwaitTask
+
+ return tmp
+ }
+
+
+ /// Creates docker container of an image given by image name (async)
+ let createContainerByImageAsync (connection:DockerClient) imageName =
+ async {
+ let param = ContainerParams.InitCreateContainerParameters(Image=imageName)
+ //param.Cmd <- Collections.Generic.List(["/bin/bash";"c";"echo 'Hello'"])
+ let! tmp =
+ connection.Containers.CreateContainerAsync(param)
+ |> Async.AwaitTask
+
+ return tmp
+ }
+
+
+ /// Creates docker container of an image given by image name (async)
+ let createContainerByImage (connection:DockerClient) imageName =
+ createContainerByImageAsync connection imageName
+ |> Async.RunSynchronously
+
+
+ /// Creates docker container with ContainerExecStartParameters for config (async)
+ let createContainerWithExecConfigAsync (connection:DockerClient) (param:ContainerExecCreateParameters) id =
+ async {
+ let! tmp =
+ connection.Containers.ExecCreateContainerAsync(id,param,CancellationToken.None)
+ |> Async.AwaitTask
+
+ return tmp
+ }
+
+
+ /// Start container with CreateContainerParameters for config (async)
+ let startContainerWithAsync (connection:DockerClient) (param:ContainerStartParameters) id =
+ async {
+ let! isRunnig =
+ connection.Containers.StartContainerAsync(id,param,CancellationToken.None)
+ |> Async.AwaitTask
+
+ return isRunnig
+ }
+
+
+ /// Start container by a given ID (async)
+ let startContainerAsync (connection:DockerClient) id =
+ startContainerWithAsync connection (ContainerParams.InitContainerStartParameters()) id
+
+
+ /// Start container by a given ID
+ let startContainer (connection:DockerClient) id =
+ startContainerWithAsync connection (ContainerParams.InitContainerStartParameters()) id
+ |> Async.RunSynchronously
+
+
+ /// Start container with ContainerExecStartParameters for config (async)
+ let startContainerWithExecConfigAsync (connection:DockerClient) (param:ContainerExecStartParameters) id =
+ async {
+ let! isRunnig =
+ connection.Containers.StartWithConfigContainerExecAsync(id,param,CancellationToken.None)
+ |> Async.AwaitTask
+
+ return isRunnig
+ }
+
+
+
+ /// Wait for container (async)
+ let waitContainerAsync (connection:DockerClient) id =
+ async {
+ let! tmp =
+ connection.Containers.WaitContainerAsync(id,CancellationToken.None)
+ |> Async.AwaitTask
+
+ return tmp
+ }
+
+ /// Wait for container
+ let waitContainer (connection:DockerClient) id =
+ waitContainerAsync connection id
+ |> Async.RunSynchronously
+
+
+ /// Get logs from container (async)
+ let getContainerLogsAsync (connection:DockerClient) param id =
+ async {
+ let! tmp =
+ connection.Containers.GetContainerLogsAsync(id,param,CancellationToken.None)
+ |> Async.AwaitTask
+
+ return tmp
+ }
+
+
+ /// Get logs from container and show StdOut
+ let getContainerLogs (connection:DockerClient) id =
+ getContainerLogsAsync connection (ContainerParams.InitContainerLogsParameters(ShowStdout=true)) id
+ |> Async.RunSynchronously
+
+
+ /// Lists available containers with ContainersListParameters for filtering (async)
+ let listContainersWithAsync (connection:DockerClient) (param:ContainersListParameters) =
+ // ImagesListParameters are only for filtering
+ async {
+ let! tmp =
+ connection.Containers.ListContainersAsync(param,CancellationToken.None)
+ |> Async.AwaitTask
+ return (tmp |> Seq.map id)
+ }
+
+
+ /// Lists all available containers (async)
+ let listAllContainersAsync (connection:DockerClient) =
+ listContainersWithAsync connection (ContainerParams.InitContainerListParameters(All=true))
+
+
+ /// Returns true if acontainer matches the dockerid (async)
+ let existsByAsync (connection:DockerClient) (dockerid:DockerId) =
+ let filter = Filters.InitContainerFilters(Ancestor=dockerid)
+ let param = ContainerParams.InitContainerListParameters(All=true,Filters=filter)
+ async {
+ let! tmp = listContainersWithAsync connection param
+
+ return (Seq.length tmp > 0)
+ }
+
+
+ /// Returns true if acontainer matches the dockerid
+ let existsByImage (connection:DockerClient) (dockerid:DockerId) =
+ existsByAsync connection dockerid
+ |> Async.RunSynchronously
+
+
+ /// Removes container
+ let removeContainerWithAsync (connection:DockerClient) (param) (dockerid:DockerId) =
+ async {
+ do!
+ connection.Containers.RemoveContainerAsync(dockerid.ToString(),param,CancellationToken.None)
+ |> Async.AwaitTask
+ }
+
+
+ /// Removes container
+ let removeContainerAsync (connection:DockerClient) (dockerid:DockerId) =
+ let param = ContainerParams.InitContainerRemoveParameters()
+ removeContainerWithAsync connection param dockerid
+
+
+ /// Creates an container that will perform the execution (async). IMPORTANT: Start exec-container using StartContainerExecAsync
+ let execCreateContainerAsync (connection:DockerClient) param id =
+ async {
+ let! tmp =
+ connection.Containers.ExecCreateContainerAsync (id,param,CancellationToken.None)
+ |> Async.AwaitTask
+
+ return tmp
+ }
+
+ /// Starts
+ let startContainerExecAsync (connection:DockerClient) id =
+ async {
+ do!
+ connection.Containers.StartContainerExecAsync (id,CancellationToken.None)
+ |> Async.AwaitTask
+ }
+
+ /// Stops
+ let stopContainerAsync (connection:DockerClient) (param) id =
+ async {
+ let! isRunning =
+ connection.Containers.StopContainerAsync(id,param,CancellationToken.None)
+ |> Async.AwaitTask
+ return isRunning
+ }
+
+ /// Writes stream into docker container
+ let extractArchiveToContainerAsync (connection:DockerClient) (param) id stream =
+ async {
+ let! isRunning =
+ connection.Containers.ExtractArchiveToContainerAsync(id,param,stream,CancellationToken.None)
+ |> Async.AwaitTask
+ return isRunning
+ }
+
+
+ ///
+ let getArchiveFromContainerAsync (connection:DockerClient) (param) statOnly id =
+ async {
+ let! isRunning =
+ connection.Containers.GetArchiveFromContainerAsync(id,param,statOnly,CancellationToken.None)
+ |> Async.AwaitTask
+ return isRunning
+ }
+
+
+ //#####################################################
+ //#####################################################
+ //#####################################################
+
+ module Volumes =
+ /// Provides a set of static methods for creating Image parameter.
+ type VolumesParams =
+
+ /// Creates ContainerStartParameters
+ static member InitVolumesCreateParameters
+ (
+ ?Driver,
+ ?DriverOpts,
+ ?Labels,
+ ?Name
+
+ ) =
+
+ let param = new VolumesCreateParameters()
+ Driver |> Option.iter param.set_Driver
+ DriverOpts |> Option.iter param.set_DriverOpts
+ Labels |> Option.iter param.set_Labels
+ Name |> Option.iter param.set_Name
+
+ param
+
+ /// Create volume with VolumesCreateParameters (async)
+ let createAsync (connection:DockerClient) param =
+ async {
+ let! tmp =
+ connection.Volumes. CreateAsync(param,CancellationToken.None)
+ |> Async.AwaitTask
+
+ return tmp
+ }
+
+
+ /// Create volume with VolumesCreateParameters
+ let create (connection:DockerClient) param =
+ createAsync connection param
+ |> Async.RunSynchronously
+
+
+ /// Inspect volume (async)
+ let inspectAsync (connection:DockerClient) name =
+ async {
+ let! tmp =
+ connection.Volumes.InspectAsync(name,CancellationToken.None)
+ |> Async.AwaitTask
+
+ return tmp
+ }
+
+
+ /// Inspect volume
+ let inspect (connection:DockerClient) name =
+ inspectAsync connection name
+ |> Async.RunSynchronously
+
+
+ /// List volume (async)
+ let listVolumesAsync (connection:DockerClient) =
+ async {
+ let! tmp =
+ connection.Volumes.ListAsync(CancellationToken.None)
+ |> Async.AwaitTask
+
+ return tmp
+ }
+
+
+ /// List volume
+ let listVolumes (connection:DockerClient) =
+ listVolumesAsync connection
+ |> Async.RunSynchronously
+
+
+
+
+
diff --git a/src/BioFSharp.BioTools/FastP.fs b/src/BioFSharp.BioTools/FastP.fs
new file mode 100644
index 00000000..669ce8c2
--- /dev/null
+++ b/src/BioFSharp.BioTools/FastP.fs
@@ -0,0 +1,496 @@
+namespace BioFSharp.BioTools
+
+module FastP =
+ open FSharpAux
+ open BioContainer
+
+ let ImageFastp = Docker.DockerId.ImageId "fastp"
+
+
+ type IOOptions =
+ ///read1 input file name
+ |Input1 of string
+ ///read1 output file name
+ |Input2 of string
+ ///read2 input file name
+ |Output1 of string
+ ///read2 output file name
+ |Output2 of string
+ ///indicate the input is using phred64 scoring (it'll be converted to phred33, so the output will still be phred33)
+ |Phred64
+ ///compression level for gzip output (1 ~ 9). 1 is fastest, 9 is smallest, default is 4.
+ |OutputCompressionLevel of int
+ ///input from STDIN. If the STDIN is interleaved paired-end FASTQ, please also add --interleaved_in.
+ |STDIN
+ ///output passing-filters reads to STDOUT. This option will result in interleaved FASTQ output for paired-end input. Disabled by default.
+ |STDOUT
+ ///indicate that Input1 is an interleaved FASTQ which contains both read1 and read2. Disabled by default.
+ |InterleavedInput
+ ///specify how many reads/pairs to be processed. Default 0 means process all reads.
+ |ReadsToProcess of int
+ ///don't overwrite existing files. Overwriting is allowed by default.
+ |DontOverwrite
+
+ static member make = function
+ |Input1 i -> sprintf "--in1 %s" i
+ |Input2 i -> sprintf "--in2 %s" i
+ |Output1 i -> sprintf "--out1 %s" i
+ |Output2 i -> sprintf "--out2 %s" i
+ |Phred64 -> "--phred64"
+ |OutputCompressionLevel i -> sprintf "--compression %i" i
+ |STDIN -> "--stdin"
+ |STDOUT -> "--stdout"
+ |InterleavedInput -> "--interleaved_in"
+ |ReadsToProcess i -> sprintf "--reads_to_process %i" i
+ |DontOverwrite -> "--dont_overwrite"
+
+ static member makeCmdWith (m:MountInfo) = function
+ |Input1 i -> [ "--in1"; (MountInfo.containerPathOf m i) ]
+ |Input2 i -> [ "--in2"; (MountInfo.containerPathOf m i) ]
+ |Output1 i -> [ "--out1"; (MountInfo.containerPathOf m i)]
+ |Output2 i -> [ "--out2"; (MountInfo.containerPathOf m i)]
+ |Phred64 -> ["--phred64"]
+ |OutputCompressionLevel i -> [ "--compression"; string i ]
+ |STDIN -> ["--stdin"]
+ |STDOUT -> ["--stdout"]
+ |InterleavedInput -> ["--interleaved_in"]
+ |ReadsToProcess i -> [ "--reads_to_process"; string i]
+ |DontOverwrite -> ["--dont_overwrite"]
+
+
+
+ type AdapterTrimmingOptions =
+ ///adapter trimming is enabled by default. If this option is specified, adapter trimming is disabled.
+ |DisableAdapterTrimming
+ ///the adapter for read1. For SE data, if not specified, the adapter will be auto-detected. For PE data, this is used if R1/R2 are found not overlapped.
+ |AdapterSequence1 of string
+ ///the adapter for read2 (PE data only). This is used if R1/R2 are found not overlapped. If not specified, it will be the same as AdapterSequence1.
+ |AdapterSequence2 of string
+
+ static member make = function
+ |DisableAdapterTrimming -> "--disable_adapter_trimming"
+ |AdapterSequence1 a1 -> sprintf "--adapter_sequence %s" a1
+ |AdapterSequence2 a2 -> sprintf "--adapter_sequence_r2 %s" a2
+
+ static member makeCmdWith (m:MountInfo) = function
+ |DisableAdapterTrimming -> ["--disable_adapter_trimming"]
+ |AdapterSequence1 a1 -> ["--adapter_sequence"; a1]
+ |AdapterSequence2 a2 -> ["--adapter_sequence_r2"; a2]
+
+
+
+ type GlobalTrimmingOptions =
+ ///trimming how many bases in front for read1, default is 0.
+ |TrimFront1 of int
+ ///trimming how many bases in tail for read1, default is 0.
+ |TrimFront2 of int
+ ///trimming how many bases in front for read2. If it's not specified, it will follow read1's settings.
+ |TrimTail1 of int
+ ///trimming how many bases in tail for read2. If it's not specified, it will follow read1's settings.
+ |TrimTail2 of int
+
+ static member make = function
+ |TrimFront1 tf1 -> sprintf "--trim_front1 %i" tf1
+ |TrimFront2 tf2 -> sprintf "--trim_front2 %i" tf2
+ |TrimTail1 tt1 -> sprintf "--trim_tail1 %i" tt1
+ |TrimTail2 tt2 -> sprintf "--trim_tail2 %i" tt2
+
+ static member makeCmdWith (m:MountInfo) = function
+ |TrimFront1 tf1 -> ["--trim_front1"; string tf1]
+ |TrimFront2 tf2 -> ["--trim_front2"; string tf2]
+ |TrimTail1 tt1 -> ["--trim_tail1"; string tt1]
+ |TrimTail2 tt2 -> ["--trim_tail2"; string tt2]
+
+
+
+ type PolyXTrimmingOptions =
+ ///force polyG tail trimming, by default trimming is automatically enabled for Illumina NextSeq/NovaSeq data.
+ |PolyGTrimming
+ ///the minimum length to detect polyG in the read tail. 10 by default.
+ |PolyGMinLength of int
+ ///disable polyG tail trimming, by default trimming is automatically enabled for Illumina NextSeq/NovaSeq data.
+ |DisablePolyGTrimming
+ ///enable polyX trimming in 3' ends.
+ |PolyXTrimming
+ ///the minimum length to detect polyX in the read tail. 10 by default.
+ |PolyXMinLength of int
+
+ static member make = function
+ |PolyGTrimming -> "--trim_poly_g "
+ |PolyGMinLength pgl -> sprintf "--poly_g_min_len %i" pgl
+ |DisablePolyGTrimming -> "--disable_trim_poly_g"
+ |PolyXTrimming -> "--trim_poly_x"
+ |PolyXMinLength pxl -> sprintf "--poly_x_min_len %i" pxl
+
+ static member makeCmdWith (m:MountInfo) = function
+ |PolyGTrimming -> ["--trim_poly_g"]
+ |PolyGMinLength pgl -> [ "--poly_g_min_len"; string pgl]
+ |DisablePolyGTrimming -> ["--disable_trim_poly_g"]
+ |PolyXTrimming -> ["--trim_poly_x"]
+ |PolyXMinLength pxl -> [ "--poly_x_min_len"; string pxl]
+
+
+
+ type QualityCuttingOptions =
+ ///enable per read cutting by quality in front (5'), default is disabled (WARNING: this will interfere deduplication for both PE/SE data).
+ |FivePrimeQualityCut
+ ///enable per read cutting by quality in tail (3'), default is disabled (WARNING: this will interfere deduplication for SE data).
+ |ThreePrimeQualityCut
+ ///the size of the sliding window for sliding window trimming, default is 4.
+ |CutWindowSize of int
+ ///the bases in the sliding window with mean quality below cutting_quality will be cut, default is Q20.
+ |CutMeanQuality of int
+
+ static member make = function
+ |FivePrimeQualityCut -> "--cut_by_quality5"
+ |ThreePrimeQualityCut -> "--cut_by_quality3"
+ |CutWindowSize cws -> sprintf "--cut_window_size %i" cws
+ |CutMeanQuality cmq -> sprintf "--cut_mean_quality %i" cmq
+
+ static member makeCmdWith (m:MountInfo) = function
+ |FivePrimeQualityCut -> ["--cut_by_quality5"]
+ |ThreePrimeQualityCut -> ["--cut_by_quality3"]
+ |CutWindowSize cws -> ["--cut_window_size"; string cws]
+ |CutMeanQuality cmq -> ["--cut_mean_quality"; string cmq]
+
+
+ type QualityFilteringOptions =
+ ///quality filtering is enabled by default. If this option is specified, quality filtering is disabled.
+ |DisableQualityFiltering
+ ///the quality value that a base is qualified. Default 15 means phred quality >=Q15 is qualified.
+ |BaseQualityThreshold of int
+ ///how many percents of bases are allowed to be unqualified (0~100). Default 40 means 40%.
+ |UnqualifiedThreshold of int
+ ///if one read's number of N base is >n_base_limit, then this read/pair is discarded. Default is 5.
+ |NBaseLimit of int
+
+ static member make = function
+ |DisableQualityFiltering -> "--disable_quality_filtering"
+ |BaseQualityThreshold bqt -> sprintf "--qualified_quality_phred %i" bqt
+ |UnqualifiedThreshold ut -> sprintf "--unqualified_percent_limit %i" ut
+ |NBaseLimit nbl -> sprintf "--n_base_limit %i" nbl
+
+ static member makeCmdWith (m:MountInfo) = function
+ |DisableQualityFiltering -> ["--disable_quality_filtering"]
+ |BaseQualityThreshold bqt -> [ "--qualified_quality_phred"; string bqt]
+ |UnqualifiedThreshold ut -> [ "--unqualified_percent_limit"; string ut]
+ |NBaseLimit nbl -> [ "--n_base_limit"; string nbl]
+
+
+ type LengthFilteringOptions =
+ ///length filtering is enabled by default. If this option is specified, length filtering is disabled.
+ |DisableLengthFiltering
+ ///reads shorter than length_required will be discarded, default is 15.
+ |RequiredLength of int
+ ///reads longer than length_limit will be discarded, default 0 means no limitation.
+ |LengthLimit of int
+
+ static member make = function
+ |DisableLengthFiltering -> "--disable_length_filtering"
+ |RequiredLength rl -> sprintf "--length_required %i" rl
+ |LengthLimit lm -> sprintf "--length_limit %i" lm
+
+ static member makeCmdWith (m:MountInfo) = function
+ |DisableLengthFiltering -> ["--disable_length_filtering"]
+ |RequiredLength rl -> [ "--length_required"; string rl]
+ |LengthLimit lm -> [ "--length_limit"; string lm]
+
+
+ type LowComplexityFilteringOptions =
+ ///enable low complexity filter. The complexity is defined as the percentage of base that is different from its next base (base[i] != base[i+1]).
+ |EnableLowComplexityFiltering
+ ///the threshold for low complexity filter (0~100). Default is 30, which means 30% complexity is required.
+ |ComplexityThreshold of int
+
+ static member make = function
+ |EnableLowComplexityFiltering -> "--low_complexity_filter"
+ |ComplexityThreshold ct -> sprintf "--complexity_threshold %i" ct
+
+ static member makeCmdWith (m:MountInfo) = function
+ |EnableLowComplexityFiltering -> ["--low_complexity_filter"]
+ |ComplexityThreshold ct -> ["--complexity_threshold"; string ct]
+
+
+ type IndexFilteringOptions =
+ ///specify a file contains a list of barcodes of index1 to be filtered out, one barcode per line.
+ |FilterByIndex1 of string
+ ///specify a file contains a list of barcodes of index2 to be filtered out, one barcode per line.
+ |FilterByIndex2 of string
+ ///the allowed difference of index barcode for index filtering, default 0 means completely identical.
+ |IndexFilterThreshold of int
+
+ static member make = function
+ |FilterByIndex1 fbi1 -> sprintf "--filter_by_index1 %s" fbi1
+ |FilterByIndex2 fbi2 -> sprintf "--filter_by_index2 %s" fbi2
+ |IndexFilterThreshold ift -> sprintf "--filter_by_index_threshold %i" ift
+
+ static member makeCmdWith (m:MountInfo) = function
+ |FilterByIndex1 fbi1 -> ["--filter_by_index1"; fbi1]
+ |FilterByIndex2 fbi2 -> ["--filter_by_index2"; fbi2]
+ |IndexFilterThreshold ift -> ["--filter_by_index_threshold %i"; string ift]
+
+
+ type BaseCorrectionOptions =
+ ///enable base correction in overlapped regions (only for PE data), default is disabled.
+ |EnableBaseCorrection
+ ///the minimum length of the overlapped region for overlap analysis based adapter trimming and correction. 30 by default.
+ |RequiredLengthOverlap of int
+ ///the maximum difference of the overlapped region for overlap analysis based adapter trimming and correction. 5 by default.
+ |MaxOverlapDifference of int
+
+ static member make = function
+ |EnableBaseCorrection -> "--correction"
+ |RequiredLengthOverlap rlo -> sprintf "--overlap_len_require %i" rlo
+ |MaxOverlapDifference modiff -> sprintf "--overlap_diff_limit %i" modiff
+
+ static member makeCmdWith (m:MountInfo) = function
+ |EnableBaseCorrection -> ["--correction"]
+ |RequiredLengthOverlap rlo -> [ "--overlap_len_require"; string rlo]
+ |MaxOverlapDifference modiff ->[ "--overlap_diff_limit"; string modiff]
+
+
+ type UMILocation =
+ |Index1
+ |Index2
+ |Read1
+ |Read2
+ |PerIndex
+ |PerRead
+ |NoLocation
+
+ static member make = function
+ |Index1 -> "index1"
+ |Index2 -> "index2"
+ |Read1 -> "read1"
+ |Read2 -> "read2"
+ |PerIndex -> "per_index"
+ |PerRead -> "per_read"
+ |NoLocation -> ""
+
+
+ static member makeCmdWith (m:MountInfo) = function
+ |Index1 -> ["index1"]
+ |Index2 -> ["index2"]
+ |Read1 -> ["read1"]
+ |Read2 -> ["read2"]
+ |PerIndex -> ["per_index"]
+ |PerRead -> ["per_read"]
+ |NoLocation -> [""]
+
+ //========================================= Helper ==========================================
+ let stringOfUMILocation (loc:UMILocation) =
+ match loc with
+ |Index1 -> "index1"
+ |Index2 -> "index2"
+ |Read1 -> "read1"
+ |Read2 -> "read2"
+ |PerIndex -> "per_index"
+ |PerRead -> "per_read"
+ |NoLocation -> ""
+ //=============================================================================================
+
+ type UMIProcessingOptions =
+ ///enable unique molecular identifer (UMI) preprocessing.
+ |EnableUMI
+ ///specify the location of UMI, can be index1/index2/read1/read2/per_index/per_read, default is none.
+ |Location of UMILocation
+ ///if the UMI is in read1/read2, its length should be provided.
+ |UMILength of int
+ ///if specified, an underline will be used to connect prefix and UMI (i.e. prefix=UMI, UMI=AATTCG, final=UMI_AATTCG). No prefix by default.
+ |UMIPrefix of string
+ ///if the UMI is in read1/read2, fastp can skip several bases following UMI, default is 0.
+ |UMISkip of int
+
+ static member make = function
+ |EnableUMI -> sprintf "--umi"
+ |Location loc -> sprintf "--umi_loc %s" (stringOfUMILocation loc) //(stringOfUMILocation loc)
+ |UMILength len -> sprintf "--umi_len %i" len
+ |UMIPrefix prefix -> sprintf "--umi_prefix %s" prefix
+ |UMISkip s -> sprintf "--umi_skip %i" s
+
+ static member makeCmdWith (m:MountInfo) = function
+ |EnableUMI -> ["--umi"]
+ |Location loc -> ["--umi_loc"; (stringOfUMILocation loc)] //(stringOfUMILocation loc)
+ |UMILength len -> ["--umi_len"; string len]
+ |UMIPrefix prefix -> ["--umi_prefix"; prefix]
+ |UMISkip s -> ["--umi_skip"; string s]
+
+
+ type OverrepresentationAnalysisOptions =
+ ///enable overrepresented sequence analysis.
+ |EnableORA
+ ///One in ORASampling reads will be computed for overrepresentation analysis (1~10000), smaller is slower, default is 20.
+ |ORASampling of int
+
+ static member make = function
+ |EnableORA -> "--overrepresentation_analysis"
+ |ORASampling smpl -> sprintf "--overrepresentation_sampling %i" smpl
+
+ static member makeCmdWith (m:MountInfo) = function
+ |EnableORA -> ["--overrepresentation_analysis"]
+ |ORASampling smpl -> [ "--overrepresentation_sampling"; string smpl]
+
+
+ type ReportingOptions =
+ ///the json format report file name. default is "fastp.json".
+ |JsonReport of string
+ ///the html format report file name. default is "fastp.html".
+ |HtmlReport of string
+ ///should be quoted with ' or ", default is "fastp report".
+ |ReportTitle of string
+
+ static member make = function
+ |JsonReport jsonrep -> sprintf "--json %s" jsonrep
+ |HtmlReport htmlrep -> sprintf "--html %s" htmlrep
+ |ReportTitle title -> sprintf "--report_title %s" title
+
+ static member makeCmdWith (m:MountInfo) = function
+ |JsonReport jsonrep -> ["--json"; (MountInfo.containerPathOf m jsonrep)]
+ |HtmlReport htmlrep -> ["--html"; (MountInfo.containerPathOf m htmlrep)]
+ |ReportTitle title -> ["--report_title"; (MountInfo.containerPathOf m title)]
+
+
+ type OutputSplittingOptions =
+ ///split output by limiting total split file number with this option (2~999), a sequential number prefix will be added to output name ( 0001.out.fq, 0002.out.fq...), disabled by default.
+ |Split of int
+ ///split output by limiting lines of each file with this option(>=1000), a sequential number prefix will be added to output name ( 0001.out.fq, 0002.out.fq...), disabled by default.
+ |SplitByLines of int64
+ ///the digits for the sequential number padding (1~10), default is 4, so the filename will be padded as 0001.xxx, 0 to disable padding.
+ |SplitPrefixDigits of int
+
+ static member make = function
+ |Split s -> sprintf "--split %i" s
+ |SplitByLines sbl -> sprintf "--split_by_lines %i" sbl
+ |SplitPrefixDigits spd -> sprintf "--split_prefix_digits %i" spd
+
+ static member makeCmdWith (m:MountInfo) = function
+ |Split s -> ["--split"; string s]
+ |SplitByLines sbl -> ["--split_by_lines"; string sbl]
+ |SplitPrefixDigits spd -> ["--split_prefix_digits"; string spd]
+
+
+
+ type FastpParams =
+ ///worker thread number, default is 2
+ |NumThreads of int
+ |IO of IOOptions list
+ |AdapterTrimming of AdapterTrimmingOptions list
+ |GlobalTrimming of GlobalTrimmingOptions list
+ |PolyXTrimming of PolyXTrimmingOptions list
+ |QualityCutting of QualityCuttingOptions list
+ |QualityFiltering of QualityFilteringOptions list
+ |LengthFiltering of LengthFilteringOptions list
+ |LowComplexityFiltering of LowComplexityFilteringOptions list
+ |IndexFiltering of IndexFilteringOptions list
+ |BaseCorrection of BaseCorrectionOptions list
+ |UMIProcessing of UMIProcessingOptions list
+ |OverrepresentationAnalysis of OverrepresentationAnalysisOptions list
+ |Reporting of ReportingOptions list
+ |OutputSplitting of OutputSplittingOptions list
+
+ static member make = function
+ |NumThreads t -> [sprintf "--thread %i" t ]
+ |IO l ->
+ let tmp = l |> Seq.map (fun p -> IOOptions.make p) |> Seq.toList
+ tmp
+ |AdapterTrimming l ->
+ let tmp = l |> Seq.map (fun p -> AdapterTrimmingOptions.make p) |> Seq.toList
+ tmp
+ | GlobalTrimming l ->
+ let tmp = l |> Seq.map (fun p -> GlobalTrimmingOptions.make p) |> Seq.toList
+ tmp
+ |PolyXTrimming l ->
+ let tmp = l |> Seq.map (fun p -> PolyXTrimmingOptions.make p) |> Seq.toList
+ tmp
+ |QualityCutting l ->
+ let tmp = l |> Seq.map (fun p -> QualityCuttingOptions.make p) |> Seq.toList
+ tmp
+ |QualityFiltering l ->
+ let tmp = l |> Seq.map (fun p -> QualityFilteringOptions.make p) |> Seq.toList
+ tmp
+ |LengthFiltering l ->
+ let tmp = l |> Seq.map (fun p -> LengthFilteringOptions.make p) |> Seq.toList
+ tmp
+ |LowComplexityFiltering l ->
+ let tmp = l |> Seq.map (fun p -> LowComplexityFilteringOptions.make p) |> Seq.toList
+ tmp
+ |IndexFiltering l ->
+ let tmp = l |> Seq.map (fun p -> IndexFilteringOptions.make p) |> Seq.toList
+ tmp
+ |BaseCorrection l ->
+ let tmp = l |> Seq.map (fun p -> BaseCorrectionOptions.make p) |> Seq.toList
+ tmp
+ |UMIProcessing l ->
+ let tmp = l |> Seq.map (fun p -> UMIProcessingOptions.make p) |> Seq.toList
+ tmp
+ |OverrepresentationAnalysis l ->
+ let tmp = l |> Seq.map (fun p -> OverrepresentationAnalysisOptions.make p) |> Seq.toList
+ tmp
+ |Reporting l ->
+ let tmp = l |> Seq.map (fun p -> ReportingOptions.make p) |> Seq.toList
+ tmp
+ |OutputSplitting l ->
+ let tmp = l |> Seq.map (fun p -> OutputSplittingOptions.make p) |> Seq.toList
+ tmp
+
+
+ static member makeCmd (m:MountInfo) = function
+ |NumThreads t -> [sprintf "--thread %i" t ]
+ |IO l ->
+ let tmp = l |> Seq.map (fun p -> IOOptions.makeCmdWith m p) |> List.concat
+ tmp
+ |AdapterTrimming l ->
+ let tmp = l |> Seq.map (fun p -> AdapterTrimmingOptions.makeCmdWith m p) |> List.concat
+ tmp
+ | GlobalTrimming l ->
+ let tmp = l |> Seq.map (fun p -> GlobalTrimmingOptions.makeCmdWith m p) |> List.concat
+ tmp
+ |PolyXTrimming l ->
+ let tmp = l |> Seq.map (fun p -> PolyXTrimmingOptions.makeCmdWith m p) |> List.concat
+ tmp
+ |QualityCutting l ->
+ let tmp = l |> Seq.map (fun p -> QualityCuttingOptions.makeCmdWith m p) |> List.concat
+ tmp
+ |QualityFiltering l ->
+ let tmp = l |> Seq.map (fun p -> QualityFilteringOptions.makeCmdWith m p) |> List.concat
+ tmp
+ |LengthFiltering l ->
+ let tmp = l |> Seq.map (fun p -> LengthFilteringOptions.makeCmdWith m p) |> List.concat
+ tmp
+ |LowComplexityFiltering l ->
+ let tmp = l |> Seq.map (fun p -> LowComplexityFilteringOptions.makeCmdWith m p) |> List.concat
+ tmp
+ |IndexFiltering l ->
+ let tmp = l |> Seq.map (fun p -> IndexFilteringOptions.makeCmdWith m p) |> List.concat
+ tmp
+ |BaseCorrection l ->
+ let tmp = l |> Seq.map (fun p -> BaseCorrectionOptions.makeCmdWith m p) |> List.concat
+ tmp
+ |UMIProcessing l ->
+ let tmp = l |> Seq.map (fun p -> UMIProcessingOptions.makeCmdWith m p) |> List.concat
+ tmp
+ |OverrepresentationAnalysis l ->
+ let tmp = l |> Seq.map (fun p -> OverrepresentationAnalysisOptions.makeCmdWith m p) |> List.concat
+ tmp
+ |Reporting l ->
+ let tmp = l |> Seq.map (fun p -> ReportingOptions.makeCmdWith m p) |> List.concat
+ tmp
+ |OutputSplitting l ->
+ let tmp = l |> Seq.map (fun p -> OutputSplittingOptions.makeCmdWith m p) |> List.concat
+ tmp
+
+
+ let runFastpAsync (bcContext:BioContainer.BcContext) (opt:FastpParams list) =
+ let cmds = (opt |> List.map (FastpParams.makeCmd bcContext.Mount))
+ let tp = "fastp"::(cmds |> List.concat)
+
+ async {
+ let! fastP = BioContainer.execAsync bcContext tp
+ return fastP
+ }
+
+
+ let runFastp (bcContext:BioContainer.BcContext) (opt:FastpParams list) =
+ runFastpAsync bcContext opt
+ |> Async.RunSynchronously
+
diff --git a/src/BioFSharp.BioTools/HMMER.fs b/src/BioFSharp.BioTools/HMMER.fs
new file mode 100644
index 00000000..3b6136b2
--- /dev/null
+++ b/src/BioFSharp.BioTools/HMMER.fs
@@ -0,0 +1,1160 @@
+namespace BioFSharp.BioTools
+
+// adapted from the hmmer usage guide at http://eddylab.org/software/hmmer/Userguide.pdf
+// HMMER: biosequence analysis using profile hidden Markov models (http://hmmer.org/)
+// HMMER source code is available on github under The 3-Clause BSD License: https://github.com/EddyRivasLab/hmmer/blob/master/LICENSE
+
+///HMMER - Biological sequence analysis using profile hidden Markov models (http://hmmer.org/)
+///Functions providing an API for the HMMER 3 biocontainer
+module HMMER =
+
+ open BioContainer
+
+ //Common parameter types for multiple CLI tools
+ type AlphabetType =
+ ///input alignment is protein sequence data
+ | AminoAcids
+ ///input alignment is DNA sequence data
+ | DNA
+ ///input alignment is RNA sequence data
+ | RNA
+
+ static member make = function
+ | AminoAcids-> ["--amino"]
+ | DNA -> ["--dna"]
+ | RNA -> ["--rna"]
+
+ ///Options controlling reporting thresholds:
+ // -E : report sequences <= this E-value threshold in output [10.0] (x>0)
+ // -T : report sequences >= this score threshold in output
+ // --domE : report domains <= this E-value threshold in output [10.0] (x>0)
+ // --domT : report domains >= this score cutoff in output
+
+ type ReportingThresholdOptions =
+ ///report sequences <= this E-value threshold in output
+ | MaxSequenceEvalue of float
+ ///report sequences >= this score threshold in output
+ | MinSequenceScore of float
+ ///report domains <= this E-value threshold in output
+ | MaxDomainEvalue of float
+ ///report domains >= this score cutoff in output
+ | MinDomainScore of float
+
+ static member make =
+ function
+ | MaxSequenceEvalue t -> ["-E"; string t]
+ | MinSequenceScore t -> ["-T"; string t]
+ | MaxDomainEvalue t -> ["--domE"; string t]
+ | MinDomainScore t -> ["--domT"; string t]
+
+ ///Options controlling inclusion (significance) thresholds:
+ // --incE : consider sequences <= this E-value threshold as significant
+ // --incT : consider sequences >= this score threshold as significant
+ // --incdomE : consider domains <= this E-value threshold as significant
+ // --incdomT : consider domains >= this score threshold as significant
+
+ type InclusionThresholdOptions =
+ ///consider sequences <= this E-value threshold as significant
+ | MaxSequenceEvalue of float
+ ///consider sequences >= this score threshold as significant
+ | MinSequenceScore of float
+ ///consider domains <= this E-value threshold as significant
+ | MaxDomainEvalue of float
+ ///consider domains >= this score threshold as significant
+ | MinDomainScore of float
+
+ static member make =
+ function
+ | MaxSequenceEvalue t -> ["-incE"; string t]
+ | MinSequenceScore t -> ["-incT"; string t]
+ | MaxDomainEvalue t -> ["--incdomE"; string t]
+ | MinDomainScore t -> ["--incdomT"; string t]
+
+ ///Options controlling model-specific thresholding:
+ // --cut_ga : use profile's GA gathering cutoffs to set all thresholding
+ // --cut_nc : use profile's NC noise cutoffs to set all thresholding
+ // --cut_tc : use profile's TC trusted cutoffs to set all thresholding
+
+ type ModelSpecificThresholdOptions =
+ ///use profile's GA gathering cutoffs to set all thresholding
+ | GatheringCutoff
+ ///use profile's NC noise cutoffs to set all thresholding
+ | NoiseCutoff
+ ///use profile's TC trusted cutoffs to set all thresholding
+ | TrustedCutoff
+
+ static member make =
+ function
+ | GatheringCutoff -> ["--cut_ga"]
+ | NoiseCutoff -> ["--cut_nc"]
+ | TrustedCutoff -> ["--cut_tc"]
+
+ ///Options controlling acceleration heuristics:
+ // --max : Turn all heuristic filters off (less speed, more power)
+ // --F1 : Stage 1 (MSV) threshold: promote hits w/ P <= F1 [0.02]
+ // --F2 : Stage 2 (Vit) threshold: promote hits w/ P <= F2 [1e-3]
+ // --F3 : Stage 3 (Fwd) threshold: promote hits w/ P <= F3 [1e-5]
+ // --nobias : turn off composition bias filter
+
+ type AccelerationHeuristicsOptions =
+ ///Turn all heuristic filters off (less speed, more power)
+ | NoFilters
+ ///Stage 1 (MSV) threshold: promote hits w/ P <= F1
+ | MSVThreshold of float
+ ///Stage 2 (Vit) threshold: promote hits w/ P <= F2
+ | VitThreshold of float
+ ///Stage 3 (Fwd) threshold: promote hits w/ P <= F3
+ | FwdThreshold of float
+ ///turn off composition bias filter
+ | NoBiasFilter
+
+ static member make =
+ function
+ | NoFilters -> ["max"]
+ | MSVThreshold t-> ["F1"; string t]
+ | VitThreshold t-> ["F2"; string t]
+ | FwdThreshold t-> ["F3"; string t]
+ | NoBiasFilter -> ["nobias"]
+
+ ///hmmbuild - construct profiles from multiple sequence alignments
+ module HMMbuild =
+ //Usage: hmmbuild [-options]
+
+ ///Options for selecting alphabet rather than guessing it:
+ // --amino : input alignment is protein sequence data
+ // --dna : input alignment is DNA sequence data
+ // --rna : input alignment is RNA sequence data
+
+
+ ///Alternative model construction strategies:
+ // --fast : assign cols w/ >= symfrac residues as consensus [default]
+ // --hand : manual construction (requires reference annotation)
+ // --symfrac : sets sym fraction controlling --fast construction [0.5]
+ // --fragthresh : if L <= x*alen, tag sequence as a fragment [0.5]
+
+ type ModelConstructionOptions =
+ | Fast
+ | Manual
+ | SymFraction of float
+ | FragmentThreshold of float
+
+ static member make = function
+ | Fast -> ["--fast"]
+ | Manual -> ["--hand"]
+ | SymFraction sf -> ["--symfrac"; string sf]
+ | FragmentThreshold ft -> ["--fragthresh"; string ft]
+
+ ///Alternative relative sequence weighting strategies:
+ // --wpb : Henikoff position-based weights [default]
+ // --wgsc : Gerstein/Sonnhammer/Chothia tree weights
+ // --wblosum : Henikoff simple filter weights
+ // --wnone : don't do any relative weighting; set all to 1
+ // --wgiven : use weights as given in MSA file
+ // --wid : for --wblosum: set identity cutoff [0.62] (0<=x<=1)
+
+ type RelativeSequenceWeightingOptions =
+ ///Henikoff position-based weights
+ | PositionBased
+ ///Gerstein/Sonnhammer/Chothia tree weights
+ | GSC
+ ///Henikoff simple filter weights
+ | Blosum
+ ///for Blosum: set identity cutoff (0<=x<=1)
+ | BlosumIdentityCutoff of float
+ ///don't do any relative weighting; set all to 1
+ | NoRelativeWeighting
+ ///use weights as given in MSA file
+ | Given
+
+ static member make = function
+ | PositionBased -> ["--wpb"]
+ | GSC -> ["--wgsc"]
+ | Blosum -> ["--wblosum"]
+ | BlosumIdentityCutoff c-> ["--wid"; string c]
+ | NoRelativeWeighting -> ["--wnone"]
+ | Given -> ["--wgiven"]
+
+ ///Alternative effective sequence weighting strategies:
+ // --eent : adjust eff seq # to achieve relative entropy target [default]
+ // --eclust : eff seq # is # of single linkage clusters
+ // --enone : no effective seq # weighting: just use nseq
+ // --eset : set eff seq # for all models to
+ // --ere : for --eent: set minimum rel entropy/position to
+ // --esigma : for --eent: set sigma param to [45.0]
+ // --eid : for --eclust: set fractional identity cutoff to [0.62]
+
+ type EffectiveSequenceWeightingOptions =
+ ///adjust eff seq # to achieve relative entropy target
+ | EntropyTargeted
+ ///for EntropyTargeted: set minimum rel entropy/position to
+ | MinimalEntropy of float
+ ///for EntropyTargeted: set sigma param to [45.0]
+ | EntropySigma of float
+ ///eff seq # is # of single linkage clusters
+ | SingleLinkageClusters
+ ///for SingleLinkageClusters: set fractional identity cutoff to [0.62]
+ | ClusterIdentityCutoff of float
+ ///no effective seq # weighting: just use nseq
+ | NoEffectiveWeighting
+ ///set eff seq # for all models to
+ | Set of float
+
+ static member make = function
+ | EntropyTargeted -> ["--eent"]
+ | SingleLinkageClusters -> ["--eclust"]
+ | NoEffectiveWeighting -> ["--enone"]
+ | Set s -> ["--eset"; string s]
+ | MinimalEntropy me -> ["--ere"; string me]
+ | EntropySigma es -> ["--esigma"; string es]
+ | ClusterIdentityCutoff c -> ["--eid"; string c]
+
+
+ ///Alternative prior strategies:
+ // --pnone : don't use any prior; parameters are frequencies
+ // --plaplace : use a Laplace +1 prior
+ // --popen : force gap open prob. (w/ --singlemx, aa default 0.02, nt 0.031)
+ // --pextend : force gap extend prob. (w/ --singlemx, aa default 0.4, nt 0.75)
+
+ type PriorOptions =
+ ///don't use any prior; parameters are frequencies
+ | NoPrior
+ ///use a Laplace +1 prio
+ | Laplace
+ ///force gap open prob.
+ | ForcePGapOpen of float
+ ///force gap extend prob
+ | ForcePGapExtend of float
+
+ static member make = function
+ |NoPrior -> ["--pnone"]
+ |Laplace -> ["--plaplace"]
+ |ForcePGapOpen f -> ["--popen"; string f]
+ |ForcePGapExtend f -> ["--pextend"; string f]
+
+ type SubstitutionScoreMatrix =
+ |PAM30
+ |PAM70
+ |PAM120
+ |PAM240
+ |BLOSUM45
+ |BLOSUM50
+ |BLOSUM62
+ |BLOSUM80
+ |BLOSUM90
+ |DNA1
+
+ static member make = function
+ |PAM30 -> "PAM30"
+ |PAM70 -> "PAM70"
+ |PAM120 -> "PAM120"
+ |PAM240 -> "PAM240"
+ |BLOSUM45-> "BLOSUM45"
+ |BLOSUM50-> "BLOSUM50"
+ |BLOSUM62-> "BLOSUM62"
+ |BLOSUM80-> "BLOSUM80"
+ |BLOSUM90-> "BLOSUM90"
+ |DNA1 -> "DNA1"
+
+ ///Handling single sequence inputs:
+ // --singlemx : use substitution score matrix for single-sequence inputs
+ // --mx : substitution score matrix (built-in matrices, with --singlemx)
+ // --mxfile : read substitution score matrix from file (with --singlemx)
+ type SingleSequenceHandlingOptions =
+ ///use substitution score matrix for single-sequence inputs
+ | UseSubstitutionMatrix
+ ///substitution score matrix (built-in matrices, use with UseSubstitutionMatrix)
+ | SMatrix of SubstitutionScoreMatrix
+ ///read substitution score matrix from file (use with UseSubstitutionMatrix)
+ | SMatrixFile of string
+
+ static member make = function
+ | UseSubstitutionMatrix -> ["--singlemx"]
+ | SMatrix s -> ["--mx"; (SubstitutionScoreMatrix.make s)]
+ | SMatrixFile path -> ["--mxfile"; path]
+
+ static member makeWith (m:MountInfo) =
+ let cPath p = (MountInfo.containerPathOf m p)
+ function
+ | UseSubstitutionMatrix -> ["--singlemx"]
+ | SMatrix s -> ["--mx"; (SubstitutionScoreMatrix.make s)]
+ | SMatrixFile path -> ["--mxfile"; cPath path]
+
+ //Control of E-value calibration:
+ // --EmL : length of sequences for MSV Gumbel mu fit [200] (n>0)
+ // --EmN : number of sequences for MSV Gumbel mu fit [200] (n>0)
+ // --EvL : length of sequences for Viterbi Gumbel mu fit [200] (n>0)
+ // --EvN : number of sequences for Viterbi Gumbel mu fit [200] (n>0)
+ // --EfL : length of sequences for Forward exp tail tau fit [100] (n>0)
+ // --EfN : number of sequences for Forward exp tail tau fit [200] (n>0)
+ // --Eft : tail mass for Forward exponential tail tau fit [0.04] (0 ["--EmL";string l]
+ | MSVNumber n -> ["--EmN";string n]
+ | ViterbiLength l -> ["--EvL";string l]
+ | ViterbiNumber n -> ["--EvN";string n]
+ | ForwardLength l -> ["--EfL";string l]
+ | ForwardNumber n -> ["--EfN";string n]
+ | ForwardTailMass m -> ["--Eft";string m]
+
+ ///Other options:
+ // --cpu : number of parallel CPU workers for multithreads
+ // --stall : arrest after start: for attaching debugger to process
+ // --informat : assert input alifile is in format (no autodetect)
+ // --seed : set RNG seed to (if 0: one-time arbitrary seed) [42]
+ // --w_beta : tail mass at which window length is determined
+ // --w_length : window length
+ // --maxinsertlen : pretend all inserts are length <=
+
+ type MiscellaneousOptions =
+ ///number of parallel CPU workers for multithreads
+ | Threads of int
+ ///arrest after start: for attaching debugger to process
+ | Stall
+ ///assert input alifile is in format (no autodetect)
+ | AssertInFormat of string
+ ///set RNG seed to (if 0: one-time arbitrary seed)
+ | Seed of int
+ ///tail mass at which window length is determined
+ | TailMass of float
+ ///window length
+ | WindowLength of int
+ ///pretend all inserts are length <=
+ | MaxInsertionLength of int
+
+ static member make = function
+ | Threads t -> ["--cpu"; string t]
+ | Stall -> ["--stall"]
+ | AssertInFormat f -> ["--informat"; f]
+ | Seed s -> ["--seed"; string s]
+ | TailMass m -> ["--w_beta"; string m]
+ | WindowLength wl -> ["--w_length"; string wl]
+ | MaxInsertionLength l -> ["--maxinsertlen"; string l]
+
+ //Basic options:
+ // -h : show brief help on version and usage
+ // -n : name the HMM
+ // -o : direct summary output to file , not stdout
+ // -O : resave annotated, possibly modified MSA to file
+
+ type HMMbuildParams =
+ ///Multiple sequence alignment input file
+ | InputMSAFile of string
+ ///HMM file to save the generated HMMs to
+ | OutputHMMFile of string
+ ///name the HMM
+ | Name of string
+ ///direct summary output to file , not stdout
+ | SummaryToFile of string
+ ///resave annotated, possibly modified MSA to file
+ | ResaveFile of string
+ ///Specify the alphabet used in the alignment file
+ | Alphabet of AlphabetType list
+ ///Alternative model construction strategies:
+ | ModelConstruction of ModelConstructionOptions list
+ ///Alternative relative sequence weighting strategies
+ | RelativeSequenceWeighting of RelativeSequenceWeightingOptions list
+ ///Alternative effective sequence weighting strategies
+ | EffectiveSequenceWeighting of EffectiveSequenceWeightingOptions list
+ ///Alternative prior strategies
+ | Prior of PriorOptions list
+ ///Handling single sequence inputs
+ | SingleSequenceHandling of SingleSequenceHandlingOptions list
+ ///Control of E-value calibration
+ | EValueControl of EValueControlOptions list
+ ///Other options
+ | Miscellaneous of MiscellaneousOptions list
+
+ static member makeCmd =
+ function
+ | InputMSAFile path -> [path]
+ | OutputHMMFile path -> [path]
+ | Name n -> ["-n";n]
+ | SummaryToFile path -> ["-o";path]
+ | ResaveFile path -> ["-O";path]
+ | Alphabet aList -> aList |> List.map AlphabetType.make |> List.concat
+ | ModelConstruction mcList -> mcList |> List.map ModelConstructionOptions.make |> List.concat
+ | RelativeSequenceWeighting rswList -> rswList |> List.map RelativeSequenceWeightingOptions.make |> List.concat
+ | EffectiveSequenceWeighting eswList -> eswList |> List.map EffectiveSequenceWeightingOptions.make |> List.concat
+ | Prior pList -> pList |> List.map PriorOptions.make |> List.concat
+ | SingleSequenceHandling sshList -> sshList |> List.map SingleSequenceHandlingOptions.make |> List.concat
+ | EValueControl evcList -> evcList |> List.map EValueControlOptions.make |> List.concat
+ | Miscellaneous mList -> mList |> List.map MiscellaneousOptions.make |> List.concat
+
+ static member makeCmdWith (m:MountInfo) =
+ let cPath p = (MountInfo.containerPathOf m p)
+ function
+ | InputMSAFile path -> [cPath path]
+ | OutputHMMFile path -> [cPath path]
+ | Name n -> ["-n";n]
+ | SummaryToFile path -> ["-o";cPath path]
+ | ResaveFile path -> ["-O";cPath path]
+ | Alphabet aList -> aList |> List.map AlphabetType.make |> List.concat
+ | ModelConstruction mcList -> mcList |> List.map ModelConstructionOptions.make |> List.concat
+ | RelativeSequenceWeighting rswList -> rswList |> List.map RelativeSequenceWeightingOptions.make |> List.concat
+ | EffectiveSequenceWeighting eswList -> eswList |> List.map EffectiveSequenceWeightingOptions.make |> List.concat
+ | Prior pList -> pList |> List.map PriorOptions.make |> List.concat
+ | SingleSequenceHandling sshList -> sshList |> List.map (SingleSequenceHandlingOptions.makeWith m) |> List.concat
+ | EValueControl evcList -> evcList |> List.map EValueControlOptions.make |> List.concat
+ | Miscellaneous mList -> mList |> List.map MiscellaneousOptions.make |> List.concat
+
+ let runHMMbuildAsync (bcContext:BioContainer.BcContext) (opt:HMMbuildParams list) =
+ //Usage: hmmbuild [-options] -> filter for in/out and move them to the end
+ let msa =
+ opt
+ |> List.filter (fun p -> match p with |InputMSAFile _ -> true |_ -> false)
+ |> fun x -> if List.isEmpty x then
+ failwith "no input sequence given"
+ else
+ HMMbuildParams.makeCmdWith bcContext.Mount x.[0]
+
+ let hmm =
+ opt
+ |> List.filter (fun p -> match p with |OutputHMMFile _ -> true |_ -> false)
+ |> fun x -> if List.isEmpty x then
+ failwith "no output sequence given"
+ else
+ HMMbuildParams.makeCmdWith bcContext.Mount x.[0]
+
+ let options = opt |> List.filter (fun p -> match p with |InputMSAFile _ |OutputHMMFile _ -> false |_ -> true)
+ let cmds = (options |> List.map (HMMbuildParams.makeCmdWith bcContext.Mount))
+ let tp = ("hmmbuild"::(cmds |> List.concat))@hmm@msa
+
+ printfn "Starting process hmmbuildn\r\nparameters:"
+ printfn "%s" msa.[0]
+ printfn "%s" hmm.[0]
+ cmds |> List.iter (fun op -> printfn "\t%s" (String.concat " " op))
+
+ async {
+ let! res = BioContainer.execAsync bcContext tp
+ return res
+
+ }
+
+ ///For each multiple sequence alignment in a msafile build a profile HMM and save it to a
+ ///new hmmfile.
+ let runHMMbuild (bcContext:BioContainer.BcContext) (opt:HMMbuildParams list) =
+ runHMMbuildAsync bcContext opt
+ |> Async.RunSynchronously
+
+ ///hmmalign - align sequences to a profile
+ module HMMalign =
+ //Usage: hmmalign [-options]
+
+ //Basic options:
+ // -h : show brief help on version and usage
+ // -o : output alignment to file , not stdout
+
+ //Less common options:
+ // --mapali : include alignment in file (same ali that HMM came from)
+ // --trim : trim terminal tails of nonaligned residues from alignment
+ // --amino : assert , both protein: no autodetection
+ // --dna : assert , both DNA: no autodetection
+ // --rna : assert , both RNA: no autodetection
+ // --informat : assert is in format : no autodetection
+ // --outformat : output alignment in format [Stockholm]
+
+ type HMMalignParams =
+ ///HMM input file
+ | InputHMMFile of string
+ ///Input sequence file to align.
+ ///Sequence input formats include: FASTA, EMBL, GenBank, UniProt
+ | InputSequenceFile of string
+ ///output alignment to file not stdout
+ ///Alignment output formats include: Stockholm, Pfam, A2M, PSIBLAST
+ | SummaryToFile of string
+ ///assert alphabet to both input hmm and seqfile
+ | Alphabet of AlphabetType list
+ ///assert is in format : no autodetection
+ | SequenceFileFormat of string
+ ///output alignment in format
+ | OutFormat of string
+ ///trim terminal tails of nonaligned residues from alignment
+ | Trim
+ ///include alignment in file (same ali that HMM came from)
+ | MapAlignmentToFile of string
+
+ static member makeCmd = function
+ | InputHMMFile path -> [path]
+ | InputSequenceFile path -> [path]
+ | SummaryToFile path -> ["-o";path]
+ | Alphabet aList -> aList |> List.map AlphabetType.make |> List.concat
+ | SequenceFileFormat f -> ["--informat" ;f]
+ | OutFormat f -> ["--outformat" ;f]
+ | Trim -> ["--trim"]
+ | MapAlignmentToFile path -> ["--mapali"; path]
+
+ static member makeCmdWith (m:MountInfo) =
+ let cPath p = (MountInfo.containerPathOf m p)
+ function
+ | InputHMMFile path -> [cPath path]
+ | InputSequenceFile path -> [cPath path]
+ | SummaryToFile path -> ["-o";cPath path]
+ | Alphabet aList -> aList |> List.map AlphabetType.make |> List.concat
+ | SequenceFileFormat f -> ["--informat" ;f]
+ | OutFormat f -> ["--outformat" ;f]
+ | Trim -> ["--trim"]
+ | MapAlignmentToFile path -> ["--mapali"; cPath path]
+
+
+
+ let runHMMalignAsync (bcContext:BioContainer.BcContext) (opt:HMMalignParams list) =
+ //Usage: hmmalign [options] hmmfile seqfile
+ let hmm =
+ opt
+ |> List.filter (fun p -> match p with |InputHMMFile _ -> true |_ -> false)
+ |> fun x -> if List.isEmpty x then
+ failwith "no input hmm given"
+ else
+ HMMalignParams.makeCmdWith bcContext.Mount x.[0]
+
+ let seqFile =
+ opt
+ |> List.filter (fun p -> match p with |InputSequenceFile _ -> true |_ -> false)
+ |> fun x -> if List.isEmpty x then
+ failwith "no input sequence file given"
+ else
+ HMMalignParams.makeCmdWith bcContext.Mount x.[0]
+
+ let options = opt |> List.filter (fun p -> match p with |InputHMMFile _ |InputSequenceFile _ -> false |_ -> true)
+ let cmds = (options |> List.map (HMMalignParams.makeCmdWith bcContext.Mount))
+ let tp = ("hmmalign"::(cmds |> List.concat))@hmm@seqFile
+
+ printfn "Starting process hmmalign\r\nparameters:"
+ printfn "%s" hmm.[0]
+ printfn "%s" seqFile.[0]
+ cmds |> List.iter (fun op -> printfn "\t%s" (String.concat " " op))
+
+ async {
+ let! res = BioContainer.execAsync bcContext tp
+ return res
+
+ }
+ ///Perform a multiple sequence alignment of all the sequences in seqfile by aligning
+ ///them individually to the profile HMM in hmmfile. The new alignment is output to
+ ///stdout (when not specified otherwise).
+ let runHMMalign (bcContext:BioContainer.BcContext) (opt:HMMalignParams list) =
+ runHMMalignAsync bcContext opt
+ |> Async.RunSynchronously
+
+ ///hmmsearch - search profile(s) against a sequence database
+ module HMMsearch =
+
+ ///Options directing output:
+ // -o : direct output to file , not stdout
+ // -A : save multiple alignment of all hits to file
+ // --tblout : save parseable table of per-sequence hits to file
+ // --domtblout : save parseable table of per-domain hits to file
+ // --pfamtblout : save table of hits and domains to file, in Pfam format
+ // --acc : prefer accessions over names in output
+ // --noali : don't output alignments, so output is smaller
+ // --notextw : unlimit ASCII text output line width
+ // --textw : set max width of ASCII text output lines [120] (n>=120)
+
+ type OutputDirectionsOptions =
+ ///direct output to file , not stdout
+ | OutputToFile of string
+ ///save multiple alignment of all hits to file
+ | AllHitsToFile of string
+ ///save parseable table of per-sequence hits to file
+ | HitsToPerSequenceTable of string
+ ///save parseable table of per-domain hits to file
+ | HitsToPerDomainTable of string
+ ///save table of hits and domains to file, in Pfam format
+ | HitsToPfam of string
+ ///prefer accessions over names in output
+ | PreferAccessionsOverNames
+ ///don't output alignments, so output is smaller
+ | NoAlignments
+ ///unlimit ASCII text output line width
+ | UnlimitedTextLineWidth
+ ///set max width of ASCII text output lines
+ | MaxTextLineWidth of int
+
+ static member make =
+ function
+ | OutputToFile path -> ["-o"; path]
+ | AllHitsToFile path -> ["-A"; path]
+ | HitsToPerSequenceTable path -> ["--tblout"; path]
+ | HitsToPerDomainTable path -> ["--domtblout"; path]
+ | HitsToPfam path -> ["--pfamtblout"; path]
+ | PreferAccessionsOverNames -> ["--acc"]
+ | NoAlignments -> ["--noali"]
+ | UnlimitedTextLineWidth -> ["--notextw "]
+ | MaxTextLineWidth lw -> ["--textw"]
+
+ static member makeWith (m:MountInfo) =
+ let cPath p = (MountInfo.containerPathOf m p)
+ function
+ | OutputToFile path -> ["-o"; cPath path]
+ | AllHitsToFile path -> ["-A"; cPath path]
+ | HitsToPerSequenceTable path -> ["--tblout"; cPath path]
+ | HitsToPerDomainTable path -> ["--domtblout"; cPath path]
+ | HitsToPfam path -> ["--pfamtblout"; cPath path]
+ | PreferAccessionsOverNames -> ["--acc"]
+ | NoAlignments -> ["--noali"]
+ | UnlimitedTextLineWidth -> ["--notextw "]
+ | MaxTextLineWidth lw -> ["--textw"]
+
+
+
+ ///Other expert options:
+ // --nonull2 : turn off biased composition score corrections
+ // -Z : set # of comparisons done, for E-value calculation
+ // --domZ : set # of significant seqs, for domain E-value calculation
+ // --seed : set RNG seed to (if 0: one-time arbitrary seed) [42]
+ // --tformat : assert target is in format : no autodetection
+ // --cpu : number of parallel CPU workers to use for multithreads
+
+ type MiscellaneousOptions =
+ ///turn off biased composition score corrections
+ | TurnOffBiasedScoreCorrections
+ ///set # of comparisons done, for E-value calculation
+ | EValueComparisons of int
+ ///set # of significant seqs, for domain E-value calculation
+ | NumberOfSigSeqs of int
+ ///set RNG seed to (if 0: one-time arbitrary seed)
+ | RNGSeed of int
+ ///assert target is in format : no autodetection
+ | SequenceFileFormat of string
+ ///number of parallel CPU workers to use for multithreads
+ | Threads of int
+
+ static member make =
+ function
+ | TurnOffBiasedScoreCorrections -> ["--nonull2"]
+ | EValueComparisons n -> ["-Z"; string n]
+ | NumberOfSigSeqs n -> ["--domZ"; string n]
+ | RNGSeed s -> ["--seed"; string s]
+ | SequenceFileFormat f -> ["--tformat"; f]
+ | Threads t -> ["--cpu"; string t]
+
+ type HMMsearchParams =
+ /// Input HMM file containing query profiles to search for
+ | InputHMMFile of string
+ /// Input sequence database to search query profiles in
+ | SequenceDB of string
+ ///Options directing output
+ | OutputDirections of OutputDirectionsOptions list
+ ///Options controlling reporting thresholds
+ | ReportingThreshold of ReportingThresholdOptions list
+ ///Options controlling inclusion (significance) thresholds
+ | InclusionThreshold of InclusionThresholdOptions list
+ ///Options controlling model-specific thresholding
+ | ModelSpecificThreshold of ModelSpecificThresholdOptions list
+ ///Options controlling acceleration heuristics
+ | AccelerationHeuristics of AccelerationHeuristicsOptions list
+ ///Other expert options
+ | Miscellaneous of MiscellaneousOptions list
+
+ static member makeCMD =
+ function
+ | InputHMMFile path -> [path]
+ | SequenceDB path -> [path]
+ | OutputDirections odList -> odList |> List.map OutputDirectionsOptions.make |> List.concat
+ | ReportingThreshold rtList -> rtList |> List.map ReportingThresholdOptions.make |> List.concat
+ | InclusionThreshold itList -> itList |> List.map InclusionThresholdOptions.make |> List.concat
+ | ModelSpecificThreshold mstList -> mstList|> List.map ModelSpecificThresholdOptions.make |> List.concat
+ | AccelerationHeuristics ahList -> ahList |> List.map AccelerationHeuristicsOptions.make |> List.concat
+ | Miscellaneous mList -> mList |> List.map MiscellaneousOptions.make |> List.concat
+
+ static member makeCmdWith (m:MountInfo) =
+ let cPath p = (MountInfo.containerPathOf m p)
+ function
+ | InputHMMFile path -> [cPath path]
+ | SequenceDB path -> [cPath path]
+ | OutputDirections odList -> odList |> List.map (OutputDirectionsOptions.makeWith m) |> List.concat
+ | ReportingThreshold rtList -> rtList |> List.map ReportingThresholdOptions.make |> List.concat
+ | InclusionThreshold itList -> itList |> List.map InclusionThresholdOptions.make |> List.concat
+ | ModelSpecificThreshold mstList -> mstList|> List.map ModelSpecificThresholdOptions.make |> List.concat
+ | AccelerationHeuristics ahList -> ahList |> List.map AccelerationHeuristicsOptions.make |> List.concat
+ | Miscellaneous mList -> mList |> List.map MiscellaneousOptions.make |> List.concat
+ //Usage: hmmsearch [options]
+
+ let runHMMsearchAsync (bcContext:BioContainer.BcContext) (opt:HMMsearchParams list) =
+ //hmmsearch [options] hmmfile seqdb
+ let hmm =
+ opt
+ |> List.filter (fun p -> match p with |InputHMMFile _ -> true |_ -> false)
+ |> fun x -> if List.isEmpty x then
+ failwith "no input hmm given"
+ else
+ HMMsearchParams.makeCmdWith bcContext.Mount x.[0]
+
+ let seqDB =
+ opt
+ |> List.filter (fun p -> match p with |SequenceDB _ -> true |_ -> false)
+ |> fun x -> if List.isEmpty x then
+ failwith "no input sequence file given"
+ else
+ HMMsearchParams.makeCmdWith bcContext.Mount x.[0]
+
+ let options = opt |> List.filter (fun p -> match p with |InputHMMFile _ |SequenceDB _ -> false |_ -> true)
+ let cmds = (options |> List.map (HMMsearchParams.makeCmdWith bcContext.Mount))
+ let tp = ("hmmsearch"::(cmds |> List.concat))@hmm@seqDB
+
+ printfn "Starting process hmmsearch\r\nparameters:"
+ printfn "%s" hmm.[0]
+ printfn "%s" seqDB.[0]
+ cmds |> List.iter (fun op -> printfn "\t%s" (String.concat " " op))
+
+ async {
+ let! res = BioContainer.execAsync bcContext tp
+ return res
+
+ }
+
+ ///hmmsearch is used to search one or more profiles against a sequence database. For each
+ ///profile in hmmfile, use that query profile to search the target database of sequences in
+ ///seqdb, and output ranked lists of the sequences with the most significant matches to
+ ///the profile
+ let runHMMalign (bcContext:BioContainer.BcContext) (opt:HMMsearchParams list) =
+ runHMMsearchAsync bcContext opt
+ |> Async.RunSynchronously
+
+ ///hmmscan - search sequence(s) against a profile database
+ module HMMscan =
+
+ //Usage: hmmscan [-options]
+
+ ///Options controlling output:
+ // -o : direct output to file , not stdout
+ // --tblout : save parseable table of per-sequence hits to file
+ // --domtblout : save parseable table of per-domain hits to file
+ // --pfamtblout : save table of hits and domains to file, in Pfam format
+ // --acc : prefer accessions over names in output
+ // --noali : don't output alignments, so output is smaller
+ // --notextw : unlimit ASCII text output line width
+ // --textw : set max width of ASCII text output lines [120] (n>=120)
+
+ //TO-DO: maybe unify overlap with output direcction options of other hmmer functions
+ type OutputDirectionsOptions =
+ ///direct output to file , not stdout
+ | OutputToFile of string
+ ///save parseable table of per-sequence hits to file
+ | HitsToPerSequenceTable of string
+ ///save parseable table of per-domain hits to file
+ | HitsToPerDomainTable of string
+ ///save table of hits and domains to file, in Pfam format
+ | HitsToPfam of string
+ ///prefer accessions over names in output
+ | PreferAccessionsOverNames
+ ///don't output alignments, so output is smaller
+ | NoAlignments
+ ///unlimit ASCII text output line width
+ | UnlimitedTextLineWidth
+ ///set max width of ASCII text output lines
+ | MaxTextLineWidth of int
+
+ static member make =
+ function
+ | OutputToFile path -> ["-o"; path]
+ | HitsToPerSequenceTable path -> ["--tblout"; path]
+ | HitsToPerDomainTable path -> ["--domtblout"; path]
+ | HitsToPfam path -> ["--pfamtblout"; path]
+ | PreferAccessionsOverNames -> ["--acc"]
+ | NoAlignments -> ["--noali"]
+ | UnlimitedTextLineWidth -> ["--notextw "]
+ | MaxTextLineWidth lw -> ["--textw"]
+
+ static member makeWith (m:MountInfo) =
+ let cPath p = (MountInfo.containerPathOf m p)
+ function
+ | OutputToFile path -> ["-o"; cPath path]
+ | HitsToPerSequenceTable path -> ["--tblout"; cPath path]
+ | HitsToPerDomainTable path -> ["--domtblout"; cPath path]
+ | HitsToPfam path -> ["--pfamtblout"; cPath path]
+ | PreferAccessionsOverNames -> ["--acc"]
+ | NoAlignments -> ["--noali"]
+ | UnlimitedTextLineWidth -> ["--notextw "]
+ | MaxTextLineWidth lw -> ["--textw"]
+
+ ///Other expert options:
+ // --nonull2 : turn off biased composition score corrections
+ // -Z : set # of comparisons done, for E-value calculation
+ // --domZ : set # of significant seqs, for domain E-value calculation
+ // --seed : set RNG seed to (if 0: one-time arbitrary seed) [42]
+ // --qformat : assert input is in format : no autodetection
+ // --daemon : run program as a daemon
+ // --cpu : number of parallel CPU workers to use for multithreads
+
+ type MiscellaneousOptions =
+ ///turn off biased composition score corrections
+ | TurnOffBiasedScoreCorrections
+ ///set # of comparisons done, for E-value calculation
+ | EValueComparisons of int
+ ///set # of significant seqs, for domain E-value calculation
+ | NumberOfSigSeqs of int
+ ///set RNG seed to (if 0: one-time arbitrary seed)
+ | RNGSeed of int
+ ///assert target is in format : no autodetection
+ | SequenceFileFormat of string
+ ///number of parallel CPU workers to use for multithreads
+ | Threads of int
+ ///run program as a daemon
+ | RunAsDaemon
+
+ static member make =
+ function
+ | TurnOffBiasedScoreCorrections -> ["--nonull2"]
+ | EValueComparisons n -> ["-Z"; string n]
+ | NumberOfSigSeqs n -> ["--domZ"; string n]
+ | RNGSeed s -> ["--seed"; string s]
+ | SequenceFileFormat f -> ["--qformat"; f]
+ | Threads t -> ["--cpu"; string t]
+ | RunAsDaemon -> ["--daemon"]
+
+
+ type HMMscanParams =
+ | InputHMMDB of string
+ | InputSequenceFile of string
+ | OutputDirections of OutputDirectionsOptions list
+ ///Options controlling reporting thresholds
+ | ReportingThreshold of ReportingThresholdOptions list
+ ///Options controlling inclusion (significance) thresholds
+ | InclusionThreshold of InclusionThresholdOptions list
+ ///Options controlling model-specific thresholding
+ | ModelSpecificThreshold of ModelSpecificThresholdOptions list
+ ///Options controlling acceleration heuristics
+ | AccelerationHeuristics of AccelerationHeuristicsOptions list
+ ///Other expert options
+ | Miscellaneous of MiscellaneousOptions list
+
+ static member makeCmd = function
+ | InputHMMDB path -> [path]
+ | InputSequenceFile path -> [path]
+ | OutputDirections pList -> pList |> List.map OutputDirectionsOptions.make |> List.concat
+ | ReportingThreshold pList -> pList |> List.map ReportingThresholdOptions.make |> List.concat
+ | InclusionThreshold pList -> pList |> List.map InclusionThresholdOptions.make |> List.concat
+ | ModelSpecificThreshold pList -> pList |> List.map ModelSpecificThresholdOptions.make |> List.concat
+ | AccelerationHeuristics pList -> pList |> List.map AccelerationHeuristicsOptions.make |> List.concat
+ | Miscellaneous pList -> pList |> List.map MiscellaneousOptions.make |> List.concat
+
+ static member makeCmdWith (m:MountInfo) =
+ let cPath p = (MountInfo.containerPathOf m p)
+ function
+ | InputHMMDB path -> [cPath path]
+ | InputSequenceFile path -> [cPath path]
+ | OutputDirections pList -> pList |> List.map (OutputDirectionsOptions.makeWith m) |> List.concat
+ | ReportingThreshold pList -> pList |> List.map ReportingThresholdOptions.make |> List.concat
+ | InclusionThreshold pList -> pList |> List.map InclusionThresholdOptions.make |> List.concat
+ | ModelSpecificThreshold pList -> pList |> List.map ModelSpecificThresholdOptions.make |> List.concat
+ | AccelerationHeuristics pList -> pList |> List.map AccelerationHeuristicsOptions.make |> List.concat
+ | Miscellaneous pList -> pList |> List.map MiscellaneousOptions.make |> List.concat
+
+ let runHMMscanAsync (bcContext:BioContainer.BcContext) (opt:HMMscanParams list) =
+ //Usage: hmmscan [-options]
+ let hmmDB =
+ opt
+ |> List.filter (fun p -> match p with |InputHMMDB _ -> true |_ -> false)
+ |> fun x -> if List.isEmpty x then
+ failwith "no input hmm db given"
+ else
+ HMMscanParams.makeCmdWith bcContext.Mount x.[0]
+
+ let inSeq =
+ opt
+ |> List.filter (fun p -> match p with |InputSequenceFile _ -> true |_ -> false)
+ |> fun x -> if List.isEmpty x then
+ failwith "no input sequence file given"
+ else
+ HMMscanParams.makeCmdWith bcContext.Mount x.[0]
+
+ let options = opt |> List.filter (fun p -> match p with |InputHMMDB _ |InputSequenceFile _ -> false |_ -> true)
+ let cmds = (options |> List.map (HMMscanParams.makeCmdWith bcContext.Mount))
+ let tp = ("hmmscan"::(cmds |> List.concat))@hmmDB@inSeq
+
+ printfn "Starting process hmmscan\r\nparameters:"
+ printfn "%s" hmmDB.[0]
+ printfn "%s" inSeq.[0]
+ cmds |> List.iter (fun op -> printfn "\t%s" (String.concat " " op))
+
+ async {
+ let! res = BioContainer.execAsync bcContext tp
+ return res
+
+ }
+
+ ///hmmscan is used to search protein sequences against collections of protein profiles.
+ ///For each sequence in seqfile, use that query sequence to search the target database
+ ///of profiles in hmmdb, and output ranked lists of the profiles with the most significant
+ ///matches to the sequence
+ let runHMMscan (bcContext:BioContainer.BcContext) (opt:HMMscanParams list) =
+ runHMMscanAsync bcContext opt
+ |> Async.RunSynchronously
+
+
+ ///hmmpress - prepare a profile database for hmmscan
+ module HMMpress =
+
+ type HMMpressParams =
+ ///HMM profile to construct the binary compressed datafiles from
+ | HMMInputFile of string
+ ///overwrites any previous hmmpress’ed datafiles
+ | ForceOverwirite
+
+ static member makeCmd =
+ function
+ | HMMInputFile path -> [path]
+ | ForceOverwirite -> ["-f"]
+
+ static member makeCmdWith (m:MountInfo) =
+ let cPath p = (MountInfo.containerPathOf m p)
+ function
+ | HMMInputFile path -> [cPath path]
+ | ForceOverwirite -> ["-f"]
+
+ let runHMMpressAsync (bcContext:BioContainer.BcContext) (opt:HMMpressParams list) =
+ //Usage: hmmpress [options] hmmfile
+ let hmm =
+ opt
+ |> List.filter (fun p -> match p with |HMMInputFile _ -> true |_ -> false)
+ |> fun x -> if List.isEmpty x then
+ failwith "no input hmm db given"
+ else
+ HMMpressParams.makeCmdWith bcContext.Mount x.[0]
+
+ let options = opt |> List.filter (fun p -> match p with |HMMInputFile _ -> false |_ -> true)
+ let cmds = (options |> List.map (HMMpressParams.makeCmdWith bcContext.Mount))
+ let tp = ("hmmpress"::(cmds |> List.concat))@hmm
+
+ printfn "Starting process hmmpress\r\nparameters:"
+ printfn "%s" hmm.[0]
+ cmds |> List.iter (fun op -> printfn "\t%s" (String.concat " " op))
+
+ async {
+ let! res = BioContainer.execAsync bcContext tp
+ return res
+
+ }
+
+ ///Constructs binary compressed datafiles for hmmscan, starting from a profile database
+ ///hmmfile in standard HMMER3 format. The hmmpress step is required for hmmscan to work.
+ let runHMMpress (bcContext:BioContainer.BcContext) (opt:HMMpressParams list) =
+ runHMMpressAsync bcContext opt
+ |> Async.RunSynchronously
+
+
+ module HMMconvert =
+
+ //Usage: hmmconvert [-options]
+
+ //Options:
+ // -h : show brief help on version and usage
+ // -a : ascii: output models in HMMER3 ASCII format [default]
+ // -b : binary: output models in HMMER3 binary format
+ // -2 : HMMER2: output backward compatible HMMER2 ASCII format (ls mode)
+ // --outfmt : choose output legacy 3.x file formats by name, such as '3/a'
+
+ type OutputFormatOptions =
+ | ASCII
+ | Binary
+ | HMMER2
+
+ static member make = function
+ | ASCII -> ["-a"]
+ | Binary -> ["-b"]
+ | HMMER2 -> ["-2"]
+
+ type HMMconvertParams =
+ | HMMInputFile of string
+ | OutputFormat of OutputFormatOptions list
+ | LegacyFormat of string
+
+ static member make = function
+ | HMMInputFile path -> [path]
+ | OutputFormat pList -> pList |> List.map OutputFormatOptions.make |> List.concat
+ | LegacyFormat f -> ["--outfmt"; f]
+
+ static member makeCmdWith (m:MountInfo) =
+ let cPath p = (MountInfo.containerPathOf m p)
+ function
+ | HMMInputFile path -> [cPath path]
+ | OutputFormat pList -> pList |> List.map OutputFormatOptions.make |> List.concat
+ | LegacyFormat f -> ["--outfmt"; f]
+
+ let runHMMconvertAsync (bcContext:BioContainer.BcContext) (opt:HMMconvertParams list) =
+ //Usage: hmmconvert [options] hmmfile
+ let hmm =
+ opt
+ |> List.filter (fun p -> match p with |HMMInputFile _ -> true |_ -> false)
+ |> fun x -> if List.isEmpty x then
+ failwith "no input hmm given"
+ else
+ HMMconvertParams.makeCmdWith bcContext.Mount x.[0]
+
+ let options = opt |> List.filter (fun p -> match p with |HMMInputFile _ -> false |_ -> true)
+ let cmds = (options |> List.map (HMMconvertParams.makeCmdWith bcContext.Mount))
+ let tp = ("hmmconvert"::(cmds |> List.concat))@hmm
+
+ printfn "Starting process hmmconvert\r\nparameters:"
+ printfn "%s" hmm.[0]
+ cmds |> List.iter (fun op -> printfn "\t%s" (String.concat " " op))
+
+ async {
+ let! res = BioContainer.execAsync bcContext tp
+ return res
+
+ }
+
+ ///The hmmconvert utility converts an input profile file to different HMMER formats
+ let runHMMconvert (bcContext:BioContainer.BcContext) (opt:HMMconvertParams list) =
+ runHMMconvertAsync bcContext opt
+ |> Async.RunSynchronously
+
+ ///hmmemit - sample sequences from a profile
+ module HMMemit =
+
+
+ ///Options controlling what to emit:
+ // -a : emit alignment
+ // -c : emit simple majority-rule consensus sequence
+ // -C : emit fancier consensus sequence (req's --minl, --minu)
+ // -p : sample sequences from profile, not core model
+
+ type EmitControllingOptions =
+ ///emit alignment
+ | EmitAlignment
+ ///emit simple majority-rule consensus sequence
+ | EmitConsensus
+ ///emit fancier consensus sequence (req's --minl, --minu)
+ | EmitConsensusFancy
+ ///sample sequences from profile, not core model
+ | SampleFromProfile
+
+ static member make = function
+ | EmitAlignment -> ["-a"]
+ | EmitConsensus -> ["-c"]
+ | EmitConsensusFancy -> ["-C"]
+ | SampleFromProfile -> ["-p"]
+
+ ///Options controlling emission from profiles with SampleFromProfile:
+ // -L : set expected length from profile to