Skip to content

Latest commit

 

History

History
77 lines (49 loc) · 3.5 KB

INSTALL.md

File metadata and controls

77 lines (49 loc) · 3.5 KB

Geneplexus Web Server: Installing the application locally

These instructions are for internal developers to get started working ont his sytem and running it locally. The author do not currently provide support for this outside of the team members.

  1. Clone the repo
   git clone https://github.com/krishnanlab/geneplexus_app.git
   cd geneplexus_app

  1. Create the virtual environment
    python -m venv venv or conda create -n geneplexus_app

    and activate it

    .\venv\Scripts\activate or conda activate geneplexus_app

  2. Install the required packages

    pip install -r requirements.txt

  3. Download local data using the instructions under "Backend Data" below If you are on-campus, you can instead mount a SMB connection to

  4. create a .env file

    • must have two entries to run the front-end (with gene validation):
JOB_PATH=/path/to/job/files
DATA_PATH=/path/to/backend/datafiles

- For the system to send emails, config the email service. See [EMAIL.md](EMAIL.md) documention for details

NOTIFIER_EMAIL_ADDRESS='[email protected]'  # the address from which the app sends emails, 
                                             # and that's configured in sendgrid 
SENDGRID_API_KEY='some key '                 # get this from our sendgrid.com account
TEST_EMAIL_RECIPIENT='[email protected]'             # the email recipient you'd like to use for testing
  1. flask run

Input File

The file "input_genes.txt" has been included in the root folder of the application to use as a sample input

Backend Data

The application will not work without at least a subset of the backend data, which is available on the HPCC at the following location.

/mnt/research/compbio/krishnanlab/projects/GenePlexus/repos/GenePlexusBackend/data_backend2

The "data_backend2" folder will need to be created under the "app" folder within the application folder structure

Beneath the "data_backend" folder on the HPCC are several subfolders (e.g. GSCs, Edgelists, Node_Orders, etc). At a minimum the entire contents of the "ID_conversion" folder will need to be downloaded. Additionally, any files from the remaining folders that contain the name of the corresponding Network, Feature and Negative Class you want to be able to run.

For example, if you want to run the "BioGRID" network with the "Embedding" feature and the "DisGeNet" negative class every file with those attributes in the name of the file will need to be downloaded (in the same folder structure/names) (NOTE: The above example is a good place to start for the minimum files to download)

Testing the model run

The model is run from methods in models.py which are encapsulated in a wrapper script runner.py which accepts command line parameters, and is used by the back-end container. You can run a test job by sending the parameters like

source .env
python test/runner.py --output_path $OUTPUT_PATH --net_type BioGRID --features Embedding --GSC GO -j $JOBNAME -d $DATA_PATH  $GENE_FILE > $OUTPUT_FILE

There is a shell script that does all this for you, and as set of test genes, in the /test folder. You must have set DATA_PATH and JOB_PATH to valide folders in .env (or modify the script). From the top project folder, run

test/test_runner_py.sh

it will create a new folder inside $JOB_PATH to hold the job output.

Note there is (currently) no route/method in that app to run a job locally from inside the apap e.g. in a flask run , but there should be~