Added environment_full.yml and a simplified version of environment.yml.

README.md

@@ -56,6 +56,12 @@ Ran 92 tests in 47.056s
 OK
 ```
 
+Keep in mind that if you want to fully reproduce the results in the manuscript,
+then you need to install the full environment using the file
+`environment_full.yml`, which has additional dependencies. The one we used
+before (`environment.yml`) has the minimum set of packages needed to run
+Clustermatch.
+
 
 ## Reproducing results
 

environment.yml

@@ -3,20 +3,13 @@ channels:
 - defaults
 dependencies:
 - tabulate=0.8
-- requests=2.18
 - ipython=6.2
 - joblib=0.11
-- matplotlib=2.1
 - numpy=1.13
 - pandas=0.21
 - python=3.6
 - scikit-learn=0.19
 - scipy=1.0.0
-- seaborn=0.8
 - xlrd=1.1
 - xlwt=1.2
 - openpyxl=2.4
-- pillow=5.1
-- pip:
-  - minepy==1.2.1
-  - pyclustering==0.8.1

environment_full.yml

@@ -0,0 +1,22 @@
+name: clustermatch
+channels:
+- defaults
+dependencies:
+- tabulate=0.8
+- requests=2.18
+- ipython=6.2
+- joblib=0.11
+- matplotlib=2.1
+- numpy=1.13
+- pandas=0.21
+- python=3.6
+- scikit-learn=0.19
+- scipy=1.0.0
+- seaborn=0.8
+- xlrd=1.1
+- xlwt=1.2
+- openpyxl=2.4
+- pillow=5.1
+- pip:
+  - minepy==1.2.1
+  - pyclustering==0.8.1

utils/output.py

@@ -8,13 +8,23 @@
 from scipy.cluster.hierarchy import fcluster
 from scipy.spatial.distance import squareform, pdist
 import pandas as pd
-import matplotlib
-import seaborn as sns
-import requests
+
+try:
+    import matplotlib
+    import seaborn as sns
+    sns.set(context="paper", font="monospace")
+    MATPLOTLIB_INSTALLED = True
+except:
+    MATPLOTLIB_INSTALLED = False
+
+try:
+    import requests
+    REQUESTS_INSTALLED = True
+except:
+    REQUESTS_INSTALLED = False
 
 from utils.misc import get_temp_file_name
 
-sns.set(context="paper", font="monospace")
 
 RESULTS_DIR = 'results'
 
@@ -38,6 +48,9 @@ def func_wrapper(*args, **kwargs):
 
 
 def get_clustergrammer_link(square_matrix, names):
+    if not REQUESTS_INSTALLED:
+        raise ValueError('requests is not installed')
+
     # save sim matrix as csv and get clustergrammer visualization
     df = pd.DataFrame(square_matrix)
     df['names'] = names
@@ -170,6 +183,9 @@ def save_partitions(partitions, timestamp, extra_columns=None, columns_order=Non
 
 @setup_results_dir
 def save_coassociation_matrix(ensemble, partition, timestamp, columns_order=None, image_format='pdf'):
+    if not MATPLOTLIB_INSTALLED:
+        raise ValueError('matplotlib is not installed')
+
     condensed_matrix = _get_condensed_distance_matrix(ensemble)
     full_matrix = squareform(condensed_matrix)
     full_matrix = 1 - full_matrix
@@ -209,6 +225,9 @@ def save_coassociation_matrix(ensemble, partition, timestamp, columns_order=None
 
 @setup_results_dir
 def save_similarity_matrix(partition, feature_names, sim_matrix, timestamp, image_format='pdf'):
+    if not MATPLOTLIB_INSTALLED:
+        raise ValueError('matplotlib is not installed')
+
     partition_idxs_sorted = np.argsort(partition)
     sim_matrix_sorted = sim_matrix[partition_idxs_sorted, :]
     sim_matrix_sorted = sim_matrix_sorted[:, partition_idxs_sorted]
@@ -238,6 +257,9 @@ def save_reps_comparison(reps_comparison, timestamp):
 
 @setup_results_dir
 def save_clustermap(sim_matrix, feature_names, sources_names, partition_linkage, timestamp, image_format='pdf'):
+    if not MATPLOTLIB_INSTALLED:
+        raise ValueError('matplotlib is not installed')
+
     font_scale = (50.0 / len(feature_names))
     # sns.set(font_scale=font_scale)