Add comments to requirements.txt (#465)

requirements.txt

@@ -1,18 +1,22 @@
+# core packages
 numpy
 pandas
-wget
+matchpy # used for optimisation of Transformer pipelines
+more_itertools
 tqdm
-pyjnius>=1.4.2
-matchpy
-deprecated
-chest
-scipy
+# used by datasets
 requests
-joblib
-more_itertools
 ir_datasets>=0.3.2
-jinja2
-statsmodels
-ir_measures>=0.3.1
-dill
+wget # used by mavenresolver to download jar files 
+pyjnius>=1.4.2 # used by pt.java to execute Java code, such as Terrier and Anserini
+deprecated # used for marking APIs as deprecated
+scipy # used for significance testing
+# evaluation related:
+ir_measures>=0.3.1 # used for evaluation measure computation
 pytrec_eval_terrier>=0.5.3
+jinja2 # used to get rendering of colour in Pandas by pt.Experiment
+statsmodels # used for multiple testing correction
+# others
+dill # used for serialisation that is better than pickle, e.g. passing wmodel functions to terrier.Retriever
+joblib # used in parallel support 
+chest # used for caching

setup.py

@@ -39,7 +39,15 @@ def get_version(rel_path):
 requirements = []
 with open('requirements.txt', 'rt') as f:
     for req in f.read().splitlines():
+        # A line that begins with # is treated as a comment and ignored. Whitespace followed 
+        # by a # causes the # and the remainder of the line to be treated as a comment.
+        if req.startswith("#"):
+            continue
+        req = req.split(" #")[0]
+        if req.strip() == "":
+            continue
         if req.startswith('git+'):
+            # support for git urls
             pkg_name = req.split('/')[-1].replace('.git', '')
             if "#egg=" in pkg_name:
                 pkg_name = pkg_name.split("#egg=")[1]