Document the functions

parser.py

@@ -6,12 +6,18 @@ from re import sub
 from pandas import DataFrame, read_html, Series
 
 
-def find_html_files(path) -> List:
+def remove_parenthesis(identifier):
-    file_list = glob(path + "/*fastqc.html")
+    """
-    return file_list
+    Remove parenthesis from the sequence identifier
+    """
+    sequence_str = "".join(map(str, identifier))
+    return sub(r"[()]", "", sequence_str)
 
 
 def extract_adapters(files) -> Tuple[Series, List]:
+    """
+    Extract the adapters sequences and statistics from the files
+    """
     all_adapters = DataFrame()
     for entry in files:
         tables = read_html(entry)
@@ -27,6 +33,9 @@ def extract_adapters(files) -> Tuple[Series, List]:
 
 
 def preprocess_dataframe(adapters) -> Series:
+    """
+    Remove empty sequences and duplicates
+    """
     na_free_adapters = adapters.dropna(axis=1)
     stacked_adapters = na_free_adapters.stack()
     duplicate_free_adapters = stacked_adapters.drop_duplicates()
@@ -34,15 +43,20 @@ def preprocess_dataframe(adapters) -> Series:
 
 
 def save_to_file(filename, adapters) -> None:
+    """
+    Save the adapter sequences as a FASTA file
+    """
     with open(filename, "w") as f:
         for index, value in adapters.iteritems():
-            sequence_str = "".join(map(str, index))
+            sequence_id = remove_parenthesis(index)
-            sequence_id = sub(r"[()]", "", sequence_str)
             fasta_entry = f">{sequence_id}\n{value}\n"
             f.write(fasta_entry)
 
 
 def parse_arguments():
+    """
+    Parse the command-line arguments
+    """
     parser = ArgumentParser()
     parser.add_argument("input", help="directory containing the fastqc reports")
     parser.add_argument("output", help="file where to export the sequences")
@@ -50,8 +64,11 @@ def parse_arguments():
 
 
 def main():
+    """
+    Extract the adapters from FASTQC reports to a FASTA file and show sequence length statistics
+    """
     args = parse_arguments()
-    file_list = find_html_files(args.input)
+    file_list = glob(args.input + "/*fastqc.html")
     adapters, stats = extract_adapters(file_list)
     save_to_file(args.output, adapters)
     print(