Added utils/logging.py, adapted setup.py, started working on revise.py

.gitignore

 .vscode/
 build/
 mossutils.egg-info/
+mossutils/utils/__pycache__/
+mossutils/__pycache__/

mossutils/moss.py

 import mosspy
 import argparse
-# import logging
-import pprint
+import logging
+
+from .utils import logging as logutils
 
 DEFAULT_USER_ID="3795777", # Malte's Moss ID
 DEFAULT_LANGUAGE="cc" # C++
@@ -12,7 +13,6 @@ DEFAULT_FILE_PATTERN="./*/main.cpp"
 DEFAULT_REPORT_FILE="moss-report.html"
 DEFAULT_REPORT_SUBDIR="_moss-report"
 DEFAULT_CONNECTIONS=8
-DEFAULT_LOGGING_LEVEL="INFO"
 
 def run_moss(
     user_id=DEFAULT_USER_ID,
@@ -32,36 +32,39 @@ def run_moss(
   comment_string = str(comments)
 
   # # A nicely formatted string representation requires a bit of hacking
+  # import pprint
   # comment_string = pprint.pformat(comments, indent=2)
   # comment_string = comment_string.replace("\n", "<br/>")
   # comment_string = comment_string.replace("  ", "&nbsp;&nbsp;")
 
   # Initialisation
-  logging.info("User id: {}".format(user_id))
-  logging.info("Language: {}".format(language))
+  logging.debug("User id: {}".format(user_id))
+  logging.debug("Language: {}".format(language))
 
   moss = mosspy.Moss(user_id, language)
 
   # Configure
-  logging.info("Ignore limit: {}".format(ignore_limit))
-  logging.info("Matches: {}".format(number_of_matching_files))
-  # logging.info("Comment: {}".format(comment_string))
+  logging.debug("Ignore limit: {}".format(ignore_limit))
+  logging.debug("Matches: {}".format(number_of_matching_files))
+  # logging.debug("Comment: {}".format(comment_string))
   
   moss.setIgnoreLimit(ignore_limit)
   moss.setNumberOfMatchingFiles(number_of_matching_files)
   moss.setCommentString(comment_string)
 
   # Add template file(s)
-  logging.info("Template: {}".format(template))
+  logging.debug("Template: {}".format(template))
 
   if template:
     moss.addBaseFile(template)
 
   # Add submission file
-  logging.info("File pattern: {}".format(file_pattern))
+  logging.debug("File pattern: {}".format(file_pattern))
 
   moss.addFilesByWildcard(file_pattern)
 
+  exit(0)
+
   logging.info("Sending files to Moss")
   url = moss.send() # Submission Report URL
 
@@ -124,12 +127,7 @@ def configure_cli_parser(parser):
     default=DEFAULT_REPORT_SUBDIR,
     help="Directory for individual report files (default: {})".format(DEFAULT_REPORT_SUBDIR))
   
-  parser.add_argument(
-    "--log-level",
-    type=str,
-    choices=("OFF", "DEBUG", "INFO", "WARNING", "ERROR"),
-    default=DEFAULT_LOGGING_LEVEL,
-    help="Log level (default: {})".format(DEFAULT_LOGGING_LEVEL))
+  logutils.add_loglevel_argument(parser)
 
   parser.add_argument(
     "pattern",
@@ -142,21 +140,7 @@ def main():
   configure_cli_parser(parser)
   args = parser.parse_args()
 
-  # See https://docs.python.org/3.8/library/logging.html#logging-levels
-  log_levels = {
-    "OFF": logging.NOTSET,
-    "DEBUG": logging.DEBUG,
-    "INFO": logging.INFO,
-    "WARNING": logging.WARNING,
-    "ERROR": logging.ERROR
-  }
-
-  log_level = log_levels[args.log_level]
-
-  # See https://docs.python.org/3.8/library/logging.html#logrecord-attributes
-  log_format = "%(message)s"
-
-  logging.basicConfig(level=log_level, format=log_format)
+  logutils.configure_level_and_format(args.log_level)
 
   run_moss(
     user_id=args.user_id,
@@ -168,5 +152,6 @@ def main():
     report_file=args.report_file,
     report_subdir=args.report_dir)
 
+
 if __name__ == '__main__':
   main()

mossutils/revise.py

+import re
+import logging
+import argparse
+from bs4 import BeautifulSoup
+
+from .utils import logging as logutils
+
+DEFAULT_REPORT_INPUT_FILE="moss-report.html"
+DEFAULT_REPORT_OUTPUT_FILE="moss-report-revised.html"
+
+def get_match_percentage(match_text):
+  percentage_pattern = r'\((\d+)%\)$'
+
+  percentage_string = re.search(percentage_pattern, match_text).group(1)
+  
+  return int(percentage_string)  
+
+# moss_report_input_filename = 'moss-report.html'
+# moss_report_ouput_filename = 'moss-report-fixed.html'
+
+def main():
+  with open(moss_report_input_filename) as input_file:
+    # We use BeautifulSoup because of it allows easy DOM manipulation and has
+    # a nice pretty-printing feature. The lxml parser is used because it is able
+    # to parse the totally broken HTML report file that Moss generates.
+    doc = BeautifulSoup(input_file, features='lxml')
+
+    doc.body.table['class'] = 'sortable'
+
+    ths = doc.find_all('th')
+    th = doc.new_tag('th')
+    th.string = 'Avg. %'
+    ths[1].insert_after(th)
+
+    for row in doc.find_all('tr'):
+      for a in row.find_all('a'):
+        a['target'] = '_blank'
+
+      # print(row)
+      cols = row.find_all('td')
+      first_match_text = cols[0].a.string
+      second_match_text = cols[1].a.string
+
+      first_percentage = get_match_percentage(first_match_text)
+      second_percentage = get_match_percentage(second_match_text)
+      avg_percentage = (first_percentage + second_percentage) / 2
+
+      td = doc.new_tag('td')
+      td.string = str(avg_percentage)
+      row.insert(2, td)
+
+    # <script src="sorttable.js"></script>
+    script = doc.new_tag('script')
+    script['src'] = 'sorttable.js' # Library from https://kryogenix.org/code/browser/sorttable/
+    doc.html.head.append(script)
+
+    # <link href="style.css" rel="stylesheet">
+    link = doc.new_tag('link')
+    link['href'] = 'style.css'
+    link['rel'] = 'stylesheet'
+    doc.html.head.append(link)
+
+    # <script src="script.js"></script>
+    script = doc.new_tag('script')
+    script['src'] = 'script.js'
+    doc.html.body.append(script)
+
+    with open(moss_report_ouput_filename, 'w') as output_file:
+      output_file.write(doc.prettify()) # 'utf-8'
+
+
+if __name__ == '__main__':
+  main()
\ No newline at end of file

mossutils/utils/logging.py

+import logging
+import argparse
+
+# See https://docs.python.org/3.8/library/logging.html#logging-levels
+LOG_LEVELS = {
+  "DEBUG": logging.DEBUG,
+  "INFO": logging.INFO,
+  "WARNING": logging.WARNING,
+  "ERROR": logging.ERROR,
+  "OFF": logging.CRITICAL + 1
+}
+
+DEFAULT_LOGGING_LEVEL="DEBUG"
+DEFAULT_LOG_FORMAT="[%(levelname)s] %(message)s"
+
+def add_loglevel_argument(parser, default_level=DEFAULT_LOGGING_LEVEL):
+  parser.add_argument(
+    "--log-level",
+    type=str,
+    choices=("OFF", "DEBUG", "INFO", "WARNING", "ERROR"),
+    default=default_level,
+    help="Log level (default: {})".format(default_level))
+
+def configure_level_and_format(level_string=DEFAULT_LOGGING_LEVEL, format=DEFAULT_LOG_FORMAT):
+  # See https://docs.python.org/3.8/library/logging.html#logrecord-attributes
+  logging.basicConfig(level=LOG_LEVELS[level_string], format=format)

setup.py

@@ -8,12 +8,20 @@ setup(
   author='Malte Schwerhoff',
   author_email='malte.schwerhoff@inf.ethz.ch',
   license='unlicensed',
-  packages=['mossutils'],
+  packages=[
+    'mossutils',
+    'mossutils.utils'
+  ],
   install_requires=[
     'mosspy',
+    'bs4',
+    'lxml'
   ],
   # scripts=['bin/mossutils-moss'],
   entry_points = {
-    "console_scripts": ['mu-moss = mossutils.moss:main']
+    "console_scripts": [
+      'mu-moss = mossutils.moss:main',
+      'mu-revise = mossutils.revise:main'
+    ]
   },
   zip_safe=False)