Android: Update modulcation LOC stat script

  - Allow running from outside the repo (--git-dir)
  - Date range for how many days to look back (--past-days)
  - Write metadata about start and end date parameters

Also moved the script under tools/android/modularization/loc

Bug: 1146478
Change-Id: Ic63ea085836d8b5750ada2299ba992221d97b9f9
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2526863
Commit-Queue: Jinsuk Kim <jinsukkim@chromium.org>
Reviewed-by: Peter Wen <wnwen@chromium.org>
Reviewed-by: Henrique Nakashima <hnakashima@chromium.org>
Cr-Commit-Position: refs/heads/master@{#826482}

tools/android/modularization-loc-stat.py → tools/android/modularization/loc/modularization_loc_stat.py

@@ -8,7 +8,9 @@
 """
 
 import argparse
+import datetime
 import json
+import os
 import subprocess
 import sys
 from collections import OrderedDict
@@ -22,14 +24,22 @@ _M12N_DIRS = [
 _LEGACY_DIR = 'chrome/android'
 
 
-def GenerateLOCStats(dates, quiet, json_format):
+def GenerateLOCStats(start_date,
+                     end_date,
+                     *,
+                     quiet=False,
+                     json_format=False,
+                     git_dir=None):
   """Generate modulazation LOC stats.
 
   Args:
-    dates: A tuple containing the range of dates for git patches to process.
+    start_date: The date to analyze the stat from.
+    end_date: The date to analyze the stat to.
     quiet: True if no message is output during the processing.
     json_format: True if the output should be in json format. Otherwise
         a plain, human-readable table is generated.
+    git_dir: Git repo directory to use for stats. If None, the current directory
+        is used.
 
   Return:
     Text string containing the stat in a specified format.
@@ -50,9 +60,11 @@ def GenerateLOCStats(dates, quiet, json_format):
   #  added-lines    deleted-lines  file-path1
   #  added-lines    deleted-lines  file-path2
   #  ...
+  repo_dir = git_dir or os.getcwd()
   command = [
-      'git', 'log', '--numstat', '--no-renames', '--format=#:%al:%cs:%s',
-      '--after=' + dates[0], '--before=' + dates[1], 'chrome', 'components'
+      'git', '-C', repo_dir, 'log', '--numstat', '--no-renames',
+      '--format=#:%al:%cs:%s', '--after=' + start_date, '--before=' + end_date,
+      'chrome', 'components'
   ]
   try:
     proc = subprocess.Popen(
@@ -111,23 +123,29 @@ def GenerateLOCStats(dates, quiet, json_format):
         'loc_modularized': total_m12n,
         'loc_legacy': total_legacy,
         'rankings': rankings,
+        'start_date': start_date,
+        'end_date': end_date,
     })
   else:
     output = []
-    percentage = 100.0 * total_m12n / (total_m12n + total_legacy)
+    total = total_m12n + total_legacy
+    percentage = 100.0 * total_m12n / total if total > 0 else 0
     output.append(f'# of lines added in modularized files: {total_m12n}')
     output.append(f'# of lines added in legacy files: {total_legacy}')
     output.append(f'% of lines landing in modularized files: {percentage:2.2f}')
 
     # Shows the top 50 contributors to modularized files.
     output.append('\nTop contributors:')
-    output.append('No  lines    %    author')
-    for rank, author in enumerate(list(rankings.keys())[:50], 1):
-      lines = rankings[author]
-      if lines == 0:
-        break
-      ratio = 100 * lines / total_m12n
-      output.append(f'{rank:2d} {lines:6d} {ratio:5.1f}  {author}')
+    if rankings:
+      output.append('No  lines    %    author')
+      for rank, author in enumerate(list(rankings.keys())[:50], 1):
+        lines = rankings[author]
+        if lines == 0:
+          break
+        ratio = 100 * lines / total_m12n
+        output.append(f'{rank:2d} {lines:6d} {ratio:5.1f}  {author}')
+    else:
+      output.append('...none found.')
     return '\n'.join(output)
 
 
@@ -151,15 +169,30 @@ def _print_progress(msg, prev_msg_len):
   print(msg, end='\r')
 
 
+def _get_date_range(args):
+  if args.date:
+    return args.date
+
+  today = datetime.date.today()
+  delta = datetime.timedelta(days=args.past_days)
+  past = datetime.datetime(today.year, today.month, today.day) - delta
+  return (past.date().isoformat(), today.isoformat())
+
+
 if __name__ == "__main__":
   parser = argparse.ArgumentParser(
       description="Generates LOC stats for modularization effort.")
-  parser.add_argument('--date',
-                      required=True,
-                      type=str,
-                      metavar=('<date-from>', '<date-to>'),
-                      nargs=2,
-                      help='date range (YYYY-MM-DD)~(YYYY-MM-DD)')
+
+  date_group = parser.add_mutually_exclusive_group(required=True)
+  date_group.add_argument('--date',
+                          type=str,
+                          metavar=('<date-from>', '<date-to>'),
+                          nargs=2,
+                          help='date range (YYYY-MM-DD)~(YYYY-MM-DD)')
+  date_group.add_argument('--past-days',
+                          type=int,
+                          help='The number of days to look back for stats. '
+                          '0 for today only.')
   parser.add_argument('-q',
                       '--quiet',
                       action='store_true',
@@ -174,9 +207,20 @@ if __name__ == "__main__":
                       type=str,
                       help='File to write the result to in json format. '
                       'If not specified, outputs to console.')
+  parser.add_argument('--git-dir',
+                      type=str,
+                      help='Root directory of the git repo to look into. '
+                      'If not specified, use the current directory.')
   args = parser.parse_args()
-  result = GenerateLOCStats(args.date, args.quiet, args.json)
-
+  if args.past_days and args.past_days < 0:
+    raise parser.error('--past-days must be non-negative.')
+
+  start_date, end_date = _get_date_range(args)
+  result = GenerateLOCStats(start_date,
+                            end_date,
+                            quiet=args.quiet,
+                            json_format=args.json,
+                            git_dir=args.git_dir)
   if args.output:
     with open(args.output, 'w') as f:
       f.write(result)