[wishlist] Produce comparison output into the terminal instead of html #32

Activity

1. blakev commented on Feb 1, 2018

   

   blakev

   on Feb 1, 2018

   More actions

   I'm experimenting with pkgdiff for my workflow and needed this too. This is the python code I wrote to extract the information; you can use it if you want.

   class Status(IntEnum):
       ADDED = 1
       CHANGED = 0
       MOVED = 3
       REMOVED = 2
       RENAMED = 4
       ERROR = 99
   
   
   status_cmp = {
       'added': Status.ADDED,
       'changed': Status.CHANGED,
       'moved': Status.MOVED,
       'removed': Status.REMOVED,
       'renamed': Status.RENAMED
   }
   
   def run_pkgdiff(path_lo, path_hi, dest=None):
       """Compare two .tar archives and generate a report of their difference.
   
       Args:
           path_lo (str):
           path_hi (str):
           dest (str):
   
       Returns:
           str:
       """
   
       if not check_tar(path_lo) and check_tar(path_hi):
           logger.error('cannot compare non tar archives, %s or %s', path_lo, path_hi)
           return None
   
       if dest is None:
           dest = tempfile.gettempdir()
   
       if os.path.isdir(dest):
           dest = os.path.join(dest, 'report.html')
   
       dest = os.path.abspath(dest)
   
       command = 'pkgdiff %s %s -hide-unchanged -list-added-removed -quick -report-path %s' % (
           path_lo, path_hi, dest)
   
       success = _run_process(command, track_output=True, success_statuses=(0, 1,))
       return dest if success else None
   
   
   def generate_diff(report_path):
       """Generate a list of files to pick/choose for the new distribution."""
   
       if not report_path or not os.path.exists(report_path):
           logger.error('cannot parse report, %s', report_path)
           return None
   
       with open(report_path, 'r') as ins_file:
           soup = BeautifulSoup(ins_file, 'lxml')
   
       logger.debug('created report %s', report_path)
   
       try:
   
           headers = list(soup.find_all('h2'))
           # ignore: Test Info, Test Results, Changes in Files . . . Packages
           headers = headers[3:-1]
   
           for section in headers:
               # each table represents all the file changes of a certain "type"
               table = section.find_next('table')
               rows = iter(table.findChildren('tr'))
   
               for row in rows:
                   # each row is an individual file that changed
                   cells = list(row.findChildren('td'))
                   if not cells or len(cells) == 0:
                       continue
                   # extract the two columns we care about
                   filepath, status = cells[:2]
   
                   filepath = os.path.join(filepath.text)
                   status = status_cmp.get(status.string.lower(), Status.ERROR)
   
                   # path attribute changes come in pairs of two rows
                   if status in (Status.RENAMED, Status.MOVED):
                       row_b = next(rows)
                       dest = row_b.findChildren('td')[0].text
                   else:
                       dest = None
   
                   yield filepath, dest, status
   
       except Exception as e:
           logger.exception(e, exc_info=True)
           logger.error('cannot parse report, possibly invalid')
           raise e