# HG changeset patch # User ymh # Date 1637836083 -3600 # Node ID 39186950a53e49b8151be0427344a98b838c9b29 # Parent 7182d370c05f9f99a5230e837833e7d6f264bfeb Correct zoom export diff -r 7182d370c05f -r 39186950a53e script/lib/iri_tweet/setup.py --- a/script/lib/iri_tweet/setup.py Wed Sep 29 13:53:53 2021 +0200 +++ b/script/lib/iri_tweet/setup.py Thu Nov 25 11:28:03 2021 +0100 @@ -2,8 +2,8 @@ import os extra = {} -if sys.version_info >= (3, 0): - extra.update(use_2to3=True) +# if sys.version_info >= (3, 0): +# extra.update(use_2to3=True) try: diff -r 7182d370c05f -r 39186950a53e script/utils/export_chat_zoom.py --- a/script/utils/export_chat_zoom.py Wed Sep 29 13:53:53 2021 +0200 +++ b/script/utils/export_chat_zoom.py Thu Nov 25 11:28:03 2021 +0100 @@ -187,14 +187,17 @@ } return int(round(datetime.timedelta(**time_params).total_seconds()*1000)) -CHAT_REGEXP = re.compile(r"^(?P\d{2}:\d{2}:\d{2})\t\sFrom\s{2}(?P.+?)\s:\s(?P.*)$", re.DOTALL) -CHAT_LINE_REGEXP = re.compile(r"^\d{2}:\d{2}:\d{2}\t\sFrom\s{2}.+?\s:") +# CHAT_REGEXP = re.compile(r"^(?P\d{2}:\d{2}:\d{2})\t\sFrom\s{2}(?P.+?)\s:\s(?P.*)$", re.DOTALL) +CHAT_REGEXP = re.compile(r"^(?P\d{2}:\d{2}:\d{2})\t(?:(?:\sFrom\s{2}(?P.+?)\s)|(?P[^:]+)):\s(?P.*)$", re.DOTALL) +CHAT_LINE_REGEXP = re.compile(r"^\d{2}:\d{2}:\d{2}\t(?:(?:\sFrom\s{2}.+?\s)|(?:[^:]+)):") CHAT_DM_REGEXP = re.compile(r"\(Direct Message\)", re.IGNORECASE) def parse_chat_line(chat_id, chat_line): if (m := CHAT_REGEXP.match(chat_line)) is not None: res = {k: v.replace('\r','\n') if k == 'text' else v for k,v in m.groupdict().items()} res['id'] = chat_id + if user_str := res.get('user_from'): + res['user'] = user_str res['tags'] = re.findall('#(\w+)',res['text']) return res else: @@ -464,7 +467,6 @@ for i,chat_line in enumerate(chat_content_lines): cht = parse_chat_line("%04d" % (i+1) ,chat_line.strip()) - #TODO parse chat line cht_ts_dt = cht['created_at'] default_date = start_date or datetime.now()