Python Response.urls_to_activity方法代码示例

# 需要导入模块: from models import Response [as 别名]
# 或者: from models.Response import urls_to_activity [as 别名]

#.........这里部分代码省略.........
                # since it probably came from the user's activity, so prefer this one.
                # background: https://github.com/snarfed/bridgy/issues/533
                existing = responses.get(id)
                if existing:
                    if source.gr_source.activity_changed(resp, existing, log=True):
                        logging.warning("Got two different versions of same response!\n%s\n%s", existing, resp)
                    resp["activities"].extend(existing.get("activities", []))

                responses[id] = resp

        #
        # Step 3: filter out responses we've already seen
        #
        # seen responses (JSON objects) for each source are stored in its entity.
        unchanged_responses = []
        if source.seen_responses_cache_json:
            for seen in json.loads(source.seen_responses_cache_json):
                id = seen["id"]
                resp = responses.get(id)
                if resp and not source.gr_source.activity_changed(seen, resp, log=True):
                    unchanged_responses.append(seen)
                    del responses[id]

        #
        # Step 4: store new responses and enqueue propagate tasks
        #
        pruned_responses = []
        for id, resp in responses.items():
            resp_type = Response.get_type(resp)
            activities = resp.pop("activities", [])
            if not activities and resp_type == "post":
                activities = [resp]
            too_long = set()
            urls_to_activity = {}
            for i, activity in enumerate(activities):
                # we'll usually have multiple responses for the same activity, and the
                # objects in resp['activities'] are shared, so cache each activity's
                # discovered webmention targets inside its object.
                if "originals" not in activity or "mentions" not in activity:
                    activity["originals"], activity["mentions"] = original_post_discovery.discover(
                        source,
                        activity,
                        fetch_hfeed=True,
                        include_redirect_sources=False,
                        already_fetched_hfeeds=fetched_hfeeds,
                    )

                targets = original_post_discovery.targets_for_response(
                    resp, originals=activity["originals"], mentions=activity["mentions"]
                )
                if targets:
                    logging.info(
                        "%s has %d webmention target(s): %s", activity.get("url"), len(targets), " ".join(targets)
                    )
                for t in targets:
                    if len(t) <= _MAX_STRING_LENGTH:
                        urls_to_activity[t] = i
                    else:
                        logging.warning("Giving up on target URL over %s chars! %s", _MAX_STRING_LENGTH, t)
                        too_long.add(t[: _MAX_STRING_LENGTH - 4] + "...")

            # store/update response entity. the prune_*() calls are important to
            # remove circular references in link responses, which are their own
            # activities. details in the step 2 comment above.
            pruned_response = util.prune_response(resp)
            pruned_responses.append(pruned_response)

# 需要导入模块: from models import Response [as 别名]
# 或者: from models.Response import urls_to_activity [as 别名]

#.........这里部分代码省略.........
      rsvps = Source.get_rsvps_from_event(obj)

      # coalesce responses. drop any without ids
      for resp in replies + likes + reactions + reposts + rsvps:
        id = resp.get('id')
        if not id:
          logging.error('Skipping response without id: %s', json.dumps(resp, indent=2))
          continue

        if source.is_blocked(resp):
          logging.info('Skipping response by blocked user: %s',
                       json.dumps(resp.get('author') or resp.get('actor'), indent=2))
          continue

        resp.setdefault('activities', []).append(activity)

        # when we find two responses with the same id, the earlier one may have
        # come from a link post or user mention, and this one is probably better
        # since it probably came from the user's activity, so prefer this one.
        # background: https://github.com/snarfed/bridgy/issues/533
        existing = responses.get(id)
        if existing:
          if source.gr_source.activity_changed(resp, existing, log=True):
            logging.warning('Got two different versions of same response!\n%s\n%s',
                            existing, resp)
          resp['activities'].extend(existing.get('activities', []))

        responses[id] = resp

    #
    # Step 3: filter out responses we've already seen
    #
    # seen responses (JSON objects) for each source are stored in its entity.
    unchanged_responses = []
    if source.seen_responses_cache_json:
      for seen in json.loads(source.seen_responses_cache_json):
        id = seen['id']
        resp = responses.get(id)
        if resp and not source.gr_source.activity_changed(seen, resp, log=True):
          unchanged_responses.append(seen)
          del responses[id]

    #
    # Step 4: store new responses and enqueue propagate tasks
    #
    pruned_responses = []
    for id, resp in responses.items():
      resp_type = Response.get_type(resp)
      activities = resp.pop('activities', [])
      if not activities and resp_type == 'post':
        activities = [resp]
      too_long = set()
      urls_to_activity = {}
      for i, activity in enumerate(activities):
        # we'll usually have multiple responses for the same activity, and the
        # objects in resp['activities'] are shared, so cache each activity's
        # discovered webmention targets inside its object.
        if 'originals' not in activity or 'mentions' not in activity:
          activity['originals'], activity['mentions'] = \
            original_post_discovery.discover(
              source, activity, fetch_hfeed=True,
              include_redirect_sources=False,
              already_fetched_hfeeds=fetched_hfeeds)

        targets = original_post_discovery.targets_for_response(
          resp, originals=activity['originals'], mentions=activity['mentions'])
        if targets:
          logging.info('%s has %d webmention target(s): %s', activity.get('url'),
                       len(targets), ' '.join(targets))
        for t in targets:
          if len(t) <= _MAX_STRING_LENGTH:
            urls_to_activity[t] = i
          else:
            logging.info('Giving up on target URL over %s chars! %s',
                         _MAX_STRING_LENGTH, t)
            too_long.add(t[:_MAX_STRING_LENGTH - 4] + '...')

      # store/update response entity. the prune_*() calls are important to
      # remove circular references in link responses, which are their own
      # activities. details in the step 2 comment above.
      pruned_response = util.prune_response(resp)
      pruned_responses.append(pruned_response)
      resp_entity = Response(
        id=id,
        source=source.key,
        activities_json=[json.dumps(util.prune_activity(a, source))
                         for a in activities],
        response_json=json.dumps(pruned_response),
        type=resp_type,
        unsent=list(urls_to_activity.keys()),
        failed=list(too_long),
        original_posts=resp.get('originals', []))
      if urls_to_activity and len(activities) > 1:
        resp_entity.urls_to_activity=json.dumps(urls_to_activity)
      resp_entity.get_or_save(source, restart=self.RESTART_EXISTING_TASKS)

    # update cache
    if pruned_responses:
      source.updates['seen_responses_cache_json'] = json.dumps(
        pruned_responses + unchanged_responses)