diff --git a/swh/scheduler/updater/ghtorrent/__init__.py b/swh/scheduler/updater/ghtorrent/__init__.py index c11b71b..5738a87 100644 --- a/swh/scheduler/updater/ghtorrent/__init__.py +++ b/swh/scheduler/updater/ghtorrent/__init__.py @@ -1,143 +1,145 @@ # Copyright (C) 2018 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import json from kombu import Connection, Exchange, Queue from kombu.common import collect_replies from swh.core.config import SWHConfig from swh.scheduler.updater.events import SWHEvent from swh.scheduler.updater.consumer import UpdaterConsumer events = { # ghtorrent events related to github events (interesting) 'evt': [ 'commitcomment', 'create', 'delete', 'deployment', 'deploymentstatus', 'download', 'follow', 'fork', 'forkapply', 'gist', 'gollum', 'issuecomment', 'issues', 'member', 'membership', 'pagebuild', 'public', 'pullrequest', 'pullrequestreviewcomment', 'push', 'release', 'repository', 'status', 'teamadd', 'watch' ], # ghtorrent events related to mongodb insert (not interesting) 'ent': [ 'commit_comments', 'commits', 'followers', 'forks', 'geo_cache', 'issue_comments', 'issue_events', 'issues', 'org_members', 'pull_request_comments', 'pull_requests', 'repo_collaborators', 'repo_labels', 'repos', 'users', 'watchers' ] } class RabbitMQConn(SWHConfig): """RabbitMQ Connection class """ CONFIG_BASE_FILENAME = 'backend/ghtorrent' DEFAULT_CONFIG = { 'conn': ('dict', { 'user': 'guest', 'pass': 'guest', 'port': 5672, 'server': 'localhost', 'exchange_name': 'ght-streams', 'routing_key': 'something', 'queue_name': 'fake-events' }) } ADDITIONAL_CONFIG = {} def _connection_string(self): """Build the connection queue string. """ return 'amqp://%s:%s@%s:%s' % ( self.config['conn']['user'], self.config['conn']['pass'], self.config['conn']['server'], self.config['conn']['port'] ) def __init__(self, **config): super().__init__() if config: self.config = config else: self.config = self.parse_config_file( additional_configs=[self.ADDITIONAL_CONFIG]) self.conn_string = self._connection_string() self.exchange = Exchange(self.config['conn']['exchange_name'], 'topic', durable=True) self.routing_key = self.config['conn']['routing_key'] self.queue = Queue(self.config['conn']['queue_name'], exchange=self.exchange, routing_key=self.routing_key, auto_delete=True) def convert_event(event): """Given ghtorrent event, convert it to an swhevent instance. """ if isinstance(event, str): event = json.loads(event) keys = ['type', 'repo', 'created_at'] for k in keys: if k not in event: raise ValueError('Event should have the \'%s\' entry defined' % k) _type = event['type'].lower().rstrip('Event') _repo_name = 'https://github.com/%s' % event['repo']['name'] return SWHEvent({ 'type': _type, 'url': _repo_name, 'last_seen': event['created_at'] }) class GHTorrentConsumer(RabbitMQConn, UpdaterConsumer): """GHTorrent events consumer """ ADDITIONAL_CONFIG = { 'debug': ('bool', False), - 'batch': ('int', 1000), + 'batch_cache_write': ('int', 1000), + 'rabbitmq_prefetch_read': ('int', 100), } def __init__(self, **config): super().__init__(**config) self.debug = self.config['debug'] - self.batch = self.config['batch'] + self.batch = self.config['batch_cache_write'] + self.prefetch_read = self.config['rabbitmq_prefetch_read'] def has_events(self): return True def convert_event(self, event): """Given ghtorrent event, convert it to an swhevent instance. """ return convert_event(event) def open_connection(self): self.conn = Connection(self.conn_string) self.conn.connect() def close_connection(self): self.conn.release() def consume(self): """Consume events from rabbitmq connection """ yield from collect_replies( - self.conn, self.conn.channel(), self.queue, no_ack=False, - limit=100) + self.conn, self.conn.channel(), self.queue, + no_ack=False, limit=self.prefetch_read)