diff options
| author | bnewbold <bnewbold@archive.org> | 2020-02-14 17:55:39 +0000 | 
|---|---|---|
| committer | bnewbold <bnewbold@archive.org> | 2020-02-14 17:55:39 +0000 | 
| commit | c6d7107dc6d8e0925d8a314a64692751725faf78 (patch) | |
| tree | 533f45d7b70eceb5d5d19d37e6a71adf0d5f124b /python/fatcat_tools/harvest/harvest_common.py | |
| parent | 07fabec32aada55a75c064e5c1e01a46da30d854 (diff) | |
| parent | a1bbf612cef473af0410f9985d9e191a4000a0f5 (diff) | |
| download | fatcat-c6d7107dc6d8e0925d8a314a64692751725faf78.tar.gz fatcat-c6d7107dc6d8e0925d8a314a64692751725faf78.zip  | |
Merge branch 'martin-harvest-print' into 'master'
harvest: log state on startup and use stderr for diagnostics
See merge request webgroup/fatcat!27
Diffstat (limited to 'python/fatcat_tools/harvest/harvest_common.py')
| -rw-r--r-- | python/fatcat_tools/harvest/harvest_common.py | 10 | 
1 files changed, 7 insertions, 3 deletions
diff --git a/python/fatcat_tools/harvest/harvest_common.py b/python/fatcat_tools/harvest/harvest_common.py index 78830a1c..310366bd 100644 --- a/python/fatcat_tools/harvest/harvest_common.py +++ b/python/fatcat_tools/harvest/harvest_common.py @@ -57,6 +57,10 @@ class HarvestState:          if catchup_days or start_date or end_date:              self.enqueue_period(start_date, end_date, catchup_days) +    def __str__(self): +        return '<HarvestState to_process={}, completed={}>'.format( +            len(self.to_process), len(self.completed)) +      def enqueue_period(self, start_date=None, end_date=None, catchup_days=14):          """          This function adds a time period to the "TODO" list, unless the dates @@ -129,7 +133,7 @@ class HarvestState:              def fail_fast(err, msg):                  if err:                      raise KafkaException(err) -            print("Commiting status to Kafka: {}".format(kafka_topic)) +            print("Commiting status to Kafka: {}".format(kafka_topic), file=sys.stderr)              producer_conf = kafka_config.copy()              producer_conf.update({                  'delivery.report.only.error': True, @@ -154,7 +158,7 @@ class HarvestState:          if not kafka_topic:              return -        print("Fetching state from kafka topic: {}".format(kafka_topic)) +        print("Fetching state from kafka topic: {}".format(kafka_topic), file=sys.stderr)          def fail_fast(err, msg):              if err:                  raise KafkaException(err) @@ -191,4 +195,4 @@ class HarvestState:          # verify that we got at least to HWM          assert c >= hwm[1] -        print("... got {} state update messages, done".format(c)) +        print("... got {} state update messages, done".format(c), file=sys.stderr)  | 
