def cross_db(docs,
source_db=None,
target_db=None,
index=None,
versions=False,
**kwargs):
"""
Migration to go between
cassandra > postgres
postgres > cassandra
cassandra > elasticsearch
postgres > elasticsearch
source db can be passed in to the migrate task, and will default to the CANONICAL_PROCESSOR specified in settings
target_db will be specified when the task is called
"""
assert target_db, 'Please specify a target db for the migration -- either postgres or elasticsearch'
assert target_db in [
'postgres', 'cassandra', 'elasticsearch'
], 'Invalid target database - please specify either postgres, cassandra or elasticsearch'
source_processor = get_processor(source_db or settings.CANONICAL_PROCESSOR)
target_processor = get_processor(target_db)
for doc in docs:
try:
if not doc.raw['doc']:
# corrupted database item has no doc element
message = 'No doc element in raw doc -- could not migrate document from {} with id {}'.format(
doc.raw.attributes['source'], doc.raw.attributes['docID'])
log_to_sentry(message)
logger.info(message)
continue
raw, normalized = doc.raw, doc.normalized
if not kwargs.get('dry'):
if versions:
for raw_version, norm_version in source_processor.get_versions(
raw['source'], raw['docID']):
target_processor.process_raw(raw_version)
if norm_version:
target_processor.process_normalized(
raw_version, norm_version)
else:
logger.info(
'Not storing migrated normalized version from {} with id {}, document is not in approved set list.'
.format(raw.attributes['source'],
raw.attributes['docID']))
else:
target_processor.process_raw(raw)
if normalized:
target_processor.process_normalized(raw, normalized)
else:
logger.info(
'Not storing migrated normalized from {} with id {}, document is not in approved set list.'
.format(raw.attributes['source'],
raw.attributes['docID']))
except Exception as e:
logger.exception(e)
log_to_sentry(e)
def record_or_load_response(method,
url,
throttle=None,
force=False,
params=None,
expected=(200, ),
**kwargs):
resp = _maybe_load_response(method, url)
if not force and resp and resp.ok:
logger.info('Return recorded response from "{}"'.format(url))
return resp
if force:
logger.warning('Force updating request to "{}"'.format(url))
else:
logger.info('Making request to "{}"'.format(url))
maybe_sleep(throttle)
response = requests.request(method, url, **kwargs)
if not response.ok:
events.log_to_sentry('Got non-ok response code.',
url=url,
method=method)
if isinstance(response.content, six.text_type):
response.content = response.content.encode('utf8')
if not resp:
return HarvesterResponse(url=url.lower(),
method=method,
ok=response.ok,
content=response.content,
encoding=response.encoding,
status_code=response.status_code,
headers_str=json.dumps(dict(
response.headers))).save()
logger.warning('Skipped recorded response from "{}"'.format(url))
return resp.update(ok=(response.ok or response.status_code in expected),
content=response.content,
encoding=response.encoding,
status_code=response.status_code,
headers_str=json.dumps(dict(response.headers))).save()
def record_or_load_response(method, url, throttle=None, force=False, params=None, **kwargs):
if params:
url = furl.furl(url).set(args=params).url
resp = _maybe_load_response(method, url)
if not force and resp and resp.ok:
logger.info('Return recorded response from "{}"'.format(url))
return resp
if force:
logger.warning('Force updating request to "{}"'.format(url))
else:
logger.info('Making request to "{}"'.format(url))
if throttle:
time.sleep(throttle)
response = requests.request(method, url, **kwargs)
if not response.ok:
events.log_to_sentry('Got non-ok response code.', url=url, method=method)
if not resp:
return HarvesterResponse(
url=url,
method=method,
ok=response.ok,
content=response.content,
encoding=response.encoding,
status_code=response.status_code,
headers_str=json.dumps(dict(response.headers))
).save()
logger.warning('Skipped recorded response from "{}"'.format(url))
return resp.update(
ok=response.ok,
content=response.content,
encoding=response.encoding,
status_code=response.status_code,
headers_str=json.dumps(dict(response.headers))
).save()
def cross_db(docs, source_db=None, target_db=None, index=None, versions=False, **kwargs):
"""
Migration to go between
cassandra > postgres
postgres > cassandra
cassandra > elasticsearch
postgres > elasticsearch
source db can be passed in to the migrate task, and will default to the CANONICAL_PROCESSOR specified in settings
target_db will be specified when the task is called
"""
assert target_db, 'Please specify a target db for the migration -- either postgres or elasticsearch'
assert target_db in ['postgres', 'cassandra', 'elasticsearch'], 'Invalid target database - please specify either postgres, cassandra or elasticsearch'
source_processor = get_processor(source_db or settings.CANONICAL_PROCESSOR)
target_processor = get_processor(target_db)
for doc in docs:
try:
if not doc.raw['doc']:
# corrupted database item has no doc element
message = 'No doc element in raw doc -- could not migrate document from {} with id {}'.format(doc.raw.attributes['source'], doc.raw.attributes['docID'])
log_to_sentry(message)
logger.info(message)
continue
raw, normalized = doc.raw, doc.normalized
if not kwargs.get('dry'):
if versions:
for raw_version, norm_version in source_processor.get_versions(raw['source'], raw['docID']):
target_processor.process_raw(raw_version)
if norm_version:
target_processor.process_normalized(raw_version, norm_version)
else:
logger.info('Not storing migrated normalized version from {} with id {}, document is not in approved set list.'.format(raw.attributes['source'], raw.attributes['docID']))
else:
target_processor.process_raw(raw)
if normalized:
target_processor.process_normalized(raw, normalized)
else:
logger.info('Not storing migrated normalized from {} with id {}, document is not in approved set list.'.format(raw.attributes['source'], raw.attributes['docID']))
except Exception as e:
logger.exception(e)
log_to_sentry(e)
def record_or_load_response(method, url, throttle=None, force=False, params=None, expected=(200,), **kwargs):
resp = _maybe_load_response(method, url)
if not force and resp and resp.ok:
logger.info('Return recorded response from "{}"'.format(url))
return resp
if force:
logger.warning('Force updating request to "{}"'.format(url))
else:
logger.info('Making request to "{}"'.format(url))
maybe_sleep(throttle)
response = requests.request(method, url, **kwargs)
if not response.ok:
events.log_to_sentry('Got non-ok response code.', url=url, method=method)
if isinstance(response.content, six.text_type):
response.content = response.content.encode('utf8')
if not resp:
return HarvesterResponse(
url=url.lower(),
method=method,
ok=response.ok,
content=response.content,
encoding=response.encoding,
status_code=response.status_code,
headers_str=json.dumps(dict(response.headers))
).save()
logger.warning('Skipped recorded response from "{}"'.format(url))
return resp.update(
ok=(response.ok or response.status_code in expected),
content=response.content,
encoding=response.encoding,
status_code=response.status_code,
headers_str=json.dumps(dict(response.headers))
).save()
