def import_data(self, data_items):
""" import a bunch of dicts together """
# keep counts of all actions
record = {
'insert': 0, 'update': 0, 'noop': 0,
'start': utcnow(),
'records': {
'insert': [],
'update': [],
'noop': [],
}
}
for json_id, data in self._prepare_imports(data_items):
obj_id, what = self.import_item(data)
self.json_to_db_id[json_id] = obj_id
record['records'][what].append(obj_id)
record[what] += 1
# all objects are loaded, a perfect time to do inter-object resolution and other tasks
self.postimport()
record['end'] = utcnow()
return {self._type: record}
def import_data(self, data_items):
""" import a bunch of dicts together """
# keep counts of all actions
record = {
'insert': 0, 'update': 0, 'noop': 0,
'start': utcnow(),
'records': {
'insert': [],
'update': [],
'noop': [],
}
}
for json_id, data in self._prepare_imports(data_items):
obj_id, what = self.import_item(data)
self.json_to_db_id[json_id] = obj_id
record['records'][what].append(obj_id)
record[what] += 1
# all objects are loaded, a perfect time to do inter-object resolution and other tasks
self.postimport()
record['end'] = utcnow()
return {self._type: record}
def do_handle(self, args, other, juris):
available_scrapers = getattr(juris, 'scrapers', {})
scrapers = OrderedDict()
if not available_scrapers:
raise CommandError('no scrapers defined on jurisdiction')
if other:
# parse arg list in format: (scraper (k:v)+)+
cur_scraper = None
for arg in other:
if '=' in arg:
if not cur_scraper:
raise CommandError('argument {} before scraper name'.format(arg))
k, v = arg.split('=', 1)
scrapers[cur_scraper][k] = v
elif arg in juris.scrapers:
cur_scraper = arg
scrapers[cur_scraper] = {}
else:
raise CommandError('no such scraper: module={} scraper={}'.format(args.module,
arg))
else:
scrapers = {key: {} for key in available_scrapers.keys()}
# modify args in-place so we can pass them around
if not args.actions:
args.actions = ALL_ACTIONS
if 'import' in args.actions:
django.setup()
# print the plan
report = {'plan': {'module': args.module, 'actions': args.actions, 'scrapers': scrapers},
'start': utils.utcnow(),
}
print_report(report)
self.check_session_list(juris)
try:
if 'scrape' in args.actions:
report['scrape'] = self.do_scrape(juris, args, scrapers)
if 'import' in args.actions:
report['import'] = self.do_import(juris, args)
report['success'] = True
except Exception as exc:
report['success'] = False
report['exception'] = exc
report['traceback'] = traceback.format_exc()
if 'import' in args.actions:
save_report(report, juris.jurisdiction_id)
raise
if 'import' in args.actions:
save_report(report, juris.jurisdiction_id)
print_report(report)
return report
def save_report(report, jurisdiction):
from pupa.models import RunPlan
# set end time
report['end'] = utils.utcnow()
plan = RunPlan.objects.create(jurisdiction_id=jurisdiction,
success=report['success'],
start_time=report['start'],
end_time=report['end'],
exception=report.get('exception', ''),
traceback=report.get('traceback', ''),
)
for scraper, details in report.get('scrape', {}).items():
args = ' '.join('{k}={v}'.format(k=k, v=v)
for k, v in report['plan']['scrapers'].get(scraper, {}).items())
sr = plan.scrapers.create(scraper=scraper, args=args,
start_time=details['start'], end_time=details['end'])
for object_type, num in details['objects'].items():
sr.scraped_objects.create(object_type=object_type, count=num)
for object_type, changes in report.get('import', {}).items():
if changes['insert'] or changes['update'] or changes['noop']:
plan.imported_objects.create(
object_type=object_type,
insert_count=changes['insert'],
update_count=changes['update'],
noop_count=changes['noop'],
start_time=changes['start'],
end_time=changes['end'],
)
def handle(self, args, other):
juris = self.get_jurisdiction(args.module)
available_scrapers = getattr(juris, 'scrapers', {})
scrapers = OrderedDict()
if not available_scrapers:
raise CommandError('no scrapers defined on jurisdiction')
if other:
# parse arg list in format: (scraper (k:v)+)+
cur_scraper = None
for arg in other:
if '=' in arg:
if not cur_scraper:
raise CommandError('argument {} before scraper name'.format(arg))
k, v = arg.split('=', 1)
scrapers[cur_scraper][k] = v
elif arg in juris.scrapers:
cur_scraper = arg
scrapers[cur_scraper] = {}
else:
raise CommandError('no such scraper: module={} scraper={}'.format(args.module,
arg))
else:
scrapers = {key: {} for key in available_scrapers.keys()}
# modify args in-place so we can pass them around
if not args.actions:
args.actions = ALL_ACTIONS
# print the plan
report = {'plan': {'module': args.module, 'actions': args.actions, 'scrapers': scrapers},
'start': utils.utcnow(),
}
print_report(report)
self.check_session_list(juris)
try:
if 'scrape' in args.actions:
report['scrape'] = self.do_scrape(juris, args, scrapers)
if 'import' in args.actions:
report['import'] = self.do_import(juris, args)
report['success'] = True
except Exception as exc:
report['success'] = False
report['exception'] = exc
report['traceback'] = traceback.format_exc()
if 'import' in args.actions:
save_report(report, juris.jurisdiction_id)
raise
if 'import' in args.actions:
save_report(report, juris.jurisdiction_id)
forward_report(report, juris.jurisdiction_id)
print_report(report)
return report
def do_scrape(self, **kwargs):
record = {'objects': defaultdict(int)}
self.output_names = defaultdict(set)
record['start'] = utils.utcnow()
for obj in self.scrape(**kwargs) or []:
if hasattr(obj, '__iter__'):
for iterobj in obj:
self.save_object(iterobj)
else:
self.save_object(obj)
record['end'] = utils.utcnow()
record['skipped'] = getattr(self, 'skipped', 0)
if not self.output_names:
raise ScrapeError('no objects returned from scrape')
for _type, nameset in self.output_names.items():
record['objects'][_type] += len(nameset)
return record
def do_scrape(self, **kwargs):
record = {'objects': defaultdict(int)}
self.output_names = defaultdict(set)
record['start'] = utils.utcnow()
for obj in self.scrape(**kwargs) or []:
if hasattr(obj, '__iter__'):
for iterobj in obj:
self.save_object(iterobj)
else:
self.save_object(obj)
record['end'] = utils.utcnow()
record['skipped'] = getattr(self, 'skipped', 0)
if not self.output_names:
raise ScrapeError('no objects returned from {} scrape'.format(self.__class__.__name__))
for _type, nameset in self.output_names.items():
record['objects'][_type] += len(nameset)
return record
def save_report(report, jurisdiction):
from pupa.models import RunPlan
from opencivicdata.core.models import Jurisdiction as JurisdictionModel
# set end time
report['end'] = utils.utcnow()
# if there's an error on the first run, the jurisdiction doesn't exist
# yet, we opt for skipping creation of RunPlan until there's been at least
# one good run
try:
JurisdictionModel.objects.get(pk=jurisdiction)
except JurisdictionModel.DoesNotExist:
logger = logging.getLogger("pupa")
logger.warning(
'could not save RunPlan, no successful runs of {} yet'.format(
jurisdiction))
return
plan = RunPlan.objects.create(
jurisdiction_id=jurisdiction,
success=report['success'],
start_time=report['start'],
end_time=report['end'],
exception=report.get('exception', ''),
traceback=report.get('traceback', ''),
)
for scraper, details in report.get('scrape', {}).items():
args = ' '.join(
'{k}={v}'.format(k=k, v=v)
for k, v in report['plan']['scrapers'].get(scraper, {}).items())
sr = plan.scrapers.create(scraper=scraper,
args=args,
start_time=details['start'],
end_time=details['end'])
for object_type, num in details['objects'].items():
sr.scraped_objects.create(object_type=object_type, count=num)
for object_type, changes in report.get('import', {}).items():
if changes['insert'] or changes['update'] or changes['noop']:
plan.imported_objects.create(
object_type=object_type,
insert_count=changes['insert'],
update_count=changes['update'],
noop_count=changes['noop'],
start_time=changes['start'],
end_time=changes['end'],
)
def save_report(report, jurisdiction):
from pupa.models import RunPlan
from opencivicdata.core.models import Jurisdiction as JurisdictionModel
# set end time
report['end'] = utils.utcnow()
# if there's an error on the first run, the jurisdiction doesn't exist
# yet, we opt for skipping creation of RunPlan until there's been at least
# one good run
try:
JurisdictionModel.objects.get(pk=jurisdiction)
except JurisdictionModel.DoesNotExist:
logger = logging.getLogger("pupa")
logger.warning('could not save RunPlan, no successful runs of {} yet'.format(
jurisdiction)
)
return
plan = RunPlan.objects.create(jurisdiction_id=jurisdiction,
success=report['success'],
start_time=report['start'],
end_time=report['end'],
exception=report.get('exception', ''),
traceback=report.get('traceback', ''),
)
for scraper, details in report.get('scrape', {}).items():
args = ' '.join('{k}={v}'.format(k=k, v=v)
for k, v in report['plan']['scrapers'].get(scraper, {}).items())
sr = plan.scrapers.create(scraper=scraper, args=args,
start_time=details['start'], end_time=details['end'])
for object_type, num in details['objects'].items():
sr.scraped_objects.create(object_type=object_type, count=num)
for object_type, changes in report.get('import', {}).items():
if changes['insert'] or changes['update'] or changes['noop']:
plan.imported_objects.create(
object_type=object_type,
insert_count=changes['insert'],
update_count=changes['update'],
noop_count=changes['noop'],
start_time=changes['start'],
end_time=changes['end'],
)
def save_report(report, jurisdiction):
from pupa.models import RunPlan
# set end time
report['end'] = utils.utcnow()
plan = RunPlan.objects.create(
jurisdiction_id=jurisdiction,
success=report['success'],
start_time=report['start'],
end_time=report['end'],
exception=report.get('exception', ''),
traceback=report.get('traceback', ''),
)
for scraper, details in report.get('scrape', {}).items():
args = ' '.join(
'{k}={v}'.format(k=k, v=v)
for k, v in report['plan']['scrapers'].get(scraper, {}).items())
sr = plan.scrapers.create(scraper=scraper,
args=args,
start_time=details['start'],
end_time=details['end'])
for object_type, num in details['objects'].items():
sr.scraped_objects.create(object_type=object_type, count=num)
for object_type, changes in report.get('import', {}).items():
if changes['insert'] or changes['update'] or changes['noop']:
plan.imported_objects.create(
object_type=object_type,
insert_count=changes['insert'],
update_count=changes['update'],
noop_count=changes['noop'],
start_time=changes['start'],
end_time=changes['end'],
)
