def test_database_store():
setup_main_database()
with tempfile.TemporaryDirectory() as tmpdir:
metadata_db = MetadataDB(tmpdir)
metadata_db.create_session_metadata("Test", True,
"http://www.test.com",
"2018-01-01-10-00-00")
# test this store
assert not database.is_store_done("test_source", "2018-01-01-10-00-00",
True)
# start it!
store_id = database.start_store("test_source", "2018-01-01-10-00-00",
True, metadata_db)
# test this store
assert database.get_id_of_store("test_source", "2018-01-01-10-00-00",
True) == store_id
# end it!
database.end_store(store_id)
# test this store
assert database.is_store_done("test_source", "2018-01-01-10-00-00",
True)
# test other stores aren't marked as done
assert not database.is_store_done("test_source", "2018-01-01-10-00-00",
False) # Not a Sample
assert not database.is_store_done("test_source", "2027-01-01-10-00-00",
True) # Different version
assert not database.is_store_done("a_different_source",
"2018-01-01-10-00-00",
True) # Different source
def test_checks_records_error():
setup_main_database()
with tempfile.TemporaryDirectory() as tmpdir:
source = EmptySource(tmpdir)
metadata_db = MetadataDB(tmpdir)
metadata_db.create_session_metadata("Test", True,
"http://www.test.com",
"2018-01-01-10-00-00")
metadata_db.add_filestatus({
'filename': 'test1.json',
'url': 'http://www.test.com',
'data_type': 'record_package'
})
# store details
source_session_id = database.start_store("test_source",
"2018-01-01-10-00-00", True,
metadata_db)
for data in metadata_db.list_filestatus():
with database.add_file(source_session_id, data) as database_file:
database_file.insert_record({'record': 'totally'}, {
'version': '0.1-does-not-exist',
'extensions': []
})
database.end_store(source_session_id)
record_id = 1
# Don't like hard coding ID in. Relies on DB assigning 1 to this new row. But I think we can assume that.
# Test
assert not database.is_record_check_done(record_id)
assert not database.is_check_done(source_session_id)
# check!
for data in metadata_db.list_filestatus():
checks.check_file(source, source_session_id, data)
# Test
assert database.is_record_check_done(record_id)
assert database.is_check_done(source_session_id)
with database.get_engine().begin() as connection:
s = sa.sql.select([database.record_check_error_table])
result = connection.execute(s)
data = result.fetchone()
assert 'The schema version in your data is not valid. Accepted values:' in data[
'error']
def test_checks_releases():
setup_main_database()
with tempfile.TemporaryDirectory() as tmpdir:
source = EmptySource(tmpdir)
metadata_db = MetadataDB(tmpdir)
metadata_db.create_session_metadata("Test", True,
"http://www.test.com",
"2018-01-01-10-00-00")
metadata_db.add_filestatus({
'filename': 'test1.json',
'url': 'http://www.test.com',
'data_type': 'release_package'
})
# store details
source_session_id = database.start_store("test_source",
"2018-01-01-10-00-00", True,
metadata_db)
for data in metadata_db.list_filestatus():
with database.add_file(source_session_id, data) as database_file:
database_file.insert_release({'release': 'totally'},
{'extensions': []})
database.end_store(source_session_id)
release_id = 1
# Don't like hard coding ID in. Relies on DB assigning 1 to this new row. But I think we can assume that.
# Test
assert not database.is_release_check_done(release_id)
# check!
for data in metadata_db.list_filestatus():
checks.check_file(source, source_session_id, data)
# Test
assert database.is_release_check_done(release_id)
with database.get_engine().begin() as connection:
s = sa.sql.select([database.release_check_table])
result = connection.execute(s)
data = result.fetchone()
assert data['cove_output']['file_type'] == 'json'
assert len(data['cove_output']['validation_errors']) > 0
def test_database_store_file():
setup_main_database()
with tempfile.TemporaryDirectory() as tmpdir:
metadata_db = MetadataDB(tmpdir)
metadata_db.create_session_metadata("Test", True,
"http://www.test.com",
"2018-01-01-10-00-00")
metadata_db.add_filestatus({
'filename': 'test1.json',
'url': 'http://www.test.com',
'data_type': 'record_package'
})
# start it!
store_id = database.start_store("test_source", "2018-01-01-10-00-00",
True, metadata_db)
# test
file_id = database.get_id_of_store_file(store_id,
{'filename': 'test1.json'})
assert file_id == 1
class Source:
publisher_name = None
url = None
output_directory = None
source_id = None
sample = False
data_version = None
ignore_release_package_json_lines_missing_releases_error = False
"""It is possible to pass extra arguments.
This specifies a list of the extra arguments possible. Each item in the list should be a dict with the keys:
* name - a name compatible with argparse. Names should be unique across all sources, so include a prefix of some kind.
* help - a help string for argparse
"""
argument_definitions = []
def __init__(self,
base_dir,
remove_dir=False,
publisher_name=None,
url=None,
sample=False,
data_version=None,
new_version=False,
config=None):
self.base_dir = base_dir
self.sample = sample
self.config = config
self.publisher_name = publisher_name or self.publisher_name
if not self.publisher_name:
raise AttributeError('A publisher name needs to be specified')
# Make sure the output directory is fully specified, including sample bit (if applicable)
self.output_directory = self.output_directory or self.source_id
if not self.output_directory:
raise AttributeError('An output directory needs to be specified')
if self.sample and not self.output_directory.endswith('_sample'):
self.output_directory += '_sample'
# Load all versions if possible, pick an existing one or set a new one.
all_versions = sorted(os.listdir(os.path.join(base_dir, self.output_directory)), reverse=True)\
if os.path.exists(os.path.join(base_dir, self.output_directory)) else []
if self.data_version:
pass
elif data_version and data_version in all_versions: # Version specified is valid
self.data_version = data_version
elif data_version: # Version specified is invalid!
raise AttributeError('A version was specified that does not exist')
elif new_version or len(all_versions) == 0: # New Version
self.data_version = datetime.datetime.utcnow().strftime(
'%Y-%m-%d-%H-%M-%S')
elif len(all_versions) > 0: # Get the latest version to resume
self.data_version = all_versions[0]
else: # Should not happen...
raise AttributeError(
'The version is unavailable on the output directory')
# Build full directory, make sure it exists
self.full_directory = os.path.join(base_dir, self.output_directory,
self.data_version)
exists = os.path.exists(self.full_directory)
try:
if exists and remove_dir:
os.rmdir(self.full_directory)
exists = False
if not exists:
os.makedirs(self.full_directory)
except OSError:
raise RuntimeError(
"Error: Write permission is needed on the directory specified (or project dir). %s"
% self.full_directory)
# Misc
self.url = url or self.url
self.metadata_db = MetadataDB(self.full_directory)
self.metadata_db.create_session_metadata(
publisher_name=self.publisher_name,
sample=self.sample,
url=self.url,
data_version=self.data_version)
self.logger = logging.getLogger('ocdskingfisher.source')
"""Returns an array with objects for each url.
The return objects includes url,filename,type and more."""
def gather_all_download_urls(self):
raise NotImplementedError()
def set_arguments(self, arguments):
pass
def run_gather(self):
self.logger.info("Starting run_gather")
metadata = self.metadata_db.get_session()
if metadata['gather_success']:
return
self.metadata_db.update_session_gather_start()
try:
for info in self.gather_all_download_urls():
if self.metadata_db.has_filestatus_filename(info['filename']):
if not self.metadata_db.compare_filestatus_to_database(
info):
raise Exception(
"Tried to add the file " + info['filename'] +
" but it clashed with a file already in the list!")
else:
self.metadata_db.add_filestatus(info)
except Exception as e:
error = repr(e)
stacktrace = traceback.format_exception(*sys.exc_info())
self.metadata_db.update_session_gather_end(False, error,
stacktrace)
return
self.metadata_db.update_session_gather_end(True, None, None)
def is_gather_finished(self):
metadata = self.metadata_db.get_session()
return bool(metadata['gather_finished_datetime'])
def run_fetch(self):
self.logger.info("Starting run_fetch")
metadata = self.metadata_db.get_session()
if metadata['fetch_success']:
return
if not metadata['gather_success']:
msg = 'Can not run fetch without a successful gather!'
if metadata['gather_errors']:
msg += ' Gather errors: ' + metadata['gather_errors']
raise Exception(msg)
self.metadata_db.update_session_fetch_start()
data = self.metadata_db.get_next_filestatus_to_fetch()
while data:
self.logger.info("Starting run_fetch for file " + data['filename'])
self.metadata_db.update_filestatus_fetch_start(data['filename'])
try:
response = self.save_url(
data['filename'], data,
os.path.join(self.full_directory, data['filename']))
if response.additional_files:
for info in response.additional_files:
if self.metadata_db.has_filestatus_filename(
info['filename']):
if not self.metadata_db.compare_filestatus_to_database(
info):
response.errors.append(
"Tried to add the file " +
info['filename'] +
" but it clashed with a file already in the list!"
)
else:
self.metadata_db.add_filestatus(info)
self.metadata_db.update_filestatus_fetch_end(
data['filename'], response.errors, response.warnings)
except Exception as e:
self.metadata_db.update_filestatus_fetch_end(
data['filename'], [repr(e)])
self.push_to_server(data)
data = self.metadata_db.get_next_filestatus_to_fetch()
self.metadata_db.update_session_fetch_end()
def push_to_server(self, data):
if data['data_type'].startswith('meta'):
return
if data['data_type'] == 'release_package_json_lines' or data[
'data_type'] == 'record_package_json_lines':
# BREAK FILE INTO INDIVIDUAL LINES AND SEND ONLY ONE AT A TIME TO THE SERVER!
try:
with open(os.path.join(self.full_directory, data['filename']),
encoding=data['encoding']) as f:
number = 0
raw_data = f.readline()
while raw_data:
print("PUSHING TO SERVER NOW NUMBER " + str(number) +
" FILE " + data['filename'] + " TO " +
self.config.server_url + " KEY " +
self.config.server_api_key)
(tmp_file,
tmp_filename) = tempfile.mkstemp(prefix="ocdskf-")
os.close(tmp_file)
f_write = open(tmp_filename, "w")
f_write.write(raw_data)
f_write.close()
r = requests.post(self.config.server_url +
'/api/v1/submit/item/?API_KEY=' +
self.config.server_api_key,
data={
'collection_source':
self.source_id,
'collection_data_version':
self.data_version,
'collection_sample':
1 if self.sample else 0,
'file_name':
data['filename'],
'file_url':
data['url'],
'file_data_type':
data['data_type'],
'file_encoding':
data['encoding'],
'number':
number,
},
files={
'file': open(tmp_filename, 'rb')
})
print(r.text)
raw_data = f.readline()
number += 1
except Exception as e:
raise e
# TODO Store error in database and make nice HTTP response!
else:
# SEND WHOLE FILE IN ONE GO TO THE SERVER!
print("PUSHING TO SERVER NOW " + data['filename'] + " TO " +
self.config.server_url + " KEY " +
self.config.server_api_key)
r = requests.post(
self.config.server_url + '/api/v1/submit/?API_KEY=' +
self.config.server_api_key,
data={
'collection_source': self.source_id,
'collection_data_version': self.data_version,
'collection_sample': 1 if self.sample else 0,
'file_name': data['filename'],
'file_url': data['url'],
'file_data_type': data['data_type'],
'file_encoding': data['encoding'],
},
files={
'file':
open(os.path.join(self.full_directory, data['filename']),
'rb')
})
print(r.text)
def is_fetch_finished(self):
metadata = self.metadata_db.get_session()
return bool(metadata['fetch_finished_datetime'])
def save_url(self, file_name, data, file_path):
save_content_response = save_content(data['url'], file_path)
return self.SaveUrlResult(errors=save_content_response.errors,
warnings=save_content_response.warnings)
class SaveUrlResult:
def __init__(self, additional_files=[], errors=[], warnings=[]):
self.additional_files = additional_files
self.errors = errors
self.warnings = warnings
"""Called with data to check before checks are run, so any problems can be fixed (See Australia)"""
def before_check_data(self, data, override_schema_version=None):
return data
"""Gather, Fetch, Store and Check data from this publisher."""
def run_all(self):
self.run_gather()
self.run_fetch()
def force_fetch_to_gather(self):
self.logger.info("Starting force_fetch_to_gather")
self.metadata_db.force_fetch_to_gather()
class Source:
publisher_name = None
url = None
output_directory = None
source_id = None
sample = False
data_version = None
"""It is possible to pass extra arguments.
This specifies a list of the extra arguments possible. Each item in the list should be a dict with the keys:
* name - a name compatible with argparse. Names should be unique across all sources, so include a prefix of some kind.
* help - a help string for argparse
"""
argument_definitions = []
def __init__(self,
base_dir,
remove_dir=False,
publisher_name=None,
url=None,
sample=False,
data_version=None,
new_version=False):
self.base_dir = base_dir
self.sample = sample
self.publisher_name = publisher_name or self.publisher_name
if not self.publisher_name:
raise AttributeError('A publisher name needs to be specified')
# Make sure the output directory is fully specified, including sample bit (if applicable)
self.output_directory = self.output_directory or self.source_id
if not self.output_directory:
raise AttributeError('An output directory needs to be specified')
if self.sample and not self.output_directory.endswith('_sample'):
self.output_directory += '_sample'
# Load all versions if possible, pick an existing one or set a new one.
all_versions = sorted(os.listdir(os.path.join(base_dir, self.output_directory)), reverse=True)\
if os.path.exists(os.path.join(base_dir, self.output_directory)) else []
if self.data_version:
pass
elif data_version and data_version in all_versions: # Version specified is valid
self.data_version = data_version
elif data_version: # Version specified is invalid!
raise AttributeError('A version was specified that does not exist')
elif new_version or len(all_versions) == 0: # New Version
self.data_version = datetime.datetime.utcnow().strftime(
'%Y-%m-%d-%H-%M-%S')
elif len(all_versions) > 0: # Get the latest version to resume
self.data_version = all_versions[0]
else: # Should not happen...
raise AttributeError(
'The version is unavailable on the output directory')
# Build full directory, make sure it exists
self.full_directory = os.path.join(base_dir, self.output_directory,
self.data_version)
exists = os.path.exists(self.full_directory)
try:
if exists and remove_dir:
os.rmdir(self.full_directory)
exists = False
if not exists:
os.makedirs(self.full_directory)
except OSError:
raise RuntimeError(
"Error: Write permission is needed on the directory specified (or project dir). %s"
% self.full_directory)
# Misc
self.url = url or self.url
self.metadata_db = MetadataDB(self.full_directory)
self.metadata_db.create_session_metadata(
publisher_name=self.publisher_name,
sample=self.sample,
url=self.url,
data_version=self.data_version)
self.logger = logging.getLogger('ocdskingfisher.source')
"""Returns an array with objects for each url.
The return objects includes url,filename,type and more."""
def gather_all_download_urls(self):
raise NotImplementedError()
def set_arguments(self, arguments):
pass
def run_gather(self):
self.logger.info("Starting run_gather")
metadata = self.metadata_db.get_session()
if metadata['gather_success']:
return
self.metadata_db.update_session_gather_start()
try:
for info in self.gather_all_download_urls():
if self.metadata_db.has_filestatus_filename(info['filename']):
if not self.metadata_db.compare_filestatus_to_database(
info):
raise Exception(
"Tried to add the file " + info['filename'] +
" but it clashed with a file already in the list!")
else:
self.metadata_db.add_filestatus(info)
except Exception as e:
error = repr(e)
stacktrace = traceback.format_exception(*sys.exc_info())
self.metadata_db.update_session_gather_end(False, error,
stacktrace)
return
self.metadata_db.update_session_gather_end(True, None, None)
def is_gather_finished(self):
metadata = self.metadata_db.get_session()
return bool(metadata['gather_finished_datetime'])
def run_fetch(self):
self.logger.info("Starting run_fetch")
metadata = self.metadata_db.get_session()
if metadata['fetch_success']:
return
if not metadata['gather_success']:
msg = 'Can not run fetch without a successful gather!'
if metadata['gather_errors']:
msg += ' Gather errors: ' + metadata['gather_errors']
raise Exception(msg)
self.metadata_db.update_session_fetch_start()
data = self.metadata_db.get_next_filestatus_to_fetch()
while data:
self.logger.info("Starting run_fetch for file " + data['filename'])
self.metadata_db.update_filestatus_fetch_start(data['filename'])
try:
response = self.save_url(
data['filename'], data,
os.path.join(self.full_directory, data['filename']))
if response.additional_files:
for info in response.additional_files:
if self.metadata_db.has_filestatus_filename(
info['filename']):
if not self.metadata_db.compare_filestatus_to_database(
info):
response.errors.append(
"Tried to add the file " +
info['filename'] +
" but it clashed with a file already in the list!"
)
else:
self.metadata_db.add_filestatus(info)
self.metadata_db.update_filestatus_fetch_end(
data['filename'], response.errors, response.warnings)
except Exception as e:
self.metadata_db.update_filestatus_fetch_end(
data['filename'], [repr(e)])
data = self.metadata_db.get_next_filestatus_to_fetch()
self.metadata_db.update_session_fetch_end()
def is_fetch_finished(self):
metadata = self.metadata_db.get_session()
return bool(metadata['fetch_finished_datetime'])
"""Uploads the fetched data as record rows to the Database"""
def run_store(self):
self.logger.info("Starting run_store")
metadata = self.metadata_db.get_session()
if not metadata['fetch_success']:
raise Exception('Can not run store without a successful fetch')
if database.is_store_done(self.source_id, self.data_version,
self.sample):
return
source_session_id = database.start_store(self.source_id,
self.data_version,
self.sample, self.metadata_db)
for data in self.metadata_db.list_filestatus():
if data['data_type'].startswith('meta'):
continue
if database.is_store_file_done(source_session_id, data):
continue
with database.add_file(source_session_id, data) as database_file:
self.logger.info("Starting run_store for file " +
data['filename'])
try:
with open(os.path.join(self.full_directory,
data['filename']),
encoding=data['encoding']) as f:
file_json_data = json.load(f)
except Exception as e:
# TODO better way of dealing with this?
raise e
return
objects_list = []
if data['data_type'] == 'record_package_list_in_results':
objects_list.extend(file_json_data['results'])
elif data['data_type'] == 'release_package_list_in_results':
objects_list.extend(file_json_data['results'])
elif data['data_type'] == 'record_package_list' or data[
'data_type'] == 'release_package_list':
objects_list.extend(file_json_data)
else:
objects_list.append(file_json_data)
del file_json_data
for json_data in objects_list:
error_msg = ''
if not isinstance(json_data, dict):
error_msg = "Can not process data in file {} as JSON is not an object".format(
data['filename'])
if data['data_type'] == 'release':
data_list = [json_data]
elif data['data_type'] == 'release_package' or \
data['data_type'] == 'release_package_list_in_results' or \
data['data_type'] == 'release_package_list':
if 'releases' not in json_data:
error_msg = "Release list not found in file {}".format(
data['filename'])
elif not isinstance(json_data['releases'], list):
error_msg = "Release list which is not a list found in file {}".format(
data['filename'])
data_list = json_data['releases']
elif data['data_type'] == 'record_package' or \
data['data_type'] == 'record_package_list_in_results' or \
data['data_type'] == 'record_package_list':
if 'records' not in json_data:
error_msg = "Record list not found in file {}".format(
data['filename'])
elif not isinstance(json_data['records'], list):
error_msg = "Record list which is not a list found in file {}".format(
data['filename'])
data_list = json_data['records']
else:
error_msg = "data_type not a known type"
if error_msg:
raise Exception(error_msg)
package_data = {}
if not data['data_type'] == 'release':
for key, value in json_data.items():
if key not in ('releases', 'records'):
package_data[key] = value
for row in data_list:
if not isinstance(row, dict):
error_msg = "Row in data is not a object {}".format(
data['filename'])
raise Exception(error_msg)
if data['data_type'] == 'record_package' or \
data['data_type'] == 'record_package_list_in_results' or \
data['data_type'] == 'record_package_list':
database_file.insert_record(row, package_data)
else:
database_file.insert_release(row, package_data)
database.end_store(source_session_id)
def save_url(self, file_name, data, file_path):
save_content_response = save_content(data['url'], file_path)
return self.SaveUrlResult(errors=save_content_response.errors,
warnings=save_content_response.warnings)
class SaveUrlResult:
def __init__(self, additional_files=[], errors=[], warnings=[]):
self.additional_files = additional_files
self.errors = errors
self.warnings = warnings
def run_check(self, override_schema_version=None):
self.logger.info("Starting run_check")
if not database.is_store_done(self.source_id, self.data_version,
self.sample):
raise Exception('Can not run check without a successful store')
source_session_id = database.get_id_of_store(self.source_id,
self.data_version,
self.sample)
for data in self.metadata_db.list_filestatus():
if data['data_type'].startswith('meta'):
continue
self.logger.info("Starting run_check for file " + data['filename'])
check_file(self, source_session_id, data, override_schema_version)
"""Called with data to check before checks are run, so any problems can be fixed (See Australia)"""
def before_check_data(self, data, override_schema_version=None):
return data
"""Gather, Fetch, Store and Check data from this publisher."""
def run_all(self):
self.run_gather()
self.run_fetch()
self.run_store()
self.run_check()
def force_fetch_to_gather(self):
self.logger.info("Starting force_fetch_to_gather")
self.metadata_db.force_fetch_to_gather()
def test_metadatabase_store():
setup_main_database()
with tempfile.TemporaryDirectory() as tmpdir:
metadata_db = MetadataDB(tmpdir)
metadata_db.create_session_metadata("Test", True,
"http://www.test.com",
"2018-01-01-10-00-00")
# No files
assert not metadata_db.has_filestatus_filename('file1.json')
# Add files
metadata_db.add_filestatus({
'url':
'https://raw.githubusercontent.com/open-contracting/sample-data/' +
'5bcbfcf48bf6e6599194b8acae61e2c6e8fb5009/fictional-example/1.1/ocds-213czf-000-00001-02-tender.json',
'filename':
'file1.json',
'data_type':
'releases'
})
# Check file
assert metadata_db.has_filestatus_filename('file1.json')
assert metadata_db.compare_filestatus_to_database({
'url':
'https://raw.githubusercontent.com/open-contracting/sample-data/' +
'5bcbfcf48bf6e6599194b8acae61e2c6e8fb5009/fictional-example/1.1/ocds-213czf-000-00001-02-tender.json',
'filename':
'file1.json',
'data_type':
'releases'
})
assert metadata_db.compare_filestatus_to_database({
'url':
'https://raw.githubusercontent.com/open-contracting/sample-data/' +
'5bcbfcf48bf6e6599194b8acae61e2c6e8fb5009/fictional-example/1.1/ocds-213czf-000-00001-02-tender.json',
'filename':
'file1.json',
'data_type':
'releases',
'priority':
1
})
assert metadata_db.compare_filestatus_to_database({
'url':
'https://raw.githubusercontent.com/open-contracting/sample-data/' +
'5bcbfcf48bf6e6599194b8acae61e2c6e8fb5009/fictional-example/1.1/ocds-213czf-000-00001-02-tender.json',
'filename':
'file1.json',
'data_type':
'releases',
'encoding':
'utf-8'
})
# .... different data_type
assert not metadata_db.compare_filestatus_to_database({
'url':
'https://raw.githubusercontent.com/open-contracting/sample-data/' +
'5bcbfcf48bf6e6599194b8acae61e2c6e8fb5009/fictional-example/1.1/ocds-213czf-000-00001-02-tender.json',
'filename':
'file1.json',
'data_type':
'releases-package'
})
# .... different url
assert not metadata_db.compare_filestatus_to_database(
{
'url': 'https://www.google.com/',
'filename': 'file1.json',
'data_type': 'releases'
})
# .... different priority
assert not metadata_db.compare_filestatus_to_database({
'url':
'https://raw.githubusercontent.com/open-contracting/sample-data/' +
'5bcbfcf48bf6e6599194b8acae61e2c6e8fb5009/fictional-example/1.1/ocds-213czf-000-00001-02-tender.json',
'filename':
'file1.json',
'data_type':
'releases',
'priority':
10
})
# .... different encoding
assert not metadata_db.compare_filestatus_to_database({
'url':
'https://raw.githubusercontent.com/open-contracting/sample-data/' +
'5bcbfcf48bf6e6599194b8acae61e2c6e8fb5009/fictional-example/1.1/ocds-213czf-000-00001-02-tender.json',
'filename':
'file1.json',
'data_type':
'releases',
'encoding':
'ascii'
})
class Source:
publisher_name = None
url = None
output_directory = None
source_id = None
sample = False
data_version = None
ignore_release_package_json_lines_missing_releases_error = False
"""It is possible to pass extra arguments.
This specifies a list of the extra arguments possible. Each item in the list should be a dict with the keys:
* name - a name compatible with argparse. Names should be unique across all sources, so include a prefix of some kind.
* help - a help string for argparse
"""
argument_definitions = []
def __init__(self, base_dir, remove_dir=False, publisher_name=None, url=None, sample=False, data_version=None, new_version=False, config=None, note=None):
self.base_dir = base_dir
self.sample = sample
self.config = config
self.note = note
self.publisher_name = publisher_name or self.publisher_name
if not self.publisher_name:
raise AttributeError('A publisher name needs to be specified')
# Make sure the output directory is fully specified, including sample bit (if applicable)
self.output_directory = self.output_directory or self.source_id
if not self.output_directory:
raise AttributeError('An output directory needs to be specified')
if self.sample and not self.output_directory.endswith('_sample'):
self.output_directory += '_sample'
# Load all versions if possible, pick an existing one or set a new one.
all_versions = sorted(os.listdir(os.path.join(base_dir, self.output_directory)), reverse=True)\
if os.path.exists(os.path.join(base_dir, self.output_directory)) else []
if self.data_version:
pass
elif data_version and data_version in all_versions: # Version specified is valid
self.data_version = data_version
elif data_version: # Version specified is invalid!
raise AttributeError('A version was specified that does not exist')
elif new_version or len(all_versions) == 0: # New Version
self.data_version = datetime.datetime.utcnow().strftime('%Y-%m-%d-%H-%M-%S')
elif len(all_versions) > 0: # Get the latest version to resume
self.data_version = all_versions[0]
else: # Should not happen...
raise AttributeError('The version is unavailable on the output directory')
# Build full directory, make sure it exists
self.full_directory = os.path.join(base_dir, self.output_directory, self.data_version)
exists = os.path.exists(self.full_directory)
try:
if exists and remove_dir:
os.rmdir(self.full_directory)
exists = False
if not exists:
os.makedirs(self.full_directory)
except OSError:
raise RuntimeError("Error: Write permission is needed on the directory specified (or project dir). %s" % self.full_directory)
# Misc
self.url = url or self.url
self.metadata_db = MetadataDB(self.full_directory)
self.metadata_db.create_session_metadata(
publisher_name=self.publisher_name,
sample=self.sample,
url=self.url,
data_version=self.data_version
)
self.logger = logging.getLogger('ocdskingfisher.source')
"""Returns an array with objects for each url.
The return objects includes url,filename,type and more."""
def gather_all_download_urls(self):
raise NotImplementedError()
def set_arguments(self, arguments):
pass
def run_gather(self):
self.logger.info("Starting run_gather")
metadata = self.metadata_db.get_session()
if metadata['gather_success']:
return
self.metadata_db.update_session_gather_start()
try:
for info in self.gather_all_download_urls():
if self.metadata_db.has_filestatus_filename(info['filename']):
if not self.metadata_db.compare_filestatus_to_database(info):
raise Exception("Tried to add the file " + info['filename'] +
" but it clashed with a file already in the list!")
else:
self.metadata_db.add_filestatus(info)
except Exception as e:
error = repr(e)
stacktrace = traceback.format_exception(*sys.exc_info())
self.metadata_db.update_session_gather_end(False, error, stacktrace)
return
self.metadata_db.update_session_gather_end(True, None, None)
def is_gather_finished(self):
metadata = self.metadata_db.get_session()
return bool(metadata['gather_finished_datetime'])
def run_fetch(self):
self.logger.info("Starting run_fetch")
metadata = self.metadata_db.get_session()
if metadata['fetch_success']:
return
if not metadata['gather_success']:
msg = 'Can not run fetch without a successful gather!'
if metadata['gather_errors']:
msg += ' Gather errors: ' + metadata['gather_errors']
raise Exception(msg)
self.metadata_db.update_session_fetch_start()
data = self.metadata_db.get_next_filestatus_to_fetch()
while data:
self.logger.info("Starting run_fetch for file " + data['filename'])
self.metadata_db.update_filestatus_fetch_start(data['filename'])
try:
response = self.save_url(data['filename'], data, os.path.join(self.full_directory, data['filename']))
if response.additional_files:
for info in response.additional_files:
if self.metadata_db.has_filestatus_filename(info['filename']):
if not self.metadata_db.compare_filestatus_to_database(info):
response.errors.append("Tried to add the file " + info['filename'] +
" but it clashed with a file already in the list!")
else:
self.metadata_db.add_filestatus(info)
self.metadata_db.update_filestatus_fetch_end(data['filename'], response.errors, response.warnings)
self._send_file_to_server(data, response.errors, response.warnings)
except Exception as e:
self.metadata_db.update_filestatus_fetch_end(data['filename'], [repr(e)])
self._send_file_to_server(data, [repr(e)])
data = self.metadata_db.get_next_filestatus_to_fetch()
self.metadata_db.update_session_fetch_end()
self._send_fetch_end_to_server()
def _send_file_to_server(self, data, errors, warnings=None):
if not self.config.server_api_key or not self.config.server_url:
print("SERVER NOT CONFIGURED")
return
if data['data_type'].startswith('meta'):
return
if errors:
self._send_file_errors_to_server(data, errors, warnings)
elif data['data_type'] == 'release_package_json_lines' or data['data_type'] == 'record_package_json_lines':
self._send_json_lines_file_success_to_server(data, warnings)
else:
self._send_file_success_to_server(data, warnings)
def _send_file_errors_to_server(self, data, errors, warnings):
print("PUSHING ERRORS TO SERVER NOW " + data[
'filename'] + " TO " + self.config.server_url)
post_data = {
'collection_source': self.source_id,
'collection_data_version': self.data_version,
'collection_sample': self.sample,
'file_name': data['filename'],
'url': data['url'],
'errors': json.dumps(errors),
}
response = requests.post(self.config.server_url + '/api/v1/submit/file_errors/',
data=post_data,
headers={'Authorization': 'ApiKey ' + self.config.server_api_key})
if not response.ok:
raise Exception("COULD NOT SEND ERRORS TO SERVER! HTTP CODE: " + str(response.status_code))
def _send_json_lines_file_success_to_server(self, data, warnings):
with open(os.path.join(self.full_directory, data['filename']), encoding=data['encoding']) as f:
number = 0
raw_data = f.readline()
while raw_data:
print("PUSHING TO SERVER NOW NUMBER " + str(number) + " FILE " + data[
'filename'] + " TO " + self.config.server_url)
post_data = {
'collection_source': self.source_id,
'collection_data_version': self.data_version,
'collection_sample': self.sample,
'file_name': data['filename'],
'url': data['url'],
'data_type': data['data_type'],
'number': number,
'data': raw_data,
'collection_note': self.note,
}
response = requests.post(self.config.server_url + '/api/v1/submit/item/',
data=post_data,
headers={'Authorization': 'ApiKey ' + self.config.server_api_key})
if not response.ok:
raise Exception("COULD NOT SEND TO SERVER! HTTP CODE: " + str(response.status_code))
raw_data = f.readline()
number += 1
def _send_file_success_to_server(self, data, warnings):
print("PUSHING TO SERVER NOW " + data[
'filename'] + " TO " + self.config.server_url)
post_data = {
'collection_source': self.source_id,
'collection_data_version': self.data_version,
'collection_sample': self.sample,
'file_name': data['filename'],
'url': data['url'],
'data_type': data['data_type'],
'encoding': data['encoding'],
'collection_note': self.note,
}
files = {
'file': ( data['filename'], open(os.path.join(self.full_directory, data['filename']), "rb"), 'application/json')
}
response = requests.post(self.config.server_url + '/api/v1/submit/file/',
data=post_data,
files=files,
headers={'Authorization': 'ApiKey ' + self.config.server_api_key})
if not response.ok:
raise Exception("COULD NOT SEND TO SERVER! HTTP CODE: " + str(response.status_code))
def _send_fetch_end_to_server(self):
print("PUSHING COLLECTION END TO SERVER NOW TO " + self.config.server_url)
post_data = {
'collection_source': self.source_id,
'collection_data_version': self.data_version,
'collection_sample': self.sample,
}
response = requests.post(self.config.server_url + '/api/v1/submit/end_collection_store/',
data=post_data,
headers={'Authorization': 'ApiKey ' + self.config.server_api_key})
if not response.ok:
raise Exception("COULD NOT SEND END TO SERVER! HTTP CODE: " + str(response.status_code))
def is_fetch_finished(self):
metadata = self.metadata_db.get_session()
return bool(metadata['fetch_finished_datetime'])
def save_url(self, file_name, data, file_path):
save_content_response = save_content(data['url'], file_path)
return self.SaveUrlResult(errors=save_content_response.errors, warnings=save_content_response.warnings)
class SaveUrlResult:
def __init__(self, additional_files=[], errors=[], warnings=[]):
self.additional_files = additional_files
self.errors = errors
self.warnings = warnings
"""Gather, Fetch, Store and Check data from this publisher."""
def run_all(self):
self.run_gather()
self.run_fetch()
def force_fetch_to_gather(self):
self.logger.info("Starting force_fetch_to_gather")
self.metadata_db.force_fetch_to_gather()