def handle(self, *args, **options):
if options['user']:
if options['user'] == 'anonymous':
owners = [None]
else:
owners = [User.objects.get(username=options['user'])]
else:
owners = [None] + list(User.objects.all())
adapter = DatabaseAdapter()
stale_jobs = []
for owner in owners:
jobs = QueryJob.objects.filter(owner=owner)
for job in jobs:
if job.phase == job.PHASE_COMPLETED:
if not adapter.fetch_table(job.schema_name, job.table_name):
stale_jobs.append(job)
if stale_jobs:
print('The following QueryJobs have no associated database table:')
for job in stale_jobs:
username = job.owner.username if job.owner else 'anonymous'
print('%s by %s -> %s.%s' % (job.id, username, job.schema_name, job.table_name))
if options['archive']:
for job in stale_jobs:
job.archive()
print('The jobs have been archived.')
else:
print('No QueryJobs without associated associated database table have been found.')
def create(self, request, *args, **kwargs):
serializer = self.get_serializer(data=request.data)
serializer.is_valid(raise_exception=True)
table = serializer.save()
if request.data.get('discover'):
adapter = DatabaseAdapter()
for column_metadata in adapter.fetch_columns(
table.schema.name, table.name):
column_metadata['table'] = table.id
column_metadata['groups'] = [
group.id for group in table.groups.all()
]
for key in ['access_level', 'metadata_access_level']:
column_metadata[key] = getattr(table, key)
column_serializer = ColumnSerializer(data=column_metadata)
if column_serializer.is_valid():
column_serializer.save()
headers = self.get_success_headers(serializer.data)
return Response(serializer.data,
status=status.HTTP_201_CREATED,
headers=headers)
def clean(self, request, resource):
if resource not in self.get_resources():
raise NotFound()
adapter = DatabaseAdapter()
self.sql = '''
SELECT %(id_field)s, %(ra_field)s, %(dec_field)s
FROM %(schema)s.%(table)s
WHERE
%(ra_field)s BETWEEN (%%(RA)s - 0.5 * %%(SR)s) AND (%%(RA)s + 0.5 * %%(SR)s)
AND
%(dec_field)s BETWEEN (%%(DEC)s - 0.5 * %%(SR)s) AND (%%(DEC)s + 0.5 * %%(SR)s)
LIMIT %(limit)s
''' % {
'id_field': adapter.escape_identifier('id'),
'ra_field': adapter.escape_identifier('ra'),
'dec_field': adapter.escape_identifier('dec'),
'schema': settings.CONESEARCH_SCHEMA,
'table': settings.CONESEARCH_TABLE,
'limit': self.max_records
}
self.args, errors = self.parse_query_dict(request)
if errors:
raise ValidationError(errors)
def discover(self, request):
schema_name = request.GET.get('schema')
table_name = request.GET.get('table')
if schema_name and table_name:
adapter = DatabaseAdapter()
table_metadata = adapter.fetch_table(schema_name, table_name)
table_metadata['nrows'] = adapter.fetch_nrows(
schema_name, table_name)
table_metadata['size'] = adapter.fetch_size(
schema_name, table_name)
return Response([table_metadata])
else:
return Response([])
def get_job_column(job, display_column_name):
try:
schema_name, table_name, column_name = \
job.metadata['display_columns'][display_column_name]
except (ValueError, KeyError):
return {}
if schema_name == settings.TAP_UPLOAD:
# for TAP_UPLOAD get the information directly from the database
return DatabaseAdapter().fetch_column(schema_name, table_name,
column_name)
else:
# for regular schemas consult the metadata store
try:
column = Column.objects.get(name=column_name,
table__name=table_name,
table__schema__name=schema_name)
return {
'name': column.name,
'description': column.description,
'unit': column.unit,
'ucd': column.ucd,
'utype': column.utype,
'datatype': column.datatype,
'arraysize': column.arraysize,
'principal': column.principal,
'indexed': False,
'std': column.std
}
except Column.DoesNotExist:
return {}
def get_datalink_list(self):
for table in self.tables:
schema_name, table_name = table.split('.')
for row in DatabaseAdapter().fetch_rows(schema_name,
table_name,
page_size=0):
yield 'datalink', row
def test_fetch_rows_filter(self):
rows = DatabaseAdapter().fetch_rows('daiquiri_data_sim',
'halos',
filters={'id': '85000000000'})
self.assertEqual(len(rows), 1)
self.assertEqual(len(rows[0]), 8)
self.assertEqual(rows[0][0], 85000000000)
def process_display_columns(processor_display_columns):
# process display_columns to expand *
display_columns = []
for processor_display_column, original_column in processor_display_columns:
if processor_display_column == '*':
schema_name, table_name, tmp = original_column
for column_name in DatabaseAdapter().fetch_column_names(
schema_name, table_name):
display_columns.append(
(column_name, (schema_name, table_name, column_name)))
else:
display_columns.append((processor_display_column, original_column))
# check for duplicate columns in display_columns
seen = set()
errors = []
for display_column_name, display_column in display_columns:
if display_column_name not in seen:
seen.add(display_column_name)
else:
errors.append(
_('Duplicate column name %(column)s') %
{'column': display_column_name})
if errors:
raise ValidationError({'query': [error for error in errors]})
return OrderedDict(display_columns)
def test_fetch_row(self):
row = DatabaseAdapter().fetch_row(
'daiquiri_data_obs',
'stars',
column_names=None,
search=None,
filters={'id': '4551299946478123136'})
self.assertEqual(row[0], 4551299946478123136)
def drop_database_table_task(schema_name, table_name):
from daiquiri.core.adapter import DatabaseAdapter
# drop the corresponding database table, but fail silently
try:
DatabaseAdapter().drop_table(schema_name, table_name)
except ProgrammingError:
pass
def abort_databae_query_task(pid):
from daiquiri.core.adapter import DatabaseAdapter
# abort the job on the database
try:
DatabaseAdapter().abort_query(pid)
except OperationalError:
# the query was probably killed before
pass
def process_query(query):
# get the adapter
adapter = DatabaseAdapter()
try:
if adapter.database_config['ENGINE'] == 'django.db.backends.mysql':
from queryparser.mysql import MySQLQueryProcessor
processor = MySQLQueryProcessor(query)
elif adapter.database_config[
'ENGINE'] == 'django.db.backends.postgresql':
from queryparser.postgresql import PostgreSQLQueryProcessor
if settings.QUERY_PROCESSOR_CACHE:
processor = cache.get_or_set('processor',
PostgreSQLQueryProcessor(), 3600)
else:
processor = PostgreSQLQueryProcessor()
# first run to replace with get_indexed_objects
processor.set_query(query)
processor.process_query(indexed_objects=get_indexed_objects(),
replace_schema_name={
'TAP_SCHEMA': settings.TAP_SCHEMA,
'tap_schema': settings.TAP_SCHEMA,
'TAP_UPLOAD': settings.TAP_UPLOAD,
'tap_upload': settings.TAP_UPLOAD,
})
# second run
processor.set_query(processor.query)
processor.process_query()
else:
raise Exception('Unknown database engine')
except QuerySyntaxError as e:
raise ValidationError({
'query': {
'messages':
[_('There has been an error while parsing your query.')],
'positions':
json.dumps(e.syntax_errors),
}
})
except QueryError as e:
raise ValidationError({'query': {
'messages': e.messages,
}})
return processor
def retrieve(self, request, pk=None):
# get database adapter
adapter = DatabaseAdapter()
# get the schema_name and the table_name from the settings
schema_name = settings.ARCHIVE_SCHEMA
table_name = settings.ARCHIVE_TABLE
# get collecions for this user
collections = [
collection.name for collection in
Collection.objects.filter_by_access_level(request.user)
]
# fetch the path for this file from the database
row = adapter.fetch_row(schema_name,
table_name, ['path'],
filters={
'id': pk,
'collection': collections
})
if row:
if request.GET.get('download', True):
# create a stats record for this download
Record.objects.create(time=now(),
resource_type='ARCHIVE_DOWNLOAD',
resource=row[0],
client_ip=get_client_ip(request),
user=request.user if
request.user.is_authenticated else None)
# send the file to the client
file_path = os.path.join(settings.ARCHIVE_BASE_PATH, row[0])
return sendfile(request, file_path, attachment=True)
else:
# send an empty response
return Response()
# if the file was not found, return 404
raise NotFound()
def discover(self, request):
schema_name = request.GET.get('schema')
table_name = request.GET.get('table')
column_name = request.GET.get('column')
if schema_name and table_name and column_name:
return Response([
DatabaseAdapter().fetch_column(schema_name, table_name,
column_name)
])
else:
return Response([])
def list(self, request, *args, **kwargs):
# get schema, table and column_names from the querystring
schema_name = self.request.GET.get('schema')
table_name = self.request.GET.get('table')
column_names = self.request.GET.getlist('column')
# get the columns which the user is allowed to access
user_columns = get_user_columns(self.request.user, schema_name, table_name)
if user_columns:
# get the row query params from the request
ordering, page, page_size, search, filters = self._get_query_params(user_columns)
# filter by input column names by the the allowed columns
if column_names:
column_names = [column.name for column in user_columns if column.name in column_names]
else:
column_names = [column.name for column in user_columns]
# get database adapter
adapter = DatabaseAdapter()
# query the database for the total number of rows
count = adapter.count_rows(schema_name, table_name, column_names, search, filters)
# query the paginated rowset
results = adapter.fetch_rows(schema_name, table_name, column_names, ordering, page, page_size, search, filters)
# return ordered dict to be send as json
return Response(OrderedDict((
('count', count),
('results', fix_for_json(results)),
('next', self._get_next_url(page, page_size, count)),
('previous', self._get_previous_url(page))
)))
# if nothing worked, return 404
raise NotFound()
def run_sync(self):
adapter = DatabaseAdapter()
self.actual_query = adapter.build_sync_query(
self.native_query, settings.QUERY_SYNC_TIMEOUT, self.max_records)
job_sources = get_job_sources(self)
# create a stats record for this job
Record.objects.create(time=now(),
resource_type='QUERY',
resource={
'job_id': None,
'job_type': self.job_type,
'query': self.query,
'query_language': self.query_language,
'sources': job_sources
},
client_ip=self.client_ip,
user=self.owner)
try:
download_adapter = DownloadAdapter()
yield from generate_votable(
adapter.fetchall(self.actual_query),
get_job_columns(self),
table=download_adapter.get_table_name(self.schema_name,
self.table_name),
infos=download_adapter.get_infos('OK', self.query,
self.query_language,
job_sources),
links=download_adapter.get_links(job_sources))
self.drop_uploads()
except (OperationalError, ProgrammingError, InternalError,
DataError) as e:
raise StopIteration()
def handle(self, *args, **options):
if options['user']:
usernames = [options['user']]
else:
usernames = ['anonymous'] + list(User.objects.values_list('username', flat=True))
adapter = DatabaseAdapter()
stale_tables = []
for username in usernames:
schema_name = settings.QUERY_USER_SCHEMA_PREFIX + username
tables = adapter.fetch_tables(schema_name)
jobs = QueryJob.objects.filter(schema_name=schema_name)
for table in tables:
job = QueryJob.objects.filter(
schema_name=schema_name,
table_name=table['name']
).first()
if job and job.phase not in [QueryJob.PHASE_EXECUTING, QueryJob.PHASE_COMPLETED]:
stale_tables.append((schema_name, table['name'], job.phase if job else None))
if stale_tables:
print('The following database tables have no associated QueryJob:')
for stale_table in stale_tables:
print('%s.%s -> %s' % stale_table)
if options['delete']:
for schema_name, table_name, phase in stale_tables:
adapter.drop_table(schema_name, table_name)
print('The tables have been deleted.')
else:
print('No tables without associated QueryJob have been found.')
def rows(self, column_names, ordering, page, page_size, search, filters):
if self.phase == self.PHASE_COMPLETED:
# check if the columns are actually in the jobs table
errors = {}
for column_name in column_names:
if column_name not in self.column_names:
errors[column_name] = _('Column not found.')
if errors:
raise ValidationError(errors)
# get database adapter
adapter = DatabaseAdapter()
try:
# query the database for the total number of rows
count = adapter.count_rows(self.schema_name, self.table_name,
column_names, search, filters)
# query the paginated rowset
rows = adapter.fetch_rows(self.schema_name, self.table_name,
column_names, ordering, page,
page_size, search, filters)
# flatten the list if only one column is retrieved
if len(column_names) == 1:
return count, [element for row in rows for element in row]
else:
return count, rows
except ProgrammingError:
return 0, []
else:
raise ValidationError({'phase': ['Job is not COMPLETED.']})
def list(self, request, *args, **kwargs):
# get the row query params from the request
ordering, page, page_size, search, filters = self._get_query_params(
settings.ARCHIVE_COLUMNS)
# get database adapter
adapter = DatabaseAdapter()
# get the schema_name and the table_name from the settings
schema_name = settings.ARCHIVE_SCHEMA
table_name = settings.ARCHIVE_TABLE
# get collecions for this user and add them to the filters
collections = [
collection.name for collection in
Collection.objects.filter_by_access_level(request.user)
]
filters['collection'] = collections
# get the name of the columns
column_names = [column['name'] for column in settings.ARCHIVE_COLUMNS]
# query the database for the total number of rows
count = adapter.count_rows(schema_name, table_name, column_names,
search, filters)
# query the paginated rowset
results = adapter.fetch_rows(schema_name, table_name, column_names,
ordering, page, page_size, search,
filters)
# return ordered dict to be send as json
return Response(
OrderedDict((('count', count), ('results', results),
('next', self._get_next_url(page, page_size, count)),
('previous', self._get_previous_url(page)))))
def get_job_columns(job):
columns = []
if job.phase == job.PHASE_COMPLETED:
database_columns = DatabaseAdapter().fetch_columns(
job.schema_name, job.table_name)
for database_column in database_columns:
column = get_job_column(job, database_column['name'])
column.update(database_column)
columns.append(column)
else:
for display_column in job.metadata['display_columns']:
columns.append(get_job_column(job, display_column))
return columns
def _test_detail_viewset(self, username):
schema_name = settings.ARCHIVE_SCHEMA
table_name = settings.ARCHIVE_TABLE
rows = DatabaseAdapter().fetch_rows(schema_name, table_name, ['id', 'collection'], None, None, 0, None, None)
for row in rows:
url = reverse(self.url_name, kwargs={'pk': row[0]})
status_code = self.get_status_code(username, row[1])
response = self.client.get(url)
msg = OrderedDict((
('username', username),
('url', url),
('row', row),
('status_code', response.status_code)
))
self.assertEqual(response.status_code, status_code, msg=msg)
def handle(self, *args, **options):
# look for completed jobs with no table
queryset = QueryJob.objects.filter(phase=QueryJob.PHASE_COMPLETED)
if options['user']:
queryset = queryset.filter(owner__username=options['user'])
for job in queryset:
try:
DatabaseAdapter().fetch_size(job.schema_name, job.table_name)
except ProgrammingError:
try:
job.phase = QueryJob.PHASE_PENDING
if options['queue']:
job.queue = options['queue']
job.process()
print('Run %s by %s again.' % (job.id, job.owner))
if not options['dry']:
job.run()
except ValidationError as e:
job.phase = QueryJob.PHASE_ERROR
job.error_summary = ''
for key, errors in e.detail.items():
try:
job.error_summary += ''.join(errors['messages'])
except TypeError:
job.error_summary += ''.join(errors)
print('Error for %s by %s: %s' %
(job.id, job.owner, job.error_summary))
if not options['dry']:
job.save()
def process(self):
if self.job.phase == self.PHASE_COMPLETED:
self.owner = self.job.owner
else:
raise ValidationError({'phase': ['Job is not COMPLETED.']})
if not self.column_name:
raise ValidationError(
{'column_name': [_('This field may not be blank.')]})
if self.column_name not in self.job.column_names:
raise ValidationError({
'column_name': [_('Unknown column "%s".') % self.column_name]
})
# get database adapter and query the paginated rowset
rows = DatabaseAdapter().fetch_rows(self.job.schema_name,
self.job.table_name,
[self.column_name],
page_size=0)
# prepare list of files for this job
files = []
for row in rows:
file_path = row[0]
# append the file to the list of files if it exists
if file_path and check_file(self.owner, file_path):
files.append(file_path)
else:
raise ValidationError({
'files': [_('One or more of the files cannot be found.')]
})
# set files for this job
self.files = files
# set clean flag
self.is_clean = True
def ingest_table(schema_name, table_name, file_path, drop_table=False):
adapter = DatabaseAdapter()
table = parse_single_table(file_path, pedantic=False)
columns = []
for field in table.fields:
columns.append({
'name': field.name,
'datatype': field.datatype,
'ucd': field.ucd,
'unit': str(field.unit),
})
if drop_table:
adapter.drop_table(schema_name, table_name)
adapter.create_table(schema_name, table_name, columns)
adapter.insert_rows(schema_name, table_name, columns, table.array,
table.array.mask)
os.remove(file_path)
return columns
def translate_query(query_language, query):
# get the adapter
adapter = DatabaseAdapter()
# translate adql -> mysql string
if query_language == 'adql-2.0':
try:
translator = cache.get_or_set('translator', ADQLQueryTranslator(),
3600)
translator.set_query(query)
if adapter.database_config['ENGINE'] == 'django.db.backends.mysql':
return translator.to_mysql()
elif adapter.database_config[
'ENGINE'] == 'django.db.backends.postgresql':
return translator.to_postgresql()
else:
raise Exception('Unknown database engine')
except QuerySyntaxError as e:
raise ValidationError({
'query': {
'messages': [
_('There has been an error while translating your query.'
)
],
'positions':
json.dumps(e.syntax_errors),
}
})
except QueryError as e:
raise ValidationError({'query': {
'messages': e.messages,
}})
else:
return query
def process(self):
# get collections for the owner of this download job
collections = [
collection.name for collection in
Collection.objects.filter_by_access_level(self.owner)
]
# get database adapter
adapter = DatabaseAdapter()
# get the schema_name and the table_name from the settings
schema_name = settings.ARCHIVE_SCHEMA
table_name = settings.ARCHIVE_TABLE
# prepare list of files for this archive job
files = []
if 'file_ids' in self.data:
if isinstance(self.data, QueryDict):
file_ids = self.data.getlist('file_ids')
else:
file_ids = self.data.get('file_ids')
for file_id in file_ids:
# validate that the file_id is a valid UUID4
try:
uuid.UUID(file_id, version=4)
except ValueError:
raise ValidationError({
'files': [
_('One or more of the identifiers are not valid UUIDs.'
)
]
})
# fetch the path for this file from the database
row = adapter.fetch_row(schema_name,
table_name, ['path'],
filters={
'id': file_id,
'collection': collections
})
# append the file to the list of files only if it exists in the database and on the filesystem
if row and os.path.isfile(
os.path.join(settings.ARCHIVE_BASE_PATH, row[0])):
files.append(row[0])
else:
raise ValidationError({
'files':
[_('One or more of the files cannot be found.')]
})
elif 'search' in self.data:
# retrieve the pathes of all file matching the search criteria
rows = adapter.fetch_rows(schema_name,
table_name,
page_size=0,
search=self.data['search'],
filters={'collection': collections})
# get the index of the path column in the row
path_index = next(
(i for i, column in enumerate(settings.ARCHIVE_COLUMNS)
if column['name'] == 'path'))
for row in rows:
# append the file to the list of files only if it exists on the filesystem
if os.path.isfile(
os.path.join(settings.ARCHIVE_BASE_PATH,
row[path_index])):
files.append(row[path_index])
else:
raise ValidationError({
'files':
[_('One or more of the files cannot be found.')]
})
else:
raise ValidationError({[_('No data received.')]})
# set files and file_path for this archive job
self.files = files
# set clean flag
self.is_clean = True
def run_database_query_task(job_id):
# always import daiquiri packages inside the task
from daiquiri.core.adapter import DatabaseAdapter
from daiquiri.query.models import QueryJob
from daiquiri.query.utils import get_quota, get_job_sources, get_job_columns, ingest_uploads
from daiquiri.stats.models import Record
# get logger
logger = logging.getLogger(__name__)
# get the job object from the database
job = QueryJob.objects.get(pk=job_id)
if job.phase == job.PHASE_QUEUED:
# get the adapter with the database specific functions
adapter = DatabaseAdapter()
# create the database of the user if it not already exists
try:
adapter.create_user_schema_if_not_exists(job.schema_name)
except OperationalError as e:
job.phase = job.PHASE_ERROR
job.error_summary = str(e)
job.save()
return job.phase
# check if the quota is exceeded
if QueryJob.objects.get_size(job.owner) > get_quota(job.owner):
job.phase = job.PHASE_ERROR
job.error_summary = str(
_('Quota is exceeded. Please remove some of your jobs.'))
job.save()
return job.phase
# set database and start time
job.pid = adapter.fetch_pid()
job.actual_query = adapter.build_query(job.schema_name, job.table_name,
job.native_query, job.timeout,
job.max_records)
job.phase = job.PHASE_EXECUTING
job.start_time = now()
job.save()
logger.info('job %s started' % job.id)
# get the actual query and submit the job to the database
try:
ingest_uploads(job.uploads, job.owner)
# this is where the work ist done (and the time is spend)
adapter.submit_query(job.actual_query)
except (ProgrammingError, InternalError, ValueError) as e:
job.phase = job.PHASE_ERROR
job.error_summary = str(e)
logger.info('job %s failed (%s)' % (job.id, job.error_summary))
except OperationalError as e:
# load the job again and check if the job was killed
job = QueryJob.objects.get(pk=job_id)
if job.phase != job.PHASE_ABORTED:
job.phase = job.PHASE_ERROR
job.error_summary = str(e)
logger.info('job %s failed (%s)' % (job.id, job.error_summary))
else:
# get additional information about the completed job
job.phase = job.PHASE_COMPLETED
logger.info('job %s completed' % job.id)
finally:
# get timing and save the job object
job.end_time = now()
# get additional information about the completed job
if job.phase == job.PHASE_COMPLETED:
job.nrows = adapter.count_rows(job.schema_name, job.table_name)
job.size = adapter.fetch_size(job.schema_name, job.table_name)
# fetch the metadata for used tables
job.metadata['sources'] = get_job_sources(job)
# fetch the metadata for the columns and fetch additional metadata from the metadata store
job.metadata['columns'] = get_job_columns(job)
# remove unneeded metadata
job.metadata.pop('display_columns', None)
job.metadata.pop('tables', None)
# create a stats record for this job
Record.objects.create(time=job.end_time,
resource_type='QUERY',
resource={
'job_id': job.id,
'job_type': job.job_type,
'query': job.query,
'query_language': job.query_language,
'sources':
job.metadata.get('sources', [])
},
client_ip=job.client_ip,
user=job.owner)
job.save()
return job.phase
def rename_database_table_task(schema_name, table_name, new_table_name):
from daiquiri.core.adapter import DatabaseAdapter
DatabaseAdapter().rename_table(schema_name, table_name, new_table_name)
def run_database_ingest_task(job_id, file_path):
from daiquiri.core.adapter import DatabaseAdapter
from daiquiri.query.models import QueryJob
from daiquiri.stats.models import Record
from daiquiri.query.utils import get_quota, ingest_table
# get logger
logger = logging.getLogger(__name__)
# get the job object from the database
job = QueryJob.objects.get(pk=job_id)
if job.phase == job.PHASE_QUEUED:
# get the adapter with the database specific functions
adapter = DatabaseAdapter()
# create the database of the user if it not already exists
try:
adapter.create_user_schema_if_not_exists(job.schema_name)
except OperationalError as e:
job.phase = job.PHASE_ERROR
job.error_summary = str(e)
job.save()
return job.phase
# check if the quota is exceeded
if QueryJob.objects.get_size(job.owner) > get_quota(job.owner):
job.phase = job.PHASE_ERROR
job.error_summary = str(
_('Quota is exceeded. Please remove some of your jobs.'))
job.save()
return job.phase
# set database and start time
job.pid = adapter.fetch_pid()
job.phase = job.PHASE_EXECUTING
job.start_time = now()
job.save()
logger.info('job %s started' % job.id)
# create the table and insert the data
try:
columns = ingest_table(job.schema_name, job.table_name, file_path)
except (ProgrammingError, InternalError, ValueError) as e:
job.phase = job.PHASE_ERROR
job.error_summary = str(e)
logger.info('job %s failed (%s)' % (job.id, job.error_summary))
except OperationalError as e:
# load the job again and check if the job was killed
job = QueryJob.objects.get(pk=job_id)
if job.phase != job.PHASE_ABORTED:
job.phase = job.PHASE_ERROR
job.error_summary = str(e)
logger.info('job %s failed (%s)' % (job.id, job.error_summary))
else:
# get additional information about the completed job
job.phase = job.PHASE_COMPLETED
logger.info('job %s completed' % job.id)
finally:
# get timing and save the job object
job.end_time = now()
# get additional information about the completed job
if job.phase == job.PHASE_COMPLETED:
job.nrows = adapter.count_rows(job.schema_name, job.table_name)
job.size = adapter.fetch_size(job.schema_name, job.table_name)
# store the metadata for the columns from the VOTable
job.metadata = {'columns': columns}
# create a stats record for this job
Record.objects.create(time=job.end_time,
resource_type='UPLOAD',
resource={
'job_id': job.id,
'job_type': job.job_type,
},
client_ip=job.client_ip,
user=job.owner)
job.save()
return job.phase
def check_permissions(user, keywords, tables, columns, functions):
messages = []
# check keywords against whitelist
for keywords in keywords:
pass
# loop over tables to check permissions on schemas/tables
for schema_name, table_name in tables:
# check permission on schema
if schema_name is None:
# schema_name must not be null, move to next table
messages.append(_('No schema given for table %s.') % table_name)
continue
elif schema_name == get_user_schema_name(user):
# all tables are allowed move to next table
continue
elif schema_name == settings.TAP_UPLOAD:
# all tables are allowed move to next table
continue
else:
# check permissions on the schema
try:
schema = Schema.objects.filter_by_access_level(user).get(
name=schema_name)
except Schema.DoesNotExist:
# schema not found or not allowed, move to next table
messages.append(_('Schema %s not found.') % schema_name)
continue
# check permission on table
if table_name is None:
# table_name must not be null, move to next table
messages.append(_('No table given for schema %s.') % schema_name)
continue
else:
try:
Table.objects.filter_by_access_level(user).filter(
schema=schema).get(name=table_name)
except Table.DoesNotExist:
# table not found or not allowed, move to next table
messages.append(_('Table %s not found.') % table_name)
continue
# loop over columns to check permissions or just to see if they are there,
# but only if no error messages where appended so far
if not messages:
for schema_name, table_name, column_name in columns:
if schema_name in [None, get_user_schema_name(user), settings.TAP_UPLOAD] \
or table_name is None \
or column_name is None:
# doesn't need to be checked, move to next column
continue
else:
if not settings.METADATA_COLUMN_PERMISSIONS:
# just check if the column exist
if column_name == '*':
# doesn't need to be checked, move to next table
continue
else:
try:
Column.objects.filter(
table__schema__name=schema_name).filter(
table__name=table_name).get(
name=column_name)
except Column.DoesNotExist:
messages.append(
_('Column %s not found.') % column_name)
continue
else:
try:
schema = Schema.objects.filter_by_access_level(
user).get(name=schema_name)
except Schema.DoesNotExist:
messages.append(
_('Schema %s not found.') % schema_name)
continue
try:
table = Table.objects.filter_by_access_level(
user).filter(schema=schema).get(name=table_name)
except Table.DoesNotExist:
messages.append(_('Table %s not found.') % table_name)
continue
if column_name == '*':
columns = Column.objects.filter_by_access_level(
user).filter(table=table)
actual_columns = DatabaseAdapter().fetch_columns(
schema_name, table_name)
column_names_set = set(
[column.name for column in columns])
actual_column_names_set = set(
[column['name'] for column in actual_columns])
if column_names_set != actual_column_names_set:
messages.append(
_('The asterisk (*) is not allowed for this table.'
))
continue
else:
try:
column = Column.objects.filter_by_access_level(
user).filter(table=table).get(name=column_name)
except Column.DoesNotExist:
messages.append(
_('Column %s not found.') % column_name)
continue
# check permissions on functions
for function_name in functions:
# check permission on function
queryset = Function.objects.filter(name=function_name)
# forbit the function if it is in metadata.functions, and the user doesn't have access.
if queryset and not queryset.filter_by_access_level(user):
messages.append(_('Function %s is not allowed.') % function_name)
else:
continue
# return the error stack
return list(set(messages))
