Changeset View
Changeset View
Standalone View
Standalone View
swh/indexer/storage/db.py
Show First 20 Lines • Show All 330 Lines • ▼ Show 20 Lines | def revision_intrinsic_metadata_delete( | ||||
for e in entries),) | for e in entries),) | ||||
def revision_intrinsic_metadata_get_from_list(self, ids, cur=None): | def revision_intrinsic_metadata_get_from_list(self, ids, cur=None): | ||||
yield from self._get_from_list( | yield from self._get_from_list( | ||||
'revision_intrinsic_metadata', ids, | 'revision_intrinsic_metadata', ids, | ||||
self.revision_intrinsic_metadata_cols, cur=cur) | self.revision_intrinsic_metadata_cols, cur=cur) | ||||
origin_intrinsic_metadata_cols = [ | origin_intrinsic_metadata_cols = [ | ||||
'id', 'origin_url', 'metadata', 'from_revision', 'mappings', | 'id', 'metadata', 'from_revision', 'mappings', | ||||
'tool_id', 'tool_name', 'tool_version', 'tool_configuration'] | 'tool_id', 'tool_name', 'tool_version', 'tool_configuration'] | ||||
origin_intrinsic_metadata_regconfig = 'pg_catalog.simple' | origin_intrinsic_metadata_regconfig = 'pg_catalog.simple' | ||||
"""The dictionary used to normalize 'metadata' and queries. | """The dictionary used to normalize 'metadata' and queries. | ||||
'pg_catalog.simple' provides no stopword, so it should be suitable | 'pg_catalog.simple' provides no stopword, so it should be suitable | ||||
for proper names and non-English content. | for proper names and non-English content. | ||||
When updating this value, make sure to add a new index on | When updating this value, make sure to add a new index on | ||||
origin_intrinsic_metadata.metadata.""" | origin_intrinsic_metadata.metadata.""" | ||||
Show All 13 Lines | def origin_intrinsic_metadata_delete( | ||||
cur = self._cursor(cur) | cur = self._cursor(cur) | ||||
cur.execute( | cur.execute( | ||||
"DELETE from origin_intrinsic_metadata " | "DELETE from origin_intrinsic_metadata " | ||||
"WHERE (id, indexer_configuration_id) IN" | "WHERE (id, indexer_configuration_id) IN" | ||||
" (VALUES %s)" % (', '.join('%s' for _ in entries)), | " (VALUES %s)" % (', '.join('%s' for _ in entries)), | ||||
tuple((e['id'], e['indexer_configuration_id']) | tuple((e['id'], e['indexer_configuration_id']) | ||||
for e in entries),) | for e in entries),) | ||||
def origin_intrinsic_metadata_get_from_list(self, orig_ids, cur=None): | def origin_intrinsic_metadata_get_from_list(self, ids, cur=None): | ||||
yield from self._get_from_list( | yield from self._get_from_list( | ||||
'origin_intrinsic_metadata', orig_ids, | 'origin_intrinsic_metadata', ids, | ||||
self.origin_intrinsic_metadata_cols, cur=cur, | self.origin_intrinsic_metadata_cols, cur=cur, | ||||
id_col='id') | id_col='id') | ||||
def origin_intrinsic_metadata_search_fulltext(self, terms, *, limit, cur): | def origin_intrinsic_metadata_search_fulltext(self, terms, *, limit, cur): | ||||
regconfig = self.origin_intrinsic_metadata_regconfig | regconfig = self.origin_intrinsic_metadata_regconfig | ||||
tsquery_template = ' && '.join("plainto_tsquery('%s', %%s)" % regconfig | tsquery_template = ' && '.join("plainto_tsquery('%s', %%s)" % regconfig | ||||
for _ in terms) | for _ in terms) | ||||
tsquery_args = [(term,) for term in terms] | tsquery_args = [(term,) for term in terms] | ||||
▲ Show 20 Lines • Show All 79 Lines • Show Last 20 Lines |