def load_mapping(collection, mapping_id, sync=False):
"""Flush and reload all entities generated by a mapping."""
mapping = Mapping.by_id(mapping_id)
if mapping is None:
return log.error("Could not find mapping: %s", mapping_id)
origin = mapping_origin(mapping.id)
aggregator = get_aggregator(collection)
aggregator.delete(origin=origin)
delete_entities(collection.id, origin=origin, sync=True)
if mapping.disabled:
return log.info("Mapping is disabled: %s", mapping_id)
publish(
Events.LOAD_MAPPING,
params={
"collection": collection,
"table": mapping.table_id
},
channels=[collection, mapping.role],
actor_id=mapping.role_id,
)
try:
map_to_aggregator(collection, mapping, aggregator)
aggregate_model(collection, aggregator)
index_aggregator(collection, aggregator, sync=sync)
mapping.set_status(status=Status.SUCCESS)
db.session.commit()
except Exception as exc:
mapping.set_status(status=Status.FAILED, error=str(exc))
db.session.commit()
aggregator.delete(origin=origin)
finally:
aggregator.close()
def bulk_write(collection, entities, unsafe=False, role_id=None, index=True):
"""Write a set of entities - given as dicts - to the index."""
# This is called mainly by the /api/2/collections/X/_bulk API.
now = datetime.utcnow().isoformat()
aggregator = get_aggregator(collection)
writer = aggregator.bulk()
entity_ids = set()
for data in entities:
if not is_mapping(data):
raise InvalidData("Failed to read input data", errors=data)
entity = model.get_proxy(data)
if entity.id is None:
raise InvalidData("No ID for entity", errors=entity.to_dict())
entity = collection.ns.apply(entity)
if not unsafe:
entity = remove_checksums(entity)
entity.context = {
'role_id': role_id,
'created_at': now,
'updated_at': now,
}
writer.put(entity, origin='bulk')
if index and len(entity_ids) < MAX_PAGE:
entity_ids.add(entity.id)
writer.flush()
if index:
if len(entity_ids) >= MAX_PAGE:
entity_ids = None
index_aggregator(collection, aggregator, entity_ids=entity_ids)
refresh_collection(collection.id)
def update_entity(collection, entity_id=None):
"""Update xref and aggregator after an entity has been edited."""
from aleph.logic.xref import xref_entity
from aleph.logic.profiles import profile_fragments
log.info("[%s] Update entity: %s", collection, entity_id)
entity = index.get_entity(entity_id)
proxy = model.get_proxy(entity)
if collection.casefile:
xref_entity(collection, proxy)
aggregator = get_aggregator(collection, origin=MODEL_ORIGIN)
profile_fragments(collection, aggregator, entity_id=entity_id)
# Inline name properties from adjacent entities. See the
# docstring on `inline_names` for a more detailed discussion.
prop = proxy.schema.get("namesMentioned")
if prop is not None:
entity_ids = proxy.get_type_values(registry.entity)
names = set()
for related in index.entities_by_ids(entity_ids):
related = model.get_proxy(related)
names.update(related.get_type_values(registry.name))
if len(names) > 0:
name_proxy = model.make_entity(proxy.schema)
name_proxy.id = proxy.id
name_proxy.add(prop, names)
aggregator.put(name_proxy, fragment="names")
index_aggregator(collection, aggregator, entity_ids=[entity_id])
refresh_entity(collection, proxy.id)
def bulk_write(collection,
entities,
safe=False,
role_id=None,
mutable=True,
index=True):
"""Write a set of entities - given as dicts - to the index."""
# This is called mainly by the /api/2/collections/X/_bulk API.
aggregator = get_aggregator(collection)
writer = aggregator.bulk()
entity_ids = set()
for data in entities:
entity = model.get_proxy(data, cleaned=False)
if entity.id is None:
raise InvalidData("No ID for entity", errors=entity.to_dict())
entity = collection.ns.apply(entity)
if safe:
entity = remove_checksums(entity)
entity.context = {"role_id": role_id, "mutable": mutable}
for field in ("created_at", "updated_at"):
timestamp = data.get(field)
if timestamp is not None:
dt = registry.date.to_datetime(timestamp)
if dt is not None:
entity.context[field] = dt.isoformat()
writer.put(entity, origin="bulk")
if index and len(entity_ids) < MAX_PAGE:
entity_ids.add(entity.id)
writer.flush()
if index:
if len(entity_ids) >= MAX_PAGE:
entity_ids = None
index_aggregator(collection, aggregator, entity_ids=entity_ids)
refresh_collection(collection.id)
def index_many(stage, collection, sync=False, entity_ids=None, batch=BATCH_SIZE):
"""Project the contents of the collections aggregator into the index."""
if entity_ids is not None:
entity_ids = ensure_list(entity_ids)
# WEIRD: Instead of indexing a single entity, this will try
# pull a whole batch of them off the queue and do it at once.
tasks = stage.get_tasks(limit=max(1, batch - len(entity_ids)))
for task in tasks:
entity_ids.extend(ensure_list(task.payload.get("entity_ids")))
stage.mark_done(len(tasks))
aggregator = get_aggregator(collection)
index_aggregator(collection, aggregator, entity_ids=entity_ids, sync=sync)
refresh_collection(collection.id)
def save_entityset_item(entityset, collection, entity_id, **data):
"""Change the association between an entity and an entityset. In the case of
a profile, this may require re-indexing of the entity to update the associated
profile_id.
"""
item = EntitySetItem.save(entityset,
entity_id,
collection_id=collection.id,
**data)
if entityset.type == EntitySet.PROFILE and entityset.collection_id == collection.id:
from aleph.logic.profiles import profile_fragments
aggregator = get_aggregator(collection)
profile_fragments(collection, aggregator, entity_id=entity_id)
index_aggregator(collection, aggregator, entity_ids=[entity_id])
refresh_entity(collection, entity_id)
refresh_entityset(entityset.id)
return item
