def _normalize_document_uris_window(session, window):
query = (session.query(models.DocumentURI).filter(
models.DocumentURI.updated.between(window.start, window.end)).order_by(
models.DocumentURI.updated.asc()))
for docuri in query:
documents = models.Document.find_by_uris(session, [docuri.uri])
if documents.count() > 1:
merge_documents(session, documents)
existing = session.query(models.DocumentURI).filter(
models.DocumentURI.id != docuri.id,
models.DocumentURI.document_id == docuri.document_id,
models.DocumentURI.claimant_normalized == uri.normalize(
docuri.claimant),
models.DocumentURI.uri_normalized == uri.normalize(docuri.uri),
models.DocumentURI.type == docuri.type,
models.DocumentURI.content_type == docuri.content_type,
)
if existing.count() > 0:
session.delete(docuri)
else:
docuri._claimant_normalized = uri.normalize(docuri.claimant)
docuri._uri_normalized = uri.normalize(docuri.uri)
session.flush()
def test_raises_retryable_error_when_flush_fails(self, db_session, merge_data, monkeypatch):
def err():
raise sa.exc.IntegrityError(None, None, None)
monkeypatch.setattr(db_session, 'flush', err)
with pytest.raises(transaction.interfaces.TransientError):
document.merge_documents(db_session, merge_data)
def _normalize_document_uris_window(session, window):
query = (
session.query(models.DocumentURI)
.filter(models.DocumentURI.updated.between(window.start, window.end))
.order_by(models.DocumentURI.updated.asc())
)
for docuri in query:
documents = models.Document.find_by_uris(session, [docuri.uri])
if documents.count() > 1:
merge_documents(session, documents)
existing = session.query(models.DocumentURI).filter(
models.DocumentURI.id != docuri.id,
models.DocumentURI.document_id == docuri.document_id,
models.DocumentURI.claimant_normalized == uri.normalize(docuri.claimant),
models.DocumentURI.uri_normalized == uri.normalize(docuri.uri),
models.DocumentURI.type == docuri.type,
models.DocumentURI.content_type == docuri.content_type,
)
if existing.count() > 0:
session.delete(docuri)
else:
docuri._claimant_normalized = uri.normalize(docuri.claimant)
docuri._uri_normalized = uri.normalize(docuri.uri)
session.flush()
def test_raises_retryable_error_when_flush_fails(self, db_session, merge_data, monkeypatch):
def err():
raise sa.exc.IntegrityError(None, None, None)
monkeypatch.setattr(db_session, 'flush', err)
with pytest.raises(transaction.interfaces.TransientError):
document.merge_documents(db_session, merge_data)
def test_merge_documents_logs_when_its_called(self, caplog, db_session,
merge_data):
caplog.set_level(logging.INFO)
document.merge_documents(db_session, merge_data)
assert caplog.record_tuples == [("h.models.document", 20,
"Merging 3 documents")]
def test_raises_retryable_error_when_flush_fails(self, db_session,
merge_data, monkeypatch):
def err():
raise sa.exc.IntegrityError(None, None, None)
monkeypatch.setattr(db_session, "flush", err)
with pytest.raises(ConcurrentUpdateError):
document.merge_documents(db_session, merge_data)
def test_merge_documents_rewires_document_meta(self, db_session, merge_data):
master, duplicate_1, duplicate_2 = merge_data
document.merge_documents(db_session, merge_data)
db_session.flush()
assert len(master.meta) == 3
assert len(duplicate_1.meta) == 0
assert len(duplicate_2.meta) == 0
def test_merge_documents_rewires_document_meta(self, db_session, merge_data):
master, duplicate_1, duplicate_2 = merge_data
document.merge_documents(db_session, merge_data)
db_session.flush()
assert len(master.meta) == 3
assert len(duplicate_1.meta) == 0
assert len(duplicate_2.meta) == 0
def move_uri(ctx, old, new):
"""
Move annotations and document equivalence data from one URL to another.
This will **replace** the annotation's ``target_uri`` and all the
document uri's ``claimant``, plus the matching ``uri`` for self-claim and
canonical uris.
"""
request = ctx.obj["bootstrap"]()
annotations = _fetch_annotations(request.db, old)
docuris_claimant = _fetch_document_uri_claimants(request.db, old)
docuris_uri = _fetch_document_uri_canonical_self_claim(request.db, old)
prompt = (
"Changing all annotations and document data matching:\n"
'"{old}"\nto:\n"{new}"\n'
"This will affect {ann_count} annotations, {doc_claimant} "
"document uri claimants, and {doc_uri} document uri self-claims "
"or canonical uris.\n"
"Are you sure? [y/N]"
).format(
old=old,
new=new,
ann_count=len(annotations),
doc_claimant=len(docuris_claimant),
doc_uri=len(docuris_uri),
)
c = click.prompt(prompt, default="n", show_default=False)
if c != "y":
print("Aborted")
return
for annotation in annotations:
annotation.target_uri = new
for docuri in docuris_claimant:
docuri.claimant = new
for docuri in docuris_uri:
docuri.uri = new
if annotations:
indexer = BatchIndexer(request.db, request.es, request)
ids = [a.id for a in annotations]
indexer.index(ids)
request.db.flush()
documents = models.Document.find_by_uris(request.db, [new])
if documents.count() > 1:
merge_documents(request.db, documents)
request.tm.commit()
def test_merge_documents_deletes_duplicate_documents(self, db_session, merge_data):
_, duplicate_1, duplicate_2 = merge_data
document.merge_documents(db_session, merge_data)
db_session.flush()
count = db_session.query(document.Document) \
.filter(document.Document.id.in_([duplicate_1.id, duplicate_2.id])) \
.count()
assert count == 0
def move_uri(ctx, old, new):
"""
Move annotations and document equivalence data from one URL to another.
This will **replace** the annotation's ``target_uri`` and all the
document uri's ``claimant``, plus the matching ``uri`` for self-claim and
canonical uris.
"""
request = ctx.obj["bootstrap"]()
annotations = _fetch_annotations(request.db, old)
docuris_claimant = _fetch_document_uri_claimants(request.db, old)
docuris_uri = _fetch_document_uri_canonical_self_claim(request.db, old)
prompt = ("Changing all annotations and document data matching:\n"
'"{old}"\nto:\n"{new}"\n'
"This will affect {ann_count} annotations, {doc_claimant} "
"document uri claimants, and {doc_uri} document uri self-claims "
"or canonical uris.\n"
"Are you sure? [y/N]").format(
old=old,
new=new,
ann_count=len(annotations),
doc_claimant=len(docuris_claimant),
doc_uri=len(docuris_uri),
)
c = click.prompt(prompt, default="n", show_default=False)
if c != "y":
print("Aborted")
return
for annotation in annotations:
annotation.target_uri = new
for docuri in docuris_claimant:
docuri.claimant = new
for docuri in docuris_uri:
docuri.uri = new
if annotations:
indexer = BatchIndexer(request.db, request.es, request)
ids = [a.id for a in annotations]
indexer.index(ids)
request.db.flush()
documents = models.Document.find_by_uris(request.db, [new])
if documents.count() > 1:
merge_documents(request.db, documents)
request.tm.commit()
def test_merge_documents_deletes_duplicate_documents(self, db_session, merge_data):
_, duplicate_1, duplicate_2 = merge_data
document.merge_documents(db_session, merge_data)
db_session.flush()
count = db_session.query(document.Document) \
.filter(document.Document.id.in_([duplicate_1.id, duplicate_2.id])) \
.count()
assert count == 0
def test_merge_documents_rewires_annotations(self, db_session, merge_data):
master, duplicate_1, duplicate_2 = merge_data
document.merge_documents(db_session, merge_data)
db_session.flush()
assert (6 == db_session.query(
models.Annotation).filter_by(document_id=master.id).count())
assert (0 == db_session.query(
models.Annotation).filter_by(document_id=duplicate_1.id).count())
assert (0 == db_session.query(
models.Annotation).filter_by(document_id=duplicate_2.id).count())
def test_merge_documents_rewires_annotations(self, db_session, merge_data):
master, duplicate_1, duplicate_2 = merge_data
document.merge_documents(db_session, merge_data)
db_session.flush()
assert 6 == \
db_session.query(models.Annotation).filter_by(document_id=master.id).count()
assert 0 == \
db_session.query(models.Annotation).filter_by(document_id=duplicate_1.id).count()
assert 0 == \
db_session.query(models.Annotation).filter_by(document_id=duplicate_2.id).count()
def test_merge_documents_returns_master(self, db_session, merge_data):
master, _, _ = merge_data
merged_master = document.merge_documents(db_session, merge_data)
assert merged_master == master
def test_merge_documents_returns_master(self, db_session, merge_data):
master, _, _ = merge_data
merged_master = document.merge_documents(db_session, merge_data)
assert merged_master == master
