mirror of
https://github.com/tiennm99/DocsGPT.git
synced 2026-10-03 07:11:56 +00:00
Query embedding moved to the Celery worker, but nothing that ships was updated to consume the queue it dispatches to. - Add `embeddings` to every worker `-Q` list (compose x3, k8s, devcontainer, sandbox README). Without it a search blocked for EMBEDDINGS_DELEGATE_TIMEOUT and then answered with no retrieved context, because classic_rag swallows the dispatch error and skips the source -- bad answers, not an error. - Skip the task_postrun heap reclaim for the embed task. The full gc.collect() was written for docling/torch parses; on a worker holding the ONNX model it measured ~86ms against ~8ms for the embed itself, a 9x slowdown of the round trip for a task that allocates a few kilobytes. - Resolve the installation pin in the re-embed script. It never imports application.app, so an install pinned in app_metadata with no EMBEDDINGS_NAME set -- every stock k8s deployment, whose manifests carry no embedding config -- would rewrite its whole index with the legacy default and stamp sources.model to match, then be told by the boot warning to run it again. - Fail fast for 30s after a failed dispatch. fanout.embed_questions falls back to letting each store embed its own query, so one dead-worker retrieval paid the timeout once in the fan-out and again per source. - Forget the task result. Nothing reads it back: the key is per-dispatch UUID, not content-addressed, so a repeated query mints another. Left alone every search leaked ~17KB for result_expires (7 days) into the Redis the broker shares -- on the bundled k8s manifest (1Gi, no maxmemory policy) that is an OOMKill that takes the broker with it. - Release the model ensure_vector_schema loads to read the width of an unregistered model, in a process that delegates and would never call it. The width still comes from the model, not the table, so the mismatch check the hook exists for keeps working. - Correct the docs that said otherwise: embeddings.md claimed the standard deployment worked unchanged, upgrading.mdx said no action was needed, and the settings table listed none of the three delegation settings.
462 lines
20 KiB
Python
462 lines
20 KiB
Python
"""Re-embed script: CLI contract, orchestration, and the FAISS rebuild path."""
|
|
|
|
from unittest.mock import MagicMock, patch
|
|
|
|
import pytest
|
|
|
|
from application.scripts import reembed
|
|
|
|
|
|
class TestCLI:
|
|
def test_defaults(self):
|
|
args = reembed.build_parser().parse_args([])
|
|
assert args.dry_run is False
|
|
assert args.sources is None
|
|
assert args.batch_size == reembed.DEFAULT_BATCH_SIZE
|
|
|
|
def test_unsupported_store_exits_without_touching_anything(self):
|
|
with patch.object(reembed.settings, "VECTOR_STORE", "qdrant", create=True):
|
|
with patch.object(reembed, "run") as run:
|
|
assert reembed.main([]) == 2
|
|
run.assert_not_called()
|
|
|
|
def test_supported_stores_are_pgvector_and_faiss(self):
|
|
assert set(reembed.SUPPORTED_STORES) == {"pgvector", "faiss"}
|
|
|
|
def test_sources_are_split_and_trimmed(self):
|
|
with patch.object(reembed.settings, "VECTOR_STORE", "faiss", create=True):
|
|
with patch.object(reembed, "run", return_value=0) as run:
|
|
reembed.main(["--sources", " a , b ,, c "])
|
|
assert run.call_args.args[1] == ["a", "b", "c"]
|
|
|
|
def test_batch_size_is_clamped_to_at_least_one(self):
|
|
with patch.object(reembed.settings, "VECTOR_STORE", "faiss", create=True):
|
|
with patch.object(reembed, "run", return_value=0) as run:
|
|
reembed.main(["--batch-size", "0"])
|
|
assert run.call_args.args[2] == 1
|
|
|
|
|
|
class TestRun:
|
|
def test_no_sources_is_a_clean_exit(self):
|
|
with patch.object(reembed, "list_source_ids", return_value=[]):
|
|
assert reembed.run("faiss", None, 8, False) == 0
|
|
|
|
def test_processes_every_discovered_source(self):
|
|
with patch.object(reembed, "list_source_ids", return_value=["a", "b"]):
|
|
with patch.object(reembed, "reembed_faiss", return_value=(3, 3)) as handler:
|
|
assert reembed.run("faiss", None, 8, False) == 0
|
|
assert [call.args[0] for call in handler.call_args_list] == ["a", "b"]
|
|
|
|
def test_explicit_sources_skip_discovery(self):
|
|
with patch.object(reembed, "list_source_ids") as discover:
|
|
with patch.object(reembed, "reembed_faiss", return_value=(1, 1)):
|
|
reembed.run("faiss", ["only-this"], 8, False)
|
|
discover.assert_not_called()
|
|
|
|
def test_one_failing_source_does_not_stop_the_others(self):
|
|
def handler(source_id, batch_size, dry_run):
|
|
if source_id == "bad":
|
|
raise RuntimeError("boom")
|
|
return (1, 1)
|
|
|
|
with patch.object(reembed, "reembed_faiss", side_effect=handler) as spy:
|
|
code = reembed.run("faiss", ["good", "bad", "also-good"], 8, False)
|
|
assert code == 1, "a failure must be reported in the exit code"
|
|
assert spy.call_count == 3, "later sources must still be attempted"
|
|
|
|
def test_dry_run_is_reported_as_success(self):
|
|
with patch.object(reembed, "reembed_faiss", return_value=(5, 0)):
|
|
assert reembed.run("faiss", ["a"], 8, True) == 0
|
|
|
|
def test_pgvector_uses_the_pgvector_handler(self):
|
|
with patch.object(reembed, "reembed_pgvector", return_value=(1, 1)) as handler:
|
|
reembed.run("pgvector", ["a"], 8, False)
|
|
handler.assert_called_once()
|
|
|
|
|
|
class TestFaissRebuild:
|
|
@pytest.fixture
|
|
def stores(self):
|
|
"""An existing store to read from and the rebuilt one written back."""
|
|
existing = MagicMock()
|
|
existing.get_chunks.return_value = [
|
|
{"doc_id": "1", "text": "alpha", "metadata": {"i": 0}},
|
|
{"doc_id": "2", "text": "beta", "metadata": {"i": 1}},
|
|
]
|
|
rebuilt = MagicMock()
|
|
with patch.object(
|
|
reembed.VectorCreator, "create_vectorstore", side_effect=[existing, rebuilt]
|
|
) as factory:
|
|
yield existing, rebuilt, factory
|
|
|
|
def test_rebuilds_from_stored_text_and_saves(self, stores):
|
|
existing, rebuilt, factory = stores
|
|
seen, written = reembed.reembed_faiss("s1", batch_size=8, dry_run=False)
|
|
|
|
assert (seen, written) == (2, 2)
|
|
rebuilt.save_local.assert_called_once()
|
|
docs = factory.call_args_list[1].kwargs["docs_init"]
|
|
assert [d.page_content for d in docs] == ["alpha", "beta"]
|
|
assert [d.metadata for d in docs] == [{"i": 0}, {"i": 1}]
|
|
|
|
def test_embeddings_key_comes_from_settings(self, stores):
|
|
"""A placeholder here is sent as the server's bearer token."""
|
|
_, _, factory = stores
|
|
with patch.object(reembed.settings, "EMBEDDINGS_KEY", "sk-real", create=True):
|
|
reembed.reembed_faiss("s1", batch_size=8, dry_run=False)
|
|
keys = [call.kwargs.get("embeddings_key") for call in factory.call_args_list]
|
|
assert keys == ["sk-real", "sk-real"]
|
|
|
|
def test_chunk_ids_are_preserved(self, stores):
|
|
"""Re-embedding must not renumber chunks.
|
|
|
|
Fresh ids orphan every GraphRAG ``graph_node_chunks`` row for the
|
|
source and invalidate any id a client already holds.
|
|
"""
|
|
_, _, factory = stores
|
|
reembed.reembed_faiss("s1", batch_size=8, dry_run=False)
|
|
assert factory.call_args_list[1].kwargs["ids"] == ["1", "2"]
|
|
|
|
def test_existing_index_is_opened_past_the_dimension_check(self, stores):
|
|
"""A width change is the main reason to run this script.
|
|
|
|
``assert_embedding_dimensions`` refuses to open an index whose width
|
|
differs from the configured model -- and its error message recommends
|
|
this script, so without the opt-out the advice failed on every source.
|
|
The chunk text lives in the sidecar, so reading it needs no match.
|
|
"""
|
|
_, _, factory = stores
|
|
reembed.reembed_faiss("s1", batch_size=8, dry_run=False)
|
|
assert factory.call_args_list[0].kwargs["skip_dimension_check"] is True
|
|
# The rebuild writes the new width, so it must still be checked.
|
|
assert "skip_dimension_check" not in factory.call_args_list[1].kwargs
|
|
|
|
def test_batch_size_is_forwarded_to_the_rebuild(self, stores):
|
|
"""On a remote embeddings server the whole index is otherwise one POST."""
|
|
_, _, factory = stores
|
|
reembed.reembed_faiss("s1", batch_size=8, dry_run=False)
|
|
assert factory.call_args_list[1].kwargs["batch_size"] == 8
|
|
|
|
def test_existing_index_is_not_deleted(self, stores):
|
|
"""The rebuild must not destroy the old index before the new one exists."""
|
|
existing, rebuilt, _ = stores
|
|
reembed.reembed_faiss("s1", batch_size=8, dry_run=False)
|
|
existing.delete_index.assert_not_called()
|
|
|
|
def test_dry_run_reads_but_never_rebuilds(self):
|
|
existing = MagicMock()
|
|
existing.get_chunks.return_value = [{"doc_id": "1", "text": "a", "metadata": {}}]
|
|
with patch.object(
|
|
reembed.VectorCreator, "create_vectorstore", return_value=existing
|
|
) as factory:
|
|
seen, written = reembed.reembed_faiss("s1", batch_size=8, dry_run=True)
|
|
assert (seen, written) == (1, 0)
|
|
assert factory.call_count == 1, "no rebuild store may be constructed"
|
|
|
|
def test_empty_index_is_a_no_op(self):
|
|
existing = MagicMock()
|
|
existing.get_chunks.return_value = []
|
|
with patch.object(
|
|
reembed.VectorCreator, "create_vectorstore", return_value=existing
|
|
):
|
|
assert reembed.reembed_faiss("s1", batch_size=8, dry_run=False) == (0, 0)
|
|
|
|
|
|
class TestPgvectorWithoutTheExtension:
|
|
"""Mocked pgvector paths.
|
|
|
|
The live tests in ``test_reembed_pgvector_live`` skip wherever the cluster
|
|
has no pgvector build -- which includes CI -- so the SQL shape and the
|
|
batching contract are pinned here too.
|
|
"""
|
|
|
|
@pytest.fixture
|
|
def store(self):
|
|
store = MagicMock()
|
|
store._table_name = "documents"
|
|
store._vector_column = "embedding"
|
|
cursor = MagicMock()
|
|
cursor.fetchall.return_value = [(1, "alpha"), (2, "beta"), (3, "gamma")]
|
|
conn = MagicMock()
|
|
conn.cursor.return_value = cursor
|
|
store._get_connection.return_value = conn
|
|
store._embedding.embed_documents.side_effect = lambda texts: [
|
|
[0.5] * 4 for _ in texts
|
|
]
|
|
with patch.object(
|
|
reembed.VectorCreator, "create_vectorstore", return_value=store
|
|
), patch.object(reembed.settings, "GRAPHRAG_ENABLED", False):
|
|
yield store, conn, cursor
|
|
|
|
def test_reads_and_rewrites_every_chunk(self, store):
|
|
_, conn, cursor = store
|
|
seen, written = reembed.reembed_pgvector("s1", batch_size=64, dry_run=False)
|
|
assert (seen, written) == (3, 3)
|
|
cursor.executemany.assert_called_once()
|
|
conn.commit.assert_called()
|
|
|
|
def test_dry_run_neither_embeds_nor_writes(self, store):
|
|
fake_store, conn, cursor = store
|
|
seen, written = reembed.reembed_pgvector("s1", batch_size=64, dry_run=True)
|
|
assert (seen, written) == (3, 0)
|
|
fake_store._embedding.embed_documents.assert_not_called()
|
|
cursor.executemany.assert_not_called()
|
|
|
|
def test_batches_commit_separately(self, store):
|
|
_, conn, cursor = store
|
|
reembed.reembed_pgvector("s1", batch_size=2, dry_run=False)
|
|
# 3 rows at batch 2 is two write transactions.
|
|
assert cursor.executemany.call_count == 2
|
|
assert conn.commit.call_count == 2
|
|
|
|
def test_failed_batch_rolls_back_and_raises(self, store):
|
|
fake_store, conn, cursor = store
|
|
cursor.executemany.side_effect = RuntimeError("write failed")
|
|
with pytest.raises(RuntimeError):
|
|
reembed.reembed_pgvector("s1", batch_size=64, dry_run=False)
|
|
conn.rollback.assert_called_once()
|
|
|
|
def test_connection_is_returned_even_on_failure(self, store):
|
|
fake_store, _, cursor = store
|
|
cursor.executemany.side_effect = RuntimeError("boom")
|
|
with pytest.raises(RuntimeError):
|
|
reembed.reembed_pgvector("s1", batch_size=64, dry_run=False)
|
|
fake_store.close.assert_called_once()
|
|
|
|
def test_null_text_does_not_crash_the_embed_call(self, store):
|
|
fake_store, _, cursor = store
|
|
cursor.fetchall.return_value = [(1, None), (2, "beta")]
|
|
seen, written = reembed.reembed_pgvector("s1", batch_size=64, dry_run=False)
|
|
assert (seen, written) == (2, 2)
|
|
assert fake_store._embedding.embed_documents.call_args.args[0] == ["", "beta"]
|
|
|
|
def test_empty_source_is_a_no_op(self, store):
|
|
fake_store, _, cursor = store
|
|
cursor.fetchall.return_value = []
|
|
assert reembed.reembed_pgvector("s1", batch_size=64, dry_run=False) == (0, 0)
|
|
fake_store._embedding.embed_documents.assert_not_called()
|
|
|
|
def test_source_discovery_returns_sorted_ids(self, store):
|
|
_, _, cursor = store
|
|
cursor.fetchall.return_value = [("b",), ("a",)]
|
|
assert reembed.list_source_ids("pgvector") == ["b", "a"]
|
|
|
|
|
|
class TestFaissSourceDiscovery:
|
|
def test_source_ids_come_from_index_directories(self):
|
|
storage = MagicMock()
|
|
storage.list_files.return_value = [
|
|
"indexes/src-a/index.faiss",
|
|
"indexes/src-a/index.pkl",
|
|
"indexes/src-b/index.faiss",
|
|
]
|
|
with patch(
|
|
"application.storage.storage_creator.StorageCreator.get_storage",
|
|
return_value=storage,
|
|
):
|
|
assert reembed.list_source_ids("faiss") == ["src-a", "src-b"]
|
|
|
|
def test_storage_failure_is_reported_as_a_usable_error(self):
|
|
storage = MagicMock()
|
|
storage.list_files.side_effect = OSError("permission denied")
|
|
with patch(
|
|
"application.storage.storage_creator.StorageCreator.get_storage",
|
|
return_value=storage,
|
|
):
|
|
with pytest.raises(reembed.ReembedError, match="permission denied"):
|
|
reembed.list_source_ids("faiss")
|
|
|
|
def test_unsupported_store_error_names_the_alternatives(self):
|
|
with patch.object(reembed.settings, "VECTOR_STORE", "milvus", create=True):
|
|
assert reembed.main([]) == 2
|
|
|
|
|
|
class TestGraphNodeReembedding:
|
|
"""``graph_nodes.name_embedding`` seeds every graph traversal.
|
|
|
|
It is written once at extraction time and never revisited, so rewriting
|
|
only the chunk table leaves the graph seeding from the previous model --
|
|
and since mpnet and granite-311m are both 768-dimensional, the column
|
|
accepts the mismatch and nothing reports it.
|
|
"""
|
|
|
|
@pytest.fixture
|
|
def graph(self):
|
|
store = MagicMock()
|
|
store._embedding.embed_documents.side_effect = lambda texts: [
|
|
[0.5] * 4 for _ in texts
|
|
]
|
|
cursor = MagicMock()
|
|
cursor.fetchone.return_value = ("graph_nodes",)
|
|
cursor.fetchall.return_value = [
|
|
("n1", "Alpha"),
|
|
("n2", "Beta"),
|
|
("n3", "Gamma"),
|
|
]
|
|
conn = MagicMock()
|
|
conn.cursor.return_value = cursor
|
|
return store, conn, cursor
|
|
|
|
def test_rewrites_every_node_name(self, graph):
|
|
store, conn, cursor = graph
|
|
written = reembed.reembed_graph_nodes(
|
|
store, conn, "s1", batch_size=64, dry_run=False
|
|
)
|
|
assert written == 3
|
|
store._embedding.embed_documents.assert_called_once_with(
|
|
["Alpha", "Beta", "Gamma"]
|
|
)
|
|
statement = cursor.executemany.call_args.args[0]
|
|
assert "graph_nodes" in statement and "name_embedding" in statement
|
|
conn.commit.assert_called()
|
|
|
|
def test_dry_run_counts_without_embedding(self, graph):
|
|
store, conn, cursor = graph
|
|
assert reembed.reembed_graph_nodes(store, conn, "s1", 64, dry_run=True) == 3
|
|
store._embedding.embed_documents.assert_not_called()
|
|
cursor.executemany.assert_not_called()
|
|
|
|
def test_missing_table_is_a_no_op(self, graph):
|
|
store, conn, cursor = graph
|
|
cursor.fetchone.return_value = (None,)
|
|
assert reembed.reembed_graph_nodes(store, conn, "s1", 64, dry_run=False) == 0
|
|
store._embedding.embed_documents.assert_not_called()
|
|
|
|
def test_batches_commit_separately(self, graph):
|
|
store, conn, cursor = graph
|
|
reembed.reembed_graph_nodes(store, conn, "s1", batch_size=2, dry_run=False)
|
|
assert cursor.executemany.call_count == 2
|
|
assert conn.commit.call_count == 2
|
|
|
|
def test_failed_batch_rolls_back_and_raises(self, graph):
|
|
store, conn, cursor = graph
|
|
cursor.executemany.side_effect = RuntimeError("write failed")
|
|
with pytest.raises(RuntimeError, match="write failed"):
|
|
reembed.reembed_graph_nodes(store, conn, "s1", 64, dry_run=False)
|
|
conn.rollback.assert_called_once()
|
|
|
|
def test_pgvector_run_skips_the_graph_when_disabled(self):
|
|
store, conn, cursor = self._pgvector_mocks()
|
|
with patch.object(
|
|
reembed.VectorCreator, "create_vectorstore", return_value=store
|
|
), patch.object(reembed.settings, "GRAPHRAG_ENABLED", False):
|
|
reembed.reembed_pgvector("s1", batch_size=64, dry_run=False)
|
|
assert not self._graph_statements(cursor)
|
|
|
|
def test_pgvector_run_reembeds_the_graph_when_enabled(self):
|
|
store, conn, cursor = self._pgvector_mocks()
|
|
with patch.object(
|
|
reembed.VectorCreator, "create_vectorstore", return_value=store
|
|
), patch.object(reembed.settings, "GRAPHRAG_ENABLED", True):
|
|
reembed.reembed_pgvector("s1", batch_size=64, dry_run=False)
|
|
assert self._graph_statements(cursor)
|
|
|
|
@staticmethod
|
|
def _pgvector_mocks():
|
|
store = MagicMock()
|
|
store._table_name = "documents"
|
|
store._vector_column = "embedding"
|
|
cursor = MagicMock()
|
|
cursor.fetchone.return_value = ("graph_nodes",)
|
|
cursor.fetchall.return_value = [(1, "alpha"), (2, "beta")]
|
|
conn = MagicMock()
|
|
conn.cursor.return_value = cursor
|
|
store._get_connection.return_value = conn
|
|
store._embedding.embed_documents.side_effect = lambda texts: [
|
|
[0.5] * 4 for _ in texts
|
|
]
|
|
return store, conn, cursor
|
|
|
|
@staticmethod
|
|
def _graph_statements(cursor):
|
|
return [
|
|
call
|
|
for call in cursor.executemany.call_args_list
|
|
if "graph_nodes" in str(call.args[0])
|
|
]
|
|
|
|
|
|
class TestEmbedsInProcess:
|
|
"""A batch job should not round-trip every chunk through the broker."""
|
|
|
|
def test_delegation_is_turned_off_for_the_run(self, monkeypatch):
|
|
from application.core.settings import settings
|
|
|
|
monkeypatch.setattr(settings, "EMBEDDINGS_DELEGATE_TO_WORKER", True, raising=False)
|
|
monkeypatch.setattr(settings, "VECTOR_STORE", "pgvector", raising=False)
|
|
seen = {}
|
|
with patch.object(reembed, "run", side_effect=lambda *a, **k: seen.setdefault(
|
|
"delegating", settings.EMBEDDINGS_DELEGATE_TO_WORKER
|
|
) or 0):
|
|
assert reembed.main(["--dry-run"]) == 0
|
|
assert seen["delegating"] is False
|
|
|
|
|
|
class TestRecordsTheModel:
|
|
"""``sources.model`` is what the boot mismatch check reads."""
|
|
|
|
def test_a_re_embedded_source_is_stamped(self, monkeypatch):
|
|
from application.core.settings import settings
|
|
|
|
monkeypatch.setattr(settings, "EMBEDDINGS_NAME", "new/model", raising=False)
|
|
conn = MagicMock()
|
|
session = MagicMock()
|
|
session.__enter__ = MagicMock(return_value=conn)
|
|
session.__exit__ = MagicMock(return_value=False)
|
|
with patch("application.storage.db.session.db_session", return_value=session):
|
|
reembed.record_source_model("src-1")
|
|
params = conn.execute.call_args.args[1]
|
|
assert params == {"model": "new/model", "id": "src-1"}
|
|
|
|
def test_a_dry_run_stamps_nothing(self, monkeypatch):
|
|
with patch.object(reembed, "record_source_model") as record, patch.object(
|
|
reembed, "list_source_ids", return_value=["a"]
|
|
), patch.object(reembed, "reembed_pgvector", return_value=(3, 0)):
|
|
reembed.run("pgvector", None, 64, True)
|
|
record.assert_not_called()
|
|
|
|
def test_a_real_run_stamps_each_source(self, monkeypatch):
|
|
with patch.object(reembed, "record_source_model") as record, patch.object(
|
|
reembed, "list_source_ids", return_value=["a", "b"]
|
|
), patch.object(reembed, "reembed_pgvector", return_value=(3, 3)):
|
|
reembed.run("pgvector", None, 64, False)
|
|
assert [c.args[0] for c in record.call_args_list] == ["a", "b"]
|
|
|
|
def test_a_failed_source_is_not_stamped(self):
|
|
with patch.object(reembed, "record_source_model") as record, patch.object(
|
|
reembed, "list_source_ids", return_value=["a"]
|
|
), patch.object(reembed, "reembed_pgvector", side_effect=RuntimeError("boom")):
|
|
assert reembed.run("pgvector", None, 64, False) == 1
|
|
record.assert_not_called()
|
|
|
|
|
|
class TestThePinIsResolved:
|
|
"""The script must embed with the model the installation is pinned to.
|
|
|
|
``resolve_embeddings_pin`` runs in ``application.app``, which this script
|
|
never imports. An install pinned in ``app_metadata`` with no
|
|
``EMBEDDINGS_NAME`` in the environment -- every stock Kubernetes
|
|
deployment, whose manifests carry no embedding config -- would otherwise
|
|
rewrite its whole index with the legacy code default and stamp
|
|
``sources.model`` to match, creating the cross-model index this script
|
|
exists to repair.
|
|
"""
|
|
|
|
def test_main_resolves_the_pin_before_reading_the_store(self):
|
|
order = []
|
|
with patch(
|
|
"application.storage.db.embeddings_pin.resolve_embeddings_pin",
|
|
side_effect=lambda *a, **k: order.append("pin"),
|
|
), patch.object(reembed.settings, "VECTOR_STORE", "pgvector", create=True), patch.object(
|
|
reembed, "run", side_effect=lambda *a, **k: (order.append("run"), 0)[1]
|
|
):
|
|
assert reembed.main([]) == 0
|
|
assert order == ["pin", "run"], "the pin must resolve before anything embeds"
|
|
|
|
def test_an_unsupported_store_still_resolved_the_pin_first(self):
|
|
with patch(
|
|
"application.storage.db.embeddings_pin.resolve_embeddings_pin"
|
|
) as pin, patch.object(reembed.settings, "VECTOR_STORE", "qdrant", create=True):
|
|
assert reembed.main([]) == 2
|
|
pin.assert_called_once()
|