refactor(test): share the Qdrant point-seeding helper
Why: - The helper turning a SeedSpec into an upsertable ChunkPoint lived inside one integration test. A second suite now needs to seed real Qdrant the same way, and a copy would let the two drift. Changes: - Move `_chunk_point` into tests/support/point_contract.py as `chunk_point_for`, beside the `build_point` read model it derives its payload from. Impact: - Pure move. No behaviour change; enables the API suite that follows. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This commit is contained in:
@@ -18,18 +18,12 @@ from collections.abc import Awaitable, Callable
|
||||
import pytest
|
||||
from qdrant_client import AsyncQdrantClient, models
|
||||
|
||||
from src.application.ingestion.models import SparseVector
|
||||
from src.application.points.models import ChunkPoint
|
||||
from src.config import QdrantSettings
|
||||
from src.infrastructure.qdrant.collection import (
|
||||
DENSE_NOMIC_DIMENSIONS,
|
||||
DENSE_OPENAI_DIMENSIONS,
|
||||
ensure_chunks_collection,
|
||||
)
|
||||
from src.infrastructure.qdrant.collection import ensure_chunks_collection
|
||||
from src.infrastructure.qdrant.point_repository import QdrantPointRepository
|
||||
from src.infrastructure.qdrant.points import QdrantPointStorage
|
||||
from tests.support import point_contract
|
||||
from tests.support.point_contract import SeedSpec, build_point, seed_specs
|
||||
from tests.support.point_contract import chunk_point_for, seed_specs
|
||||
|
||||
pytestmark = [
|
||||
pytest.mark.integration,
|
||||
@@ -54,33 +48,13 @@ IDS = {
|
||||
}
|
||||
|
||||
|
||||
def _chunk_point(spec: SeedSpec) -> ChunkPoint:
|
||||
"""The contract's `Point` re-expressed as something upsertable.
|
||||
|
||||
`build_point` produces the read model; Qdrant needs vectors and a flat
|
||||
payload dict, so the payload is taken straight off the model to guarantee
|
||||
the two representations cannot drift apart.
|
||||
"""
|
||||
point = build_point(spec)
|
||||
payload = point.model_dump(mode="json", exclude={"point_id", "vectors"})
|
||||
return ChunkPoint(
|
||||
point_id=point.point_id,
|
||||
dense={
|
||||
"dense_nomic": [0.1] * DENSE_NOMIC_DIMENSIONS,
|
||||
"dense_openai": [0.2] * DENSE_OPENAI_DIMENSIONS,
|
||||
},
|
||||
sparse=SparseVector(indices=[1, 2], values=[0.5, 0.25]),
|
||||
payload=payload,
|
||||
)
|
||||
|
||||
|
||||
async def _seeded(
|
||||
qdrant_client: AsyncQdrantClient, qdrant_settings: QdrantSettings
|
||||
) -> QdrantPointRepository:
|
||||
await ensure_chunks_collection(qdrant_client, collection=qdrant_settings.collection)
|
||||
storage = QdrantPointStorage(qdrant_client, collection=qdrant_settings.collection)
|
||||
await storage.upsert_points(
|
||||
[_chunk_point(spec) for spec in seed_specs(TENANT_A, TENANT_B, FILE_A, FILE_B, FILE_C)]
|
||||
[chunk_point_for(spec) for spec in seed_specs(TENANT_A, TENANT_B, FILE_A, FILE_B, FILE_C)]
|
||||
)
|
||||
# Payload indexes are built asynchronously; without waiting, a full-text or
|
||||
# ordered query can run against a half-built index and return short.
|
||||
|
||||
@@ -21,8 +21,14 @@ from dataclasses import dataclass
|
||||
from datetime import UTC, datetime
|
||||
|
||||
from src.application.ingestion.chunking import chunk_id_for
|
||||
from src.application.ingestion.models import SparseVector
|
||||
from src.application.points.models import ChunkPoint
|
||||
from src.application.points.point import Point
|
||||
from src.application.ports.point_repository import PayloadPatch, PointRepository
|
||||
from src.infrastructure.qdrant.collection import (
|
||||
DENSE_NOMIC_DIMENSIONS,
|
||||
DENSE_OPENAI_DIMENSIONS,
|
||||
)
|
||||
|
||||
# Two tenants and two files, so every scenario can assert both that the right
|
||||
# rows come back and that the wrong ones do not.
|
||||
@@ -76,6 +82,26 @@ def build_point(spec: SeedSpec) -> Point:
|
||||
)
|
||||
|
||||
|
||||
def chunk_point_for(spec: SeedSpec) -> ChunkPoint:
|
||||
"""A `SeedSpec` as something upsertable into real Qdrant.
|
||||
|
||||
The payload is taken straight off `build_point`'s read model rather than
|
||||
hand-written, so the write shape and the read shape cannot drift apart. The
|
||||
vectors are constant filler: nothing in plan 002's read paths scores by
|
||||
similarity, so their values are irrelevant and their dimensions are not.
|
||||
"""
|
||||
point = build_point(spec)
|
||||
return ChunkPoint(
|
||||
point_id=point.point_id,
|
||||
dense={
|
||||
"dense_nomic": [0.1] * DENSE_NOMIC_DIMENSIONS,
|
||||
"dense_openai": [0.2] * DENSE_OPENAI_DIMENSIONS,
|
||||
},
|
||||
sparse=SparseVector(indices=[1, 2], values=[0.5, 0.25]),
|
||||
payload=point.model_dump(mode="json", exclude={"point_id", "vectors"}),
|
||||
)
|
||||
|
||||
|
||||
def seed_specs(
|
||||
tenant_a: uuid.UUID,
|
||||
tenant_b: uuid.UUID,
|
||||
|
||||
Reference in New Issue
Block a user