Codebase intelligence for AI and humans: code health scores, auto-generated docs, git analytics, dead code detection, and architectural decisions via MCP.
Python aicode-complexitycode-healthcode-intelligencecode-qualitydead-codedeveloper-toolsdocumentationgit-analyticsmcpopen-sourcepythonrefactoringstatic-analysistechnical-debt
"""``graph_edges`` upserts must keep the max confidence on collision.
A node pair can carry several resolved calls of differing confidence. The
in-memory resolver keeps the highest (``_resolvers.py:504-505``); the persistence
upsert historically did last-write, so a later low-confidence write could stamp a
real call below ``_FLOW_CALLS_CONF_FLOOR`` (0.5) and drop it from flow-path
answers. The upsert must mirror the resolver and keep the max.
"""
from __future__ import annotations
from sqlalchemy import select
from repowise.core.persistence import batch_upsert_graph_edges
from repowise.core.persistence.models import GraphEdge
from tests.unit.persistence.helpers import insert_repo
async def _edge_conf(session, repo_id: str) -> float:
row = (
await session.execute(select(GraphEdge).where(GraphEdge.repository_id == repo_id))
).scalar_one()
return row.confidence
async def test_edge_upsert_keeps_max_confidence(async_session, tmp_path):
repo = await insert_repo(async_session)
edge = {"source_node_id": "a.py::f", "target_node_id": "b.py::g", "edge_type": "calls"}
await batch_upsert_graph_edges(async_session, repo.id, [{**edge, "confidence": 0.9}])
await async_session.commit()
assert await _edge_conf(async_session, repo.id) == 0.9
# A later, lower-confidence write for the same pair must not lower it below
# the flow-path floor.
await batch_upsert_graph_edges(async_session, repo.id, [{**edge, "confidence": 0.3}])
await async_session.commit()
assert await _edge_conf(async_session, repo.id) == 0.9
# A higher-confidence write still wins.
await batch_upsert_graph_edges(async_session, repo.id, [{**edge, "confidence": 0.95}])
await async_session.commit()
assert await _edge_conf(async_session, repo.id) == 0.95