pryscraper/routers/reconciliation.py
cryptorugmunch 8b52f14774
Some checks failed
CI / Secret scan (gitleaks) (push) Successful in 34s
CI / Security audit (bandit) (push) Successful in 35s
CI / test (push) Successful in 1m23s
CI / lint (push) Failing after 49s
CI / typecheck (push) Successful in 55s
feat(pry): phase 0 — split routers, add tests, apify schema, pry api key (#5)
2026-07-03 03:43:02 +02:00

71 lines
2.1 KiB
Python

"""Pry — Reconciliation router (remaining api.py routes).
Auto-extracted from api.py during the router-split refactor.
"""
# SPDX-License-Identifier: MIT
# Copyright (c) 2026 Rug Munch Media LLC
from __future__ import annotations
import logging
from typing import Any
from fastapi import APIRouter, Body
from errors import InvalidRequestError
logger = logging.getLogger(__name__)
router = APIRouter(tags=["Reconciliation"])
@router.post(
"/v1/reconcile",
tags=["Reconciliation"],
summary="Reconcile records from multiple sources into unified entities",
)
async def reconcile_endpoint(
records: list[dict[str, Any]] = Body(...),
vertical: str = Body("product"),
threshold: float = Body(0.7),
) -> dict[str, Any]:
"""Reconcile records from multiple sources into matched entities.
Matches records across sources using identity field similarity,
normalizes to a unified vertical schema, and returns entity groups
with confidence scores.
Verticals: product, job, real_estate, review
"""
from reconciliation import VERTICAL_SCHEMAS, reconcile
if vertical not in VERTICAL_SCHEMAS:
raise InvalidRequestError(
f"Unknown vertical: {vertical}. Supported: {list(VERTICAL_SCHEMAS.keys())}"
)
result = await reconcile(records, vertical, threshold)
return {"success": True, "data": result}
@router.get(
"/v1/reconcile/schemas",
tags=["Reconciliation"],
summary="List supported reconciliation schemas",
)
async def list_schemas() -> dict[str, Any]:
"""List all supported vertical schemas for entity reconciliation."""
from reconciliation import VERTICAL_SCHEMAS
schemas = {}
for key, schema in VERTICAL_SCHEMAS.items():
schemas[key] = {
"name": schema["name"],
"fields": {
k: {fk: fv for fk, fv in v.items() if fk != "type"}
for k, v in schema["fields"].items()
},
"identity_fields": schema["identity_fields"],
}
return {"success": True, "data": {"schemas": schemas, "total": len(schemas)}}