add fastapi rest api and celery workers

This commit is contained in:
qananasikq
2026-04-08 22:54:33 +03:00
parent 6e97991c68
commit f4b9d20191
15 changed files with 982 additions and 0 deletions

View File

@@ -0,0 +1,3 @@
from .app import create_app
__all__ = ["create_app"]

41
iaai_scraper/api/app.py Normal file
View File

@@ -0,0 +1,41 @@
"""FastAPI app."""
from contextlib import asynccontextmanager
from fastapi import FastAPI
from ..core.config import Settings
from ..storage.db import PersistenceService
from .routes import cars, health, tasks
@asynccontextmanager
async def lifespan(app: FastAPI):
"""Жизненный цикл."""
persistence: PersistenceService = app.state.persistence
persistence.create_tables()
yield
def create_app(settings: Settings | None = None) -> FastAPI:
_settings = settings or Settings()
app = FastAPI(
title="IAAI Scraper API",
description="REST API для управления задачами скрапинга IAAI и просмотра данных",
version="1.0.0",
lifespan=lifespan,
)
app.state.settings = _settings
app.state.persistence = PersistenceService(_settings)
app.include_router(health.router, tags=["health"])
app.include_router(cars.router, prefix="/api/v1", tags=["cars"])
app.include_router(tasks.router, prefix="/api/v1", tags=["tasks"])
return app
# Для запуска через uvicorn.
app = create_app()

9
iaai_scraper/api/deps.py Normal file
View File

@@ -0,0 +1,9 @@
"""FastAPI deps."""
from fastapi import Request
from ..storage.db import PersistenceService
def get_persistence(request: Request) -> PersistenceService:
return request.app.state.persistence

View File

View File

@@ -0,0 +1,144 @@
"""Cars endpoints."""
from fastapi import APIRouter, Depends, HTTPException, Query
from sqlalchemy import func, select
from ..deps import get_persistence
from ...storage.db import PersistenceService
from ...storage.models import Car, Image
router = APIRouter()
@router.get("/cars")
def list_cars(
page: int = Query(1, ge=1),
per_page: int = Query(20, ge=1, le=100),
brand: str | None = None,
model: str | None = None,
year_min: int | None = None,
year_max: int | None = None,
is_sold: bool | None = None,
persistence: PersistenceService = Depends(get_persistence),
):
"""Список автомобилей с пагинацией и фильтрами."""
with persistence.session_scope() as session:
query = select(Car)
if brand:
query = query.where(Car.brand.ilike(f"%{brand}%"))
if model:
query = query.where(Car.model.ilike(f"%{model}%"))
if year_min is not None:
query = query.where(Car.year >= year_min)
if year_max is not None:
query = query.where(Car.year <= year_max)
if is_sold is not None:
query = query.where(Car.is_sold == is_sold)
# Общее количество.
count_query = select(func.count()).select_from(query.subquery())
total = session.execute(count_query).scalar() or 0
# Пагинация.
offset = (page - 1) * per_page
cars = session.execute(
query.order_by(Car.last_seen_at.desc()).offset(offset).limit(per_page)
).scalars().all()
return {
"total": total,
"page": page,
"per_page": per_page,
"pages": (total + per_page - 1) // per_page if per_page else 0,
"items": [_car_to_dict(car) for car in cars],
}
@router.get("/cars/{car_id}")
def get_car(
car_id: int,
persistence: PersistenceService = Depends(get_persistence),
):
"""Детальная информация об автомобиле с изображениями."""
with persistence.session_scope() as session:
car = session.get(Car, car_id)
if car is None:
raise HTTPException(status_code=404, detail="Car not found")
return _car_to_dict(car, include_images=True)
@router.get("/cars/by-origin/{origin_id}")
def get_car_by_origin(
origin_id: str,
persistence: PersistenceService = Depends(get_persistence),
):
"""Поиск автомобиля по origin_id."""
with persistence.session_scope() as session:
car = session.execute(
select(Car).where(Car.origin_id == origin_id)
).scalars().first()
if car is None:
raise HTTPException(status_code=404, detail="Car not found")
return _car_to_dict(car, include_images=True)
@router.get("/stats")
def get_stats(persistence: PersistenceService = Depends(get_persistence)):
"""Общая статистика по БД."""
with persistence.session_scope() as session:
total_cars = session.execute(select(func.count(Car.id))).scalar() or 0
total_images = session.execute(select(func.count(Image.id))).scalar() or 0
brands = session.execute(
select(Car.brand, func.count(Car.id))
.group_by(Car.brand)
.order_by(func.count(Car.id).desc())
.limit(20)
).all()
return {
"total_cars": total_cars,
"total_images": total_images,
"top_brands": [{"brand": b, "count": c} for b, c in brands],
}
def _car_to_dict(car: Car, include_images: bool = False) -> dict:
"""Сериализация Car в dict."""
result = {
"id": car.id,
"parser_id": car.parser_id,
"brand": car.brand,
"model": car.model,
"year": car.year,
"price": car.price,
"currency": car.currency,
"mileage": car.mileage,
"country": car.country,
"is_sold": car.is_sold,
"color": car.color,
"drive": car.drive,
"gearbox": car.gearbox,
"steering_wheel": car.steering_wheel,
"body_type": car.body_type,
"engine_volume": car.engine_volume,
"selling_type": car.selling_type,
"origin": car.origin,
"origin_url": car.origin_url,
"origin_id": car.origin_id,
"is_damaged": car.is_damaged,
"slug": car.slug,
"last_seen_at": car.last_seen_at.isoformat() if car.last_seen_at else None,
"content_hash": car.content_hash,
}
if include_images:
result["images"] = [
{
"id": img.id,
"fullres_image": img.fullres_image,
"preview_image": img.preview_image,
"order_index": img.order_index,
}
for img in sorted(car.images, key=lambda i: i.order_index)
]
return result

View File

@@ -0,0 +1,27 @@
"""Health endpoint."""
from fastapi import APIRouter, Depends
from sqlalchemy import text
from ..deps import get_persistence
from ...storage.db import PersistenceService
router = APIRouter()
@router.get("/health")
def health_check(persistence: PersistenceService = Depends(get_persistence)):
"""Проверка API и БД."""
db_ok = False
try:
with persistence.session_scope() as session:
session.execute(text("SELECT 1"))
db_ok = True
except Exception:
pass
return {
"status": "ok" if db_ok else "degraded",
"service": "iaai-scraper-api",
"database": "connected" if db_ok else "unavailable",
}

View File

@@ -0,0 +1,174 @@
"""Task endpoints."""
from datetime import datetime, timezone
from fastapi import APIRouter, Depends, HTTPException, Query
from pydantic import BaseModel
from sqlalchemy import select, func
from ..deps import get_persistence
from ...storage.db import PersistenceService
from ...storage.models import ScrapeTask, SyncRun
from ...worker.tasks import sync_vehicle_task, sync_listing_task
router = APIRouter()
# --- Схемы.
class SyncVehicleRequest(BaseModel):
vehicle_url: str
lane: str = "iaai"
class SyncListingRequest(BaseModel):
make: str | None = None
model: str | None = None
lane: str = "iaai_cars"
limit: int | None = None
only_new: bool | None = None
# --- Эндпоинты.
@router.post("/tasks/sync-vehicle")
def start_sync_vehicle(
body: SyncVehicleRequest,
persistence: PersistenceService = Depends(get_persistence),
):
"""Запустить задачу скрапинга одного автомобиля через Celery."""
result = sync_vehicle_task.delay(body.vehicle_url, lane=body.lane)
# Регистрируем задачу в БД.
persistence.create_scrape_task(
celery_task_id=result.id,
task_type="sync_vehicle",
vehicle_url=body.vehicle_url,
)
return {
"task_id": result.id,
"status": "queued",
"vehicle_url": body.vehicle_url,
}
@router.post("/tasks/sync-listing")
def start_sync_listing(
body: SyncListingRequest,
persistence: PersistenceService = Depends(get_persistence),
):
"""Запустить задачу полного цикла листинга через Celery."""
result = sync_listing_task.delay(
make=body.make,
model=body.model,
lane=body.lane,
limit=body.limit,
only_new=body.only_new,
)
persistence.create_scrape_task(
celery_task_id=result.id,
task_type="sync_listing",
)
return {
"task_id": result.id,
"status": "queued",
}
@router.get("/tasks/{task_id}")
def get_task_status(
task_id: str,
persistence: PersistenceService = Depends(get_persistence),
):
"""Статус Celery-задачи."""
task_info = persistence.get_scrape_task(task_id)
if task_info is None:
raise HTTPException(status_code=404, detail="Task not found")
return task_info
@router.get("/tasks")
def list_tasks(
page: int = Query(1, ge=1),
per_page: int = Query(20, ge=1, le=100),
status: str | None = None,
task_type: str | None = None,
persistence: PersistenceService = Depends(get_persistence),
):
"""Список задач с пагинацией."""
with persistence.session_scope() as session:
query = select(ScrapeTask)
if status:
query = query.where(ScrapeTask.status == status)
if task_type:
query = query.where(ScrapeTask.task_type == task_type)
total = session.execute(
select(func.count()).select_from(query.subquery())
).scalar() or 0
offset = (page - 1) * per_page
tasks = session.execute(
query.order_by(ScrapeTask.created_at.desc()).offset(offset).limit(per_page)
).scalars().all()
return {
"total": total,
"page": page,
"per_page": per_page,
"items": [
{
"id": t.id,
"celery_task_id": t.celery_task_id,
"task_type": t.task_type,
"status": t.status,
"vehicle_url": t.vehicle_url,
"created_at": t.created_at.isoformat() if t.created_at else None,
"started_at": t.started_at.isoformat() if t.started_at else None,
"finished_at": t.finished_at.isoformat() if t.finished_at else None,
"result_summary": t.result_summary,
"error_message": t.error_message,
}
for t in tasks
],
}
@router.get("/sync-runs")
def list_sync_runs(
page: int = Query(1, ge=1),
per_page: int = Query(20, ge=1, le=100),
persistence: PersistenceService = Depends(get_persistence),
):
"""История запусков синхронизации."""
with persistence.session_scope() as session:
total = session.execute(select(func.count(SyncRun.id))).scalar() or 0
offset = (page - 1) * per_page
runs = session.execute(
select(SyncRun).order_by(SyncRun.started_at.desc()).offset(offset).limit(per_page)
).scalars().all()
return {
"total": total,
"page": page,
"per_page": per_page,
"items": [
{
"id": r.id,
"started_at": r.started_at.isoformat() if r.started_at else None,
"finished_at": r.finished_at.isoformat() if r.finished_at else None,
"status": r.status,
"lane": r.lane,
"ids_fetched": r.ids_fetched,
"cars_upserted": r.cars_upserted,
"cars_failed": r.cars_failed,
"images_upserted": r.images_upserted,
"error_summary": r.error_summary,
}
for r in runs
],
}