Files
rf4-spotter/apps/api/app/main.py
T
ik 870d9cc7f9
CI / backend-and-migrations (push) Canceled after 0s
CI / astro-build (push) Canceled after 0s
CI / compose-e2e (push) Canceled after 0s
feat: harden production data and backups
2026-09-06 14:08:07 +07:00

385 lines
21 KiB
Python

from __future__ import annotations
from collections import Counter
from datetime import datetime, timedelta, timezone
import hashlib
import hmac
import logging
import secrets
import time as time_module
from typing import Annotated, Literal
from uuid import UUID
import uuid
import httpx
from fastapi import Depends, FastAPI, File, Header, HTTPException, Query, Request, Response, UploadFile
from fastapi.middleware.cors import CORSMiddleware
from fastapi.responses import JSONResponse
from sqlalchemy import delete, func, select, text
from sqlalchemy.orm import Session, joinedload
from .activity import activity_rows
from .database import get_session
from .config import settings
from .community_review import ExternalReviewError, map_observation, publish_observation, reject_observation
from .importer import ImportSourceError, import_records, normalize
from .logging_config import configure_logging
from .models import Bait, BaitKind, CatchReport, ExternalObservation, Fish, ModerationEvent, ModerationStatus, OfficialRecordImport, SourceType, Spot, SubmissionAttempt, Waterbody
from .readiness import readiness_report
from .schemas import ActivityOut, AdminCatchReportOut, BaitOut, CatchOut, CatchReportAccepted, CatchReportCreate, CatchReportCreated, ExternalObservationDecision, ExternalObservationMapping, ExternalObservationOut, ExternalObservationPublished, FishOut, ImportRunOut, ModerationUpdate, OfficialRecordOut, SpotOut, WaterbodyOut
from .storage import ScreenshotError, client as storage_client, delete_screenshot, signed_screenshot_url, upload_screenshot
configure_logging(settings.log_level)
logger = logging.getLogger("rf4.api")
app = FastAPI(title="RF4 Spotter API", version="0.1.0")
app.add_middleware(
CORSMiddleware,
allow_origins=settings.cors_origins,
allow_methods=["GET", "POST", "PATCH", "DELETE"],
allow_headers=["Authorization", "Content-Type"],
)
Db = Annotated[Session, Depends(get_session)]
@app.middleware("http")
async def structured_request_log(request: Request, call_next):
request_id = uuid.uuid4().hex
started = time_module.perf_counter()
status_code = 500
try:
response = await call_next(request)
status_code = response.status_code
response.headers["X-Request-ID"] = request_id
response.headers["X-Content-Type-Options"] = "nosniff"
response.headers["Referrer-Policy"] = "strict-origin-when-cross-origin"
response.headers["Permissions-Policy"] = "camera=(), microphone=(), geolocation=()"
if settings.deployment_environment == "production":
response.headers["Strict-Transport-Security"] = "max-age=31536000; includeSubDomains"
return response
except Exception as exc:
logger.error("request failed", extra={"request_id": request_id, "error_type": type(exc).__name__})
raise
finally:
logger.log(
logging.DEBUG if request.url.path in {"/health", "/ready"} else logging.INFO,
"request completed",
extra={
"request_id": request_id, "method": request.method,
"path": request.url.path, "status_code": status_code,
"duration_ms": round((time_module.perf_counter() - started) * 1000, 2),
},
)
@app.get("/health")
def health() -> dict[str, str]:
return {"status": "ok"}
@app.get("/ready")
def ready(db: Db) -> JSONResponse:
is_ready, components = readiness_report(
db, storage_client(), import_required=settings.official_import_required,
import_interval_seconds=settings.import_interval_seconds,
)
return JSONResponse(
status_code=200 if is_ready else 503,
content={"status": "ready" if is_ready else "not_ready", "components": components},
)
@app.get("/api/v1/fishes", response_model=list[FishOut])
def fishes(db: Db) -> list[Fish]:
return list(db.scalars(select(Fish).order_by(Fish.name_ru)))
@app.get("/api/v1/waterbodies", response_model=list[WaterbodyOut])
def waterbodies(db: Db) -> list[Waterbody]:
return list(db.scalars(select(Waterbody).order_by(Waterbody.name_ru)))
@app.get("/api/v1/baits", response_model=list[BaitOut])
def baits(db: Db) -> list[Bait]:
return list(db.scalars(select(Bait).order_by(Bait.name)))
@app.get("/api/v1/activity", response_model=list[ActivityOut])
def activity(
db: Db, hours: int = Query(24),
waterbody: str | None = None, fish: str | None = None,
method: str | None = None,
sort: Literal["activity", "confidence", "freshness"] = "activity",
limit: int = Query(20, ge=1, le=100), offset: int = Query(0, ge=0),
) -> list[ActivityOut]:
if hours not in {6, 12, 24, 72}:
raise HTTPException(status_code=422, detail="hours must be one of: 6, 12, 24, 72")
rows = activity_rows(db, hours=hours, waterbody=waterbody, fish=fish, method=method)
keys = {"activity": lambda r: r.activity_score, "confidence": lambda r: r.confidence_score, "freshness": lambda r: r.last_confirmed_at}
rows.sort(key=keys[sort], reverse=True)
return rows[offset:offset + limit]
def _spot_or_404(db: Session, spot_id: UUID) -> Spot:
spot = db.scalar(select(Spot).options(joinedload(Spot.waterbody)).where(Spot.id == spot_id))
if spot is None:
raise HTTPException(status_code=404, detail="spot not found")
return spot
@app.get("/api/v1/spots/{spot_id}", response_model=SpotOut)
def spot_detail(spot_id: UUID, db: Db) -> SpotOut:
spot = _spot_or_404(db, spot_id)
reports = list(db.scalars(select(CatchReport).options(joinedload(CatchReport.bait)).where(CatchReport.spot_id == spot.id, CatchReport.moderation_status == ModerationStatus.approved, CatchReport.deleted_at.is_(None))))
now = datetime.now(timezone.utc)
def count_since(delta: timedelta) -> int:
return sum(_aware(r.reported_at) >= now - delta for r in reports)
bait_counts = Counter(r.bait.name for r in reports if r.bait)
return SpotOut(id=spot.id, waterbody_slug=spot.waterbody.slug, waterbody=spot.waterbody.name_ru, x=spot.x, y=spot.y, description=spot.description, catches_24h=count_since(timedelta(hours=24)), catches_3d=count_since(timedelta(days=3)), catches_7d=count_since(timedelta(days=7)), top_baits=[name for name, _ in bait_counts.most_common(5)])
@app.get("/api/v1/spots/{spot_id}/catches", response_model=list[CatchOut])
def spot_catches(spot_id: UUID, db: Db, limit: int = Query(50, ge=1, le=100), offset: int = Query(0, ge=0)) -> list[CatchOut]:
_spot_or_404(db, spot_id)
reports = list(db.scalars(select(CatchReport).options(joinedload(CatchReport.fish), joinedload(CatchReport.bait)).where(CatchReport.spot_id == spot_id, CatchReport.moderation_status == ModerationStatus.approved, CatchReport.deleted_at.is_(None)).order_by(CatchReport.reported_at.desc()).offset(offset).limit(limit)))
return [CatchOut(id=r.id, fish=r.fish.name_ru, weight_g=r.weight_g, bait=r.bait.name if r.bait else None, player_name=r.player_name, caught_at=r.caught_at, reported_at=r.reported_at, retrieve_method=r.retrieve_method, retrieve_speed=r.retrieve_speed) for r in reports]
@app.get("/api/v1/records", response_model=list[OfficialRecordOut])
def records(
db: Db, fish: str | None = None, waterbody: str | None = None,
category: str | None = None, limit: int = Query(50, ge=1, le=100),
offset: int = Query(0, ge=0),
) -> list[OfficialRecordOut]:
query = select(CatchReport).options(joinedload(CatchReport.fish), joinedload(CatchReport.waterbody), joinedload(CatchReport.bait)).where(CatchReport.source_type == SourceType.official_record)
if fish:
query = query.join(CatchReport.fish).where(Fish.slug == fish)
if waterbody:
query = query.join(CatchReport.waterbody).where(Waterbody.slug == waterbody)
items = list(db.scalars(query.order_by(CatchReport.caught_at.desc(), CatchReport.weight_g.desc()).offset(offset).limit(limit)))
if category:
items = [item for item in items if (item.raw_payload or {}).get("category") == category]
return [OfficialRecordOut(id=r.id, fish=r.fish.name_ru, weight_g=r.weight_g, waterbody=r.waterbody.name_ru, bait=r.bait.name if r.bait else None, player_name=r.player_name, record_date=r.caught_at, category=(r.raw_payload or {}).get("category"), region=(r.raw_payload or {}).get("region"), source_url=r.source_url) for r in items]
def _admin(authorization: Annotated[str | None, Header()] = None) -> str:
if not authorization or authorization != f"Bearer {settings.admin_token}":
raise HTTPException(status_code=401, detail="invalid admin token", headers={"WWW-Authenticate": "Bearer"})
return "admin"
@app.get("/api/v1/imports", response_model=list[ImportRunOut])
def imports(db: Db, limit: int = Query(20, ge=1, le=100)) -> list[OfficialRecordImport]:
return list(db.scalars(select(OfficialRecordImport).order_by(OfficialRecordImport.started_at.desc()).limit(limit)))
@app.get("/api/v1/admin/imports", response_model=list[ImportRunOut])
def admin_imports(
db: Db,
_: Annotated[str, Depends(_admin)],
limit: int = Query(20, ge=1, le=100),
offset: int = Query(0, ge=0),
) -> list[OfficialRecordImport]:
query = select(OfficialRecordImport).order_by(
OfficialRecordImport.started_at.desc(), OfficialRecordImport.id.desc()
).offset(offset).limit(limit)
return list(db.scalars(query))
@app.post("/api/v1/admin/imports/official-records", response_model=ImportRunOut, status_code=201)
def admin_start_official_import(db: Db, _: Annotated[str, Depends(_admin)]) -> OfficialRecordImport:
try:
return import_records(
db,
url=settings.official_records_url,
region=settings.official_records_region,
category=settings.official_records_category,
)
except (ImportSourceError, httpx.HTTPError) as exc:
raise HTTPException(status_code=502, detail=f"official records import failed: {exc}") from exc
def _external_out(item: ExternalObservation) -> ExternalObservationOut:
return ExternalObservationOut(
id=item.id, source_system=item.source_system, source_external_id=item.source_external_id,
source_url=item.source_url, fish_name=item.fish_name, fish_external_id=item.fish_external_id,
waterbody_name=item.waterbody_name, waterbody_external_id=item.waterbody_external_id,
x=item.x, y=item.y, weight_g=item.weight_g, published_at=item.published_at,
last_seen_at=item.last_seen_at, status=item.status,
fish_slug=item.fish.slug if item.fish else None,
waterbody_slug=item.waterbody.slug if item.waterbody else None,
catch_report_id=item.catch_report_id, review_note=item.review_note,
)
@app.get("/api/v1/admin/external-observations", response_model=list[ExternalObservationOut])
def admin_external_observations(
db: Db, _: Annotated[str, Depends(_admin)], status: str | None = None,
source_system: str | None = None, limit: int = Query(50, ge=1, le=200), offset: int = Query(0, ge=0),
) -> list[ExternalObservationOut]:
query = select(ExternalObservation).options(
joinedload(ExternalObservation.fish), joinedload(ExternalObservation.waterbody),
)
if status:
query = query.where(ExternalObservation.status == status)
if source_system:
query = query.where(ExternalObservation.source_system == source_system)
items = db.scalars(query.order_by(ExternalObservation.last_seen_at.desc()).offset(offset).limit(limit))
return [_external_out(item) for item in items]
@app.patch("/api/v1/admin/external-observations/{observation_id}/mapping", response_model=ExternalObservationOut)
def admin_map_external_observation(
observation_id: UUID, payload: ExternalObservationMapping, db: Db,
_: Annotated[str, Depends(_admin)],
) -> ExternalObservationOut:
observation = db.get(ExternalObservation, observation_id)
fish = db.scalar(select(Fish).where(Fish.slug == payload.fish_slug))
waterbody = db.scalar(select(Waterbody).where(Waterbody.slug == payload.waterbody_slug))
if observation is None:
raise HTTPException(status_code=404, detail="external observation not found")
if fish is None or waterbody is None:
raise HTTPException(status_code=422, detail="unknown fish or waterbody")
try:
return _external_out(map_observation(db, observation, fish, waterbody, note=payload.note))
except ExternalReviewError as exc:
raise HTTPException(status_code=409, detail=str(exc)) from exc
@app.post("/api/v1/admin/external-observations/{observation_id}/publish", response_model=ExternalObservationPublished)
def admin_publish_external_observation(
observation_id: UUID, db: Db, _: Annotated[str, Depends(_admin)],
) -> ExternalObservationPublished:
observation = db.get(ExternalObservation, observation_id)
if observation is None:
raise HTTPException(status_code=404, detail="external observation not found")
try:
report = publish_observation(db, observation)
except ExternalReviewError as exc:
raise HTTPException(status_code=409, detail=str(exc)) from exc
return ExternalObservationPublished(observation_id=observation.id, catch_report_id=report.id, status=observation.status)
@app.patch("/api/v1/admin/external-observations/{observation_id}/reject", response_model=ExternalObservationOut)
def admin_reject_external_observation(
observation_id: UUID, payload: ExternalObservationDecision, db: Db,
_: Annotated[str, Depends(_admin)],
) -> ExternalObservationOut:
observation = db.get(ExternalObservation, observation_id)
if observation is None:
raise HTTPException(status_code=404, detail="external observation not found")
try:
return _external_out(reject_observation(db, observation, reason=payload.reason))
except ExternalReviewError as exc:
raise HTTPException(status_code=409, detail=str(exc)) from exc
@app.post("/api/v1/catch-reports", response_model=CatchReportAccepted, status_code=201)
def create_catch_report(payload: CatchReportCreate, request: Request, db: Db) -> CatchReportAccepted:
if payload.website:
raise HTTPException(status_code=400, detail="invalid submission")
_check_rate_limit(request.client.host if request.client else "unknown", db)
fish = db.scalar(select(Fish).where(Fish.slug == payload.fish_slug))
waterbody = db.scalar(select(Waterbody).where(Waterbody.slug == payload.waterbody_slug))
if fish is None or waterbody is None:
raise HTTPException(status_code=422, detail="unknown fish or waterbody")
spot = db.scalar(select(Spot).where(Spot.waterbody_id == waterbody.id, Spot.x == payload.x, Spot.y == payload.y))
if spot is None:
spot = Spot(waterbody=waterbody, x=payload.x, y=payload.y)
db.add(spot)
bait = None
if payload.bait_name and payload.bait_name.strip():
key = normalize(payload.bait_name)
bait = db.scalar(select(Bait).where(Bait.normalized_name == key))
if bait is None:
bait = Bait(name=payload.bait_name.strip(), normalized_name=key, kind=BaitKind.unknown)
db.add(bait)
upload_token = secrets.token_urlsafe(32)
report = CatchReport(fish=fish, spot=spot, waterbody=waterbody, bait=bait, weight_g=payload.weight_g, fishing_method=payload.fishing_method, rig_type=payload.rig_type, retrieve_method=payload.retrieve_method, retrieve_speed=payload.retrieve_speed, caught_at=payload.caught_at, reported_at=datetime.now(timezone.utc), player_name=payload.player_name, source_type=SourceType.user, source_url=payload.source_url, source_confidence=60, moderation_status=ModerationStatus.pending, raw_payload={"comment": payload.comment} if payload.comment else None, screenshot_upload_token_hash=hashlib.sha256(upload_token.encode()).hexdigest())
db.add(report)
db.commit()
return CatchReportAccepted(id=report.id, moderation_status=report.moderation_status.value, screenshot_upload_token=upload_token)
@app.post("/api/v1/catch-reports/{report_id}/screenshot", status_code=204, response_class=Response)
def add_screenshot(
report_id: UUID, db: Db, screenshot: UploadFile = File(),
upload_token: Annotated[str | None, Header(alias="X-Upload-Token")] = None,
) -> Response:
report = db.get(CatchReport, report_id)
if report is None or report.source_type != SourceType.user or report.moderation_status != ModerationStatus.pending:
raise HTTPException(status_code=404, detail="pending catch report not found")
supplied_hash = hashlib.sha256((upload_token or "").encode()).hexdigest()
if not report.screenshot_upload_token_hash or not hmac.compare_digest(report.screenshot_upload_token_hash, supplied_hash):
raise HTTPException(status_code=401, detail="invalid screenshot upload token")
if report.screenshot_key:
raise HTTPException(status_code=409, detail="screenshot already uploaded")
raw = screenshot.file.read(settings.screenshot_max_bytes + 1)
try:
report.screenshot_key = upload_screenshot(raw, filename=screenshot.filename, content_type=screenshot.content_type)
except ScreenshotError as exc:
raise HTTPException(status_code=422, detail=str(exc)) from exc
report.screenshot_upload_token_hash = None
db.commit()
return Response(status_code=204)
@app.get("/api/v1/admin/catch-reports", response_model=list[AdminCatchReportOut])
def admin_reports(db: Db, _: Annotated[str, Depends(_admin)], status: ModerationStatus = ModerationStatus.pending, limit: int = Query(50, ge=1, le=100)) -> list[AdminCatchReportOut]:
reports = list(db.scalars(select(CatchReport).options(joinedload(CatchReport.fish), joinedload(CatchReport.waterbody), joinedload(CatchReport.spot), joinedload(CatchReport.bait)).where(CatchReport.source_type == SourceType.user, CatchReport.moderation_status == status, CatchReport.deleted_at.is_(None)).order_by(CatchReport.reported_at).limit(limit)))
return [AdminCatchReportOut(id=r.id, fish=r.fish.name_ru, waterbody=r.waterbody.name_ru, coordinates=f"{r.spot.x}:{r.spot.y}" if r.spot else "—", weight_g=r.weight_g, bait=r.bait.name if r.bait else None, player_name=r.player_name, reported_at=r.reported_at, moderation_status=r.moderation_status.value, comment=(r.raw_payload or {}).get("comment"), screenshot_url=signed_screenshot_url(r.screenshot_key) if r.screenshot_key else None) for r in reports]
@app.patch("/api/v1/admin/catch-reports/{report_id}", response_model=CatchReportCreated)
def moderate_report(report_id: UUID, payload: ModerationUpdate, db: Db, moderator: Annotated[str, Depends(_admin)]) -> CatchReportCreated:
report = db.get(CatchReport, report_id)
if report is None or report.source_type != SourceType.user or report.deleted_at is not None:
raise HTTPException(status_code=404, detail="catch report not found")
previous = report.moderation_status
report.moderation_status = ModerationStatus(payload.status)
db.add(ModerationEvent(catch_report=report, created_at=datetime.now(timezone.utc), previous_status=previous, new_status=report.moderation_status, moderator=moderator, reason=payload.reason))
db.commit()
return CatchReportCreated(id=report.id, moderation_status=report.moderation_status.value)
@app.delete("/api/v1/admin/catch-reports/{report_id}", status_code=204, response_class=Response)
def delete_report(report_id: UUID, db: Db, moderator: Annotated[str, Depends(_admin)]) -> Response:
report = db.get(CatchReport, report_id)
if report is None or report.source_type != SourceType.user or report.deleted_at is not None:
raise HTTPException(status_code=404, detail="catch report not found")
previous = report.moderation_status
if report.screenshot_key:
try:
delete_screenshot(report.screenshot_key)
except Exception as exc:
raise HTTPException(status_code=502, detail="screenshot deletion failed") from exc
report.moderation_status = ModerationStatus.rejected
report.deleted_at = datetime.now(timezone.utc)
report.player_name = None
report.source_url = None
report.screenshot_key = None
report.raw_payload = None
db.add(ModerationEvent(catch_report=report, created_at=report.deleted_at, previous_status=previous, new_status=ModerationStatus.rejected, moderator=moderator, reason="user report deleted and anonymized"))
db.commit()
return Response(status_code=204)
def _check_rate_limit(client: str, db: Session) -> None:
now = datetime.now(timezone.utc)
cutoff = now - timedelta(minutes=10)
client_hash = hmac.new(settings.rate_limit_secret.encode(), client.encode(), hashlib.sha256).hexdigest()
if db.get_bind().dialect.name == "postgresql":
lock_key = int(client_hash[:16], 16) & 0x7FFF_FFFF_FFFF_FFFF
db.execute(text("SELECT pg_advisory_xact_lock(:lock_key)"), {"lock_key": lock_key})
db.execute(delete(SubmissionAttempt).where(SubmissionAttempt.created_at < now - timedelta(days=1)))
recent = db.scalar(select(func.count()).select_from(SubmissionAttempt).where(SubmissionAttempt.client_hash == client_hash, SubmissionAttempt.created_at >= cutoff)) or 0
if recent >= 5:
db.commit()
raise HTTPException(status_code=429, detail="too many submissions")
db.add(SubmissionAttempt(client_hash=client_hash, created_at=now))
db.commit()
def _aware(value: datetime) -> datetime:
return value if value.tzinfo else value.replace(tzinfo=timezone.utc)