feat: harden production data and backups

apps/api/app/config.py

@@ -16,6 +16,7 @@ class Settings(BaseSettings):
     official_records_region: str = "RU"
     official_records_category: str = "records"
     official_import_required: bool = False
+    seed_demo_data: bool = True
     import_interval_seconds: int = Field(default=3600, ge=3600)
     rate_limit_secret: str = "change-rate-limit-secret"
     log_level: str = "INFO"
@@ -39,6 +40,8 @@ class Settings(BaseSettings):
             raise ValueError("production CORS_ORIGINS must contain only HTTPS origins")
         if not self.s3_public_endpoint_url.startswith("https://"):
             raise ValueError("production S3_PUBLIC_ENDPOINT_URL must use HTTPS")
+        if self.seed_demo_data:
+            raise ValueError("SEED_DEMO_DATA must be false in production")
         return self
 
 

apps/api/app/main.py

@@ -5,6 +5,7 @@ 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
@@ -25,7 +26,7 @@ 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, CatchReportCreate, CatchReportCreated, ExternalObservationDecision, ExternalObservationMapping, ExternalObservationOut, ExternalObservationPublished, FishOut, ImportRunOut, ModerationUpdate, OfficialRecordOut, SpotOut, WaterbodyOut
+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
 
 
@@ -273,8 +274,8 @@ def admin_reject_external_observation(
         raise HTTPException(status_code=409, detail=str(exc)) from exc
 
 
-@app.post("/api/v1/catch-reports", response_model=CatchReportCreated, status_code=201)
-def create_catch_report(payload: CatchReportCreate, request: Request, db: Db) -> CatchReportCreated:
+@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)
@@ -293,17 +294,24 @@ def create_catch_report(payload: CatchReportCreate, request: Request, db: Db) ->
         if bait is None:
             bait = Bait(name=payload.bait_name.strip(), normalized_name=key, kind=BaitKind.unknown)
             db.add(bait)
-    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)
+    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 CatchReportCreated(id=report.id, moderation_status=report.moderation_status.value)
+    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()) -> 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)
@@ -311,6 +319,7 @@ def add_screenshot(report_id: UUID, db: Db, screenshot: UploadFile = File()) ->
         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)
 

apps/api/app/models.py

@@ -92,6 +92,7 @@ class CatchReport(Base):
     source_confidence: Mapped[int]
     moderation_status: Mapped[ModerationStatus] = mapped_column(Enum(ModerationStatus))
     screenshot_key: Mapped[str | None] = mapped_column(Text)
+    screenshot_upload_token_hash: Mapped[str | None] = mapped_column(String(64))
     deleted_at: Mapped[datetime | None] = mapped_column(DateTime(timezone=True))
     raw_payload: Mapped[dict | None] = mapped_column(JSON)
     fish: Mapped[Fish] = relationship()

apps/api/app/schemas.py

@@ -137,6 +137,10 @@ class CatchReportCreated(BaseModel):
     moderation_status: str
 
 
+class CatchReportAccepted(CatchReportCreated):
+    screenshot_upload_token: str
+
+
 class AdminCatchReportOut(BaseModel):
     id: UUID
     fish: str

apps/api/app/seed.py

@@ -5,6 +5,7 @@ from uuid import UUID
 
 from sqlalchemy import select
 
+from .config import settings
 from .database import SessionLocal
 from .models import Bait, BaitKind, CatchReport, Fish, ModerationStatus, SourceType, Spot, Waterbody
 
@@ -23,19 +24,38 @@ IDS = {
 
 def seed() -> None:
     with SessionLocal.begin() as db:
-        if db.scalar(select(Fish.id).limit(1)) is not None:
+        def entity(model, key: str, value: str, **values):
+            item = db.scalar(select(model).where(getattr(model, key) == value))
+            if item is None:
+                item = model(**values)
+                db.add(item)
+                db.flush()
+            return item
+
+        vyunok = entity(Waterbody, "slug", "vyunok", id=IDS["vyunok"], slug="vyunok", name_ru="Вьюнок", unlock_level=1)
+        kuori = entity(Waterbody, "slug", "kuori", id=IDS["kuori"], slug="kuori", name_ru="Куори", unlock_level=16)
+        pike = entity(Fish, "slug", "pike", id=IDS["pike"], slug="pike", name_ru="Щука", trophy_weight_g=10_000)
+        trout = entity(Fish, "slug", "lake-trout", id=IDS["trout"], slug="lake-trout", name_ru="Озёрная форель", trophy_weight_g=10_000)
+        spiker = entity(Bait, "normalized_name", "spiker #2 01-015", id=IDS["spiker"], name="Spiker #2 01-015", normalized_name="spiker #2 01-015", kind=BaitKind.lure)
+        shad = entity(Bait, "normalized_name", "salmon t1 shad 12 005", id=IDS["shad"], name="Salmon T1 Shad 12 005", normalized_name="salmon t1 shad 12 005", kind=BaitKind.lure)
+
+        def spot(waterbody: Waterbody, identity: str, x: int, y: int, description: str) -> Spot:
+            item = db.scalar(select(Spot).where(Spot.waterbody_id == waterbody.id, Spot.x == x, Spot.y == y))
+            if item is None:
+                item = Spot(id=IDS[identity], waterbody=waterbody, x=x, y=y, description=description)
+                db.add(item)
+                db.flush()
+            return item
+
+        spot1 = spot(vyunok, "spot1", 110, 103, "Кромка травы у северного берега")
+        spot2 = spot(kuori, "spot2", 85, 92, "Свальчик в глубину")
+        if not settings.seed_demo_data:
             return
-        vyunok = Waterbody(id=IDS["vyunok"], slug="vyunok", name_ru="Вьюнок", unlock_level=1)
-        kuori = Waterbody(id=IDS["kuori"], slug="kuori", name_ru="Куори", unlock_level=16)
-        pike = Fish(id=IDS["pike"], slug="pike", name_ru="Щука", trophy_weight_g=10_000)
-        trout = Fish(id=IDS["trout"], slug="lake-trout", name_ru="Озёрная форель", trophy_weight_g=10_000)
-        spiker = Bait(id=IDS["spiker"], name="Spiker #2 01-015", normalized_name="spiker #2 01-015", kind=BaitKind.lure)
-        shad = Bait(id=IDS["shad"], name="Salmon T1 Shad 12 005", normalized_name="salmon t1 shad 12 005", kind=BaitKind.lure)
-        spot1 = Spot(id=IDS["spot1"], waterbody=vyunok, x=110, y=103, description="Кромка травы у северного берега")
-        spot2 = Spot(id=IDS["spot2"], waterbody=kuori, x=85, y=92, description="Свальчик в глубину")
-        db.add_all([vyunok, kuori, pike, trout, spiker, shad, spot1, spot2])
         now = datetime.now(timezone.utc)
         for index in range(12):
+            external_id = f"seed:pike:{index}"
+            if db.scalar(select(CatchReport.id).where(CatchReport.source_external_id == external_id)):
+                continue
             db.add(CatchReport(
                 fish=pike, spot=spot1, waterbody=vyunok, bait=spiker,
                 weight_g=2_600 + index * 480, fishing_method="spinning",
@@ -43,9 +63,13 @@ def seed() -> None:
                 caught_at=now - timedelta(minutes=25 + index * 47),
                 reported_at=now - timedelta(minutes=20 + index * 47),
                 player_name=f"DemoPlayer{index % 7 + 1}", source_type=SourceType.manual_import,
+                source_external_id=external_id,
                 source_confidence=80 + index % 3 * 5, moderation_status=ModerationStatus.approved,
             ))
         for index in range(5):
+            external_id = f"seed:trout:{index}"
+            if db.scalar(select(CatchReport.id).where(CatchReport.source_external_id == external_id)):
+                continue
             db.add(CatchReport(
                 fish=trout, spot=spot2, waterbody=kuori, bait=shad,
                 weight_g=4_200 + index * 900, fishing_method="spinning",
@@ -53,6 +77,7 @@ def seed() -> None:
                 caught_at=now - timedelta(hours=2 + index * 4),
                 reported_at=now - timedelta(hours=2 + index * 4),
                 player_name=f"DemoAngler{index + 1}", source_type=SourceType.manual_import,
+                source_external_id=external_id,
                 source_confidence=85, moderation_status=ModerationStatus.approved,
             ))