Sync from GitHub via hub-sync

api/v1/analyze.py

@@ -433,8 +433,9 @@ async def analyze_image(
     except Exception as e:  # noqa: BLE001
         logger.warning(f"media save failed: {e}")
         media_path = None
-    thumbnail_url = make_image_thumbnail(pil, media_hash)
+    thumbnail_url, thumbnail_b64 = make_image_thumbnail(pil, media_hash)
     resp.thumbnail_url = thumbnail_url
+    resp.thumbnail_b64 = thumbnail_b64
     if media_path:
         resp.media_path = media_path
 
@@ -648,8 +649,9 @@ async def analyze_video_endpoint(
     except Exception as e:  # noqa: BLE001
         logger.warning(f"video media save failed: {e}")
         media_path = None
-    thumbnail_url = make_video_thumbnail(path, media_hash)
+    thumbnail_url, thumbnail_b64 = make_video_thumbnail(path, media_hash)
     resp.thumbnail_url = thumbnail_url
+    resp.thumbnail_b64 = thumbnail_b64
 
     record = AnalysisRecord(
         user_id=user.id if user else None,
@@ -996,8 +998,9 @@ async def analyze_screenshot_endpoint(
     except Exception as e:  # noqa: BLE001
         logger.warning(f"screenshot media save failed: {e}")
         media_path = None
-    thumbnail_url = make_image_thumbnail(pil, media_hash)
+    thumbnail_url, thumbnail_b64 = make_image_thumbnail(pil, media_hash)
     resp.thumbnail_url = thumbnail_url
+    resp.thumbnail_b64 = thumbnail_b64
 
     record = AnalysisRecord(
         user_id=user.id if user else None,
@@ -1157,8 +1160,9 @@ async def analyze_video_async(
             except Exception as e:  # noqa: BLE001
                 logger.warning(f"async video media save failed: {e}")
                 media_path = None
-            thumb = make_video_thumbnail(path, media_hash)
-            resp.thumbnail_url = thumb
+            thumb_url, thumb_b64 = make_video_thumbnail(path, media_hash)
+            resp.thumbnail_url = thumb_url
+            resp.thumbnail_b64 = thumb_b64
 
             rec = AnalysisRecord(
                 user_id=user_id,
@@ -1168,7 +1172,7 @@ async def analyze_video_async(
                 result_json=json.dumps(resp.model_dump()),
                 media_hash=media_hash,
                 media_path=media_path,
-                thumbnail_url=thumb,
+                thumbnail_url=thumb_url,
             )
             local_db.add(rec)
             local_db.commit()

api/v1/history.py

@@ -26,6 +26,7 @@ class HistoryItem(BaseModel):
     authenticity_score: float
     created_at: datetime
     thumbnail_url: str | None = None
+    thumbnail_b64: str | None = None  # inline data URL; preferred over thumbnail_url
     media_path: str | None = None
     text_preview: str | None = None
 
@@ -33,6 +34,7 @@ class HistoryItem(BaseModel):
 class HistoryListResponse(BaseModel):
     items: list[HistoryItem]
     total: int
+    cache_hits: int = 0
 
 
 class HistoryDeleteAllResponse(BaseModel):
@@ -129,6 +131,42 @@ def _history_text_preview(record: AnalysisRecord, limit: int = 260) -> str | Non
     return text[: limit - 3].rstrip() + "..."
 
 
+def _thumbnail_b64_from_record(r: AnalysisRecord) -> str | None:
+    """Extract inline thumbnail base64 from result_json if present."""
+    try:
+        payload = json.loads(r.result_json)
+        b64 = payload.get("thumbnail_b64")
+        if b64 and str(b64).startswith("data:image"):
+            return b64
+    except Exception:
+        pass
+    return None
+
+
+def _count_cache_hits(db: Session, user_id: int) -> int:
+    """Count analyses whose media_hash matches an earlier record for this user.
+    Each duplicate re-submission that was served from cache is a cache hit.
+    """
+    from sqlalchemy import func, text as sa_text
+    try:
+        result = db.execute(
+            sa_text(
+                "SELECT COUNT(*) FROM analyses a1 "
+                "WHERE a1.user_id = :uid AND a1.media_hash IS NOT NULL "
+                "AND EXISTS ("
+                "  SELECT 1 FROM analyses a2 "
+                "  WHERE a2.user_id = :uid AND a2.media_hash = a1.media_hash "
+                "  AND a2.id < a1.id"
+                ")"
+            ),
+            {"uid": user_id},
+        )
+        row = result.fetchone()
+        return int(row[0]) if row else 0
+    except Exception:
+        return 0
+
+
 @router.get("", response_model=HistoryListResponse)
 def list_history(
     limit: int = Query(default=50, ge=1, le=200),
@@ -136,10 +174,10 @@ def list_history(
     user: User = Depends(get_current_user),
     db: Session = Depends(get_db),
 ) -> HistoryListResponse:
-    from sqlalchemy.orm import defer
     q = db.query(AnalysisRecord).filter(AnalysisRecord.user_id == user.id)
     total = q.count()
-    rows = q.options(defer(AnalysisRecord.result_json)).order_by(AnalysisRecord.created_at.desc()).offset(offset).limit(limit).all()
+    # Load result_json so we can extract thumbnail_b64
+    rows = q.order_by(AnalysisRecord.created_at.desc()).offset(offset).limit(limit).all()
     items = [
         HistoryItem(
             id=r.id,
@@ -148,12 +186,14 @@ def list_history(
             authenticity_score=r.authenticity_score,
             created_at=r.created_at,
             thumbnail_url=_make_asset_url(r, "thumbnail") if r.thumbnail_url else None,
+            thumbnail_b64=_thumbnail_b64_from_record(r),
             media_path=_make_asset_url(r, "media") if r.media_path else None,
             text_preview=_history_text_preview(r),
         )
         for r in rows
     ]
-    return HistoryListResponse(items=items, total=total)
+    cache_hits = _count_cache_hits(db, user.id)
+    return HistoryListResponse(items=items, total=total, cache_hits=cache_hits)
 
 
 @router.get("/{record_id}")

schemas/analyze.py

@@ -104,6 +104,7 @@ class ScreenshotAnalysisResponse(BaseModel):
     record_id: int = 0
     cached: bool = False
     thumbnail_url: str | None = None
+    thumbnail_b64: str | None = None
     media_type: str = "screenshot"
     timestamp: str
     verdict: Verdict
@@ -179,6 +180,7 @@ class VideoAnalysisResponse(BaseModel):
     record_id: int = 0
     cached: bool = False
     thumbnail_url: str | None = None
+    thumbnail_b64: str | None = None
     media_type: str = "video"
     timestamp: str
     verdict: Verdict
@@ -197,6 +199,7 @@ class ImageAnalysisResponse(BaseModel):
     record_id: int = 0
     cached: bool = False
     thumbnail_url: str | None = None
+    thumbnail_b64: str | None = None  # inline data URL, survives without file storage
     media_path: str | None = None
     media_type: str = "image"
     timestamp: str
@@ -230,6 +233,7 @@ class AudioAnalysisResponse(BaseModel):
     record_id: int = 0
     cached: bool = False
     thumbnail_url: str | None = None
+    thumbnail_b64: str | None = None
     media_type: str = "audio"
     timestamp: str
     verdict: Verdict

services/storage.py

@@ -11,6 +11,7 @@ from __future__ import annotations
 
 import base64
 import hashlib
+import io
 import os
 from pathlib import Path
 
@@ -115,15 +116,11 @@ def make_image_thumbnail(pil: Image.Image, sha: str) -> tuple[str | None, str |
     return url_path, data_url
 
 
-def make_video_thumbnail(video_path: str, sha: str) -> str | None:
-    """Grab a frame ~1s in as the video thumbnail."""
+def make_video_thumbnail(video_path: str, sha: str) -> tuple[str | None, str | None]:
+    """Grab a frame ~1s in as the video thumbnail. Returns (url_path, data_url)."""
     try:
         import cv2  # lazy import — heavy
 
-        _ensure_dirs()
-        dest = THUMB_DIR / f"{sha}_400.jpg"
-        if dest.exists():
-            return f"/media/thumbs/{sha}_400.jpg"
         cap = cv2.VideoCapture(video_path)
         try:
             fps = cap.get(cv2.CAP_PROP_FPS) or 25
@@ -133,17 +130,31 @@ def make_video_thumbnail(video_path: str, sha: str) -> str | None:
                 cap.set(cv2.CAP_PROP_POS_FRAMES, 0)
                 ok, frame = cap.read()
             if not ok:
-                return None
+                return None, None
         finally:
             cap.release()
         rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
         im = Image.fromarray(rgb)
         im.thumbnail((THUMB_MAX, THUMB_MAX))
-        im.save(dest, "JPEG", quality=82, optimize=True)
-        return f"/media/thumbs/{sha}_400.jpg"
+        buf = io.BytesIO()
+        im.save(buf, "JPEG", quality=75, optimize=True)
+        b64 = base64.b64encode(buf.getvalue()).decode("ascii")
+        data_url = f"data:image/jpeg;base64,{b64}"
     except Exception as e:  # noqa: BLE001
         logger.warning(f"video thumbnail failed for {sha}: {e}")
-        return None
+        return None, None
+
+    url_path: str | None = None
+    try:
+        _ensure_dirs()
+        dest = THUMB_DIR / f"{sha}_400.jpg"
+        if not dest.exists():
+            dest.write_bytes(buf.getvalue())
+        url_path = f"/media/thumbs/{sha}_400.jpg"
+    except Exception as e:  # noqa: BLE001
+        logger.warning(f"video thumbnail file save failed for {sha}: {e}")
+
+    return url_path, data_url
 
 
 def save_overlay(data_url: str, sha: str, suffix: str) -> str | None: