#!/usr/bin/env python3 """make_gallery.py — fetch actor images for a movie and build gallery.h5. Fetches the cast from TMDB, downloads actor profile images, embeds them via the sae_embed module (SCRFD + ArcFace, same models as scene_analyze, loaded once), then writes gallery.h5. Requirements: pip install requests Pillow Usage: # By IMDB movie ID (most natural — resolves to TMDB automatically): python scripts/make_gallery.py \\ --tmdb-key YOUR_KEY \\ --imdb-id tt0137523 \\ --output gallery.h5 # Or directly with a TMDB movie ID: python scripts/make_gallery.py \\ --tmdb-key YOUR_KEY \\ --movie-id 550 \\ --output gallery.h5 # Additional options: # --build-dir build/ build dir containing sae_embed module # --models-dir models/ directory with ONNX models # --images-per-actor 3 profile images to download per actor # --image-dir /tmp/gallery_imgs where to cache downloaded images Get a free TMDB API key at: https://www.themoviedb.org/settings/api """ import argparse import sys import time from pathlib import Path sys.path.insert(0, str(Path(__file__).resolve().parent)) from sae_embed_loader import load_embedder, resolve_arcface from sae_gallery import (download_images, embedder_stamp, save_gallery, wikidata_image_urls) from sae_tmdb import TMDB_IMG, tmdb_get, tmdb_id_from_imdb def fetch_cast(movie_id: int, key: str) -> list[dict]: """Return list of {id, name, imdb_id, profile_images: [...url...]}.""" credits = tmdb_get(f"/movie/{movie_id}/credits", key) cast = credits.get("cast", []) actors = [] for member in cast: person_id = member["id"] # Get IMDB ID for this person ext = tmdb_get(f"/person/{person_id}/external_ids", key) imdb_id = ext.get("imdb_id") or "" # Get profile images (sorted by vote_average desc by TMDB) images_data = tmdb_get(f"/person/{person_id}/images", key) profiles = images_data.get("profiles", []) image_urls = [TMDB_IMG + p["file_path"] for p in profiles if p.get("file_path")] if not image_urls: image_urls = wikidata_image_urls(imdb_id) if image_urls: print(f" [info] no TMDB images for {member['name']}, " f"found {len(image_urls)} via Wikidata", file=sys.stderr) if not image_urls: print(f" [warn] no images for {member['name']}", file=sys.stderr) actors.append({ "id": person_id, "name": member["name"], "imdb_id": imdb_id, "tmdb_id": str(person_id), "profile_images": image_urls, }) time.sleep(0.05) # be polite to TMDB return actors # ── Gallery assembly ───────────────────────────────────────────────────────── def build_gallery(movie_id: int, key: str, embedder, images_per_actor: int, image_root: Path) -> tuple[dict, list[dict]]: """Fetch cast, download images, embed, return (gallery dict, actors needing more images).""" print(f"Fetching cast for TMDB movie {movie_id}…", file=sys.stderr) actors = fetch_cast(movie_id, key) print(f"Found {len(actors)} cast member(s)", file=sys.stderr) gallery_actors = [] missing = [] for actor in actors: safe_name = actor["name"].replace(" ", "_") dir_id = actor["imdb_id"] or f"tmdb_{actor['tmdb_id']}" actor_dir = image_root / f"{dir_id}_{safe_name}" print(f"\n{actor['name']} ({dir_id})", file=sys.stderr) if not actor["profile_images"]: print(" no images found, skipping", file=sys.stderr) missing.append({"name": actor["name"], "imdb_id": actor["imdb_id"], "tmdb_id": actor["tmdb_id"], "reason": "no images found"}) continue image_paths = download_images(actor["profile_images"], actor_dir, images_per_actor) if not image_paths: print(" no images downloaded, skipping", file=sys.stderr) missing.append({"name": actor["name"], "imdb_id": actor["imdb_id"], "tmdb_id": actor["tmdb_id"], "reason": "download failed"}) continue print(f" embedding {len(image_paths)} image(s)…", file=sys.stderr) embeddings = [] source_images = [] for path in image_paths: res = embedder.embed(str(path)) if not res.ok: print(f" [skip] {path.name}: {res.error}", file=sys.stderr) continue embeddings.append(res.embedding) source_images.append(path.name) print(f" [ok] {path.name} conf={res.confidence:.2f}", file=sys.stderr) if not embeddings: print(" no valid embeddings, skipping actor", file=sys.stderr) missing.append({"name": actor["name"], "imdb_id": actor["imdb_id"], "tmdb_id": actor["tmdb_id"], "reason": "no valid embeddings"}) continue gallery_actors.append({ "imdb_id": actor["imdb_id"], "tmdb_id": actor["tmdb_id"], "jellyfin_id": "", "name": actor["name"], "source_images": source_images, "embeddings": embeddings, }) print(f" → {len(embeddings)} embedding(s) stored", file=sys.stderr) return {"actors": gallery_actors}, missing # ── Entry point ─────────────────────────────────────────────────────────────── def main(): parser = argparse.ArgumentParser( description="Fetch TMDB cast images and build gallery.h5 via sae_embed") parser.add_argument("--tmdb-key", required=True, help="TMDB Bearer token (API Read Access Token from themoviedb.org/settings/api)") group = parser.add_mutually_exclusive_group(required=True) group.add_argument("--imdb-id", help="IMDB movie ID, e.g. tt0137523 — looked up via TMDB automatically") group.add_argument("--movie-id", type=int, help="TMDB movie ID (alternative to --imdb-id)") parser.add_argument("--output", required=True, help="Output gallery.h5 path") parser.add_argument("--build-dir", default="build", help="Build directory containing the sae_embed module (default: build)") parser.add_argument("--models-dir", default="models", help="Directory containing ONNX models (default: models/)") parser.add_argument("--arcface", default=None, help="Path to ArcFace ONNX model (overrides --models-dir selection)") parser.add_argument("--images-per-actor",type=int, default=3, help="Profile images to download per actor (default: 3)") parser.add_argument("--image-dir", default=None, help="Where to store downloaded images (default: /images)") parser.add_argument("--keep-images", action="store_true", help="Do not delete downloaded images after embedding") args = parser.parse_args() # Resolve paths output = Path(args.output) image_root = Path(args.image_dir) if args.image_dir else output.parent / "images" # TRACES: GR-004 | SR-001 # stamp with the model actually loaded, resolved # through the same helper load_embedder uses so the two cannot diverge. arcface_path = resolve_arcface(args.models_dir, args.arcface) embedder = load_embedder(args.build_dir, args.models_dir, args.arcface) stamp = embedder_stamp(arcface_path) # Resolve movie ID movie_id = args.movie_id if movie_id is None: print(f"Resolving IMDB ID {args.imdb_id} → TMDB…", file=sys.stderr) movie_id = tmdb_id_from_imdb(args.imdb_id, args.tmdb_key) print(f"TMDB movie ID: {movie_id}", file=sys.stderr) # Build gallery gallery, missing = build_gallery( movie_id = movie_id, key = args.tmdb_key, embedder = embedder, images_per_actor = args.images_per_actor, image_root = image_root, ) n_actors = len(gallery["actors"]) n_embeddings = sum(len(a["embeddings"]) for a in gallery["actors"]) print(f"\nGallery: {n_actors} actors, {n_embeddings} total embeddings", file=sys.stderr) if n_actors == 0: sys.exit("No actors could be processed — check models and images.") save_gallery(gallery, missing, output, embedder=stamp) if __name__ == "__main__": main()