Upload 2 files

app.py

@@ -1,202 +1,248 @@
-import os
-import io
-import json
-import yaml
-import shutil
-import sys
-import subprocess
-from pathlib import Path
-from passlib.hash import bcrypt
-try:
-    import tomllib
-except ModuleNotFoundError:  # Py<3.11
-    import tomli as tomllib
-import streamlit as st
-
-try:
-    from moviepy.editor import VideoFileClip
-except ModuleNotFoundError:  # HF runtime safeguard
-    subprocess.check_call([sys.executable, "-m", "pip", "install", "moviepy", "imageio", "imageio-ffmpeg"])
-    from moviepy.editor import VideoFileClip
-
-from database import set_db_path, init_schema, get_user, create_video, update_video_status, list_videos, get_video, get_all_users, upsert_result, get_results, add_feedback, get_feedback_for_video, get_feedback_stats
-from api_client import APIClient
-from utils import ensure_dirs, save_bytes, save_text, human_size
-
-
-# -- Move DB ---
-os.environ["STREAMLIT_DATA_DIRECTORY"] = "/tmp/.streamlit"
-Path("/tmp/.streamlit").mkdir(parents=True, exist_ok=True)
-Path("/tmp/data").mkdir(parents=True, exist_ok=True)
-source_db = "init_data/veureu.db"
-target_db = "/tmp/data/app.db"
-if not os.path.exists(target_db) and os.path.exists(source_db):
-    shutil.copy(source_db, target_db)
-
-static_videos = Path(__file__).parent / "videos"
-runtime_videos = Path("/tmp/data/videos")
-if not runtime_videos.exists():
-    shutil.copytree(static_videos, runtime_videos, dirs_exist_ok=True)
-
-
-# --- Config ---
-def _load_yaml(path="config.yaml") -> dict:
-    with open(path, "r", encoding="utf-8") as f:
-        cfg = yaml.safe_load(f) or {}
-    # interpolación sencilla de ${VARS} si las usas en el YAML
-    def _subst(s: str) -> str:
-        return os.path.expandvars(s) if isinstance(s, str) else s
-
-    # aplica sustitución en los campos que te interesan
-    if "api" in cfg:
-        cfg["api"]["base_url"] = _subst(cfg["api"].get("base_url", ""))
-        cfg["api"]["token"]    = _subst(cfg["api"].get("token", ""))
-
-    if "storage" in cfg and "root_dir" in cfg["storage"]:
-        cfg["storage"]["root_dir"] = _subst(cfg["storage"]["root_dir"])
-
-    if "sqlite" in cfg and "path" in cfg["sqlite"]:
-        cfg["sqlite"]["path"] = _subst(cfg["sqlite"]["path"])
-
-    return cfg
-
-CFG = _load_yaml("config.yaml")
-
-# Ajuste de variables según tu esquema YAML
-DATA_DIR = CFG.get("storage", {}).get("root_dir", "data")
-BACKEND_BASE_URL = CFG.get("api", {}).get("base_url", "http://localhost:8000")
-USE_MOCK = bool(CFG.get("app", {}).get("use_mock", False))  # si no la tienes en el yaml, queda False
-API_TOKEN = CFG.get("api", {}).get("token") or os.getenv("API_SHARED_TOKEN")
-
-os.makedirs(DATA_DIR, exist_ok=True)
-ensure_dirs(DATA_DIR)
-DB_PATH = os.path.join(DATA_DIR, "app.db")
-set_db_path(DB_PATH)
-init_schema()
-
-api = APIClient(BACKEND_BASE_URL, use_mock=USE_MOCK, data_dir=DATA_DIR, token=API_TOKEN)
-
-st.set_page_config(page_title="Veureu — Audiodescripció", page_icon="🎬", layout="wide")
-
-# --- Session: auth ---
-# print("Usuarios disponibles:", get_all_users()) # Descomentar para depurar
-if "user" not in st.session_state:
-    st.session_state.user = None  # dict with {username, role, id(optional)}
-
-def require_login():
-    if not st.session_state.user:
-        st.info("Por favor, inicia sesión para continuar.")
-        login_form()
-        st.stop()
-
-def verify_password(password: str, pw_hash: str) -> bool:
-    try:
-        return bcrypt.verify(password, pw_hash)
-    except Exception:
-        return False
-        
-# --- Sidebar (only after login) ---
-role = st.session_state.user["role"] if st.session_state.user else None
-with st.sidebar:
-    st.title("Veureu")
-    if st.session_state.user:
-        st.write(f"Usuari: **{st.session_state.user['username']}** (rol: {st.session_state.user['role']})")
-        if st.button("Tancar sessió"):
-            st.session_state.user = None
-            st.rerun()
-    if st.session_state.user:
-        page = st.radio("Navegació", ["Analitzar video-transcripcions","Processar vídeo nou","Estadístiques"], index=0)
-    else:
-        page = None
-
-# --- Pre-login screen ---
-if not st.session_state.user:
-    st.title("Veureu — Audiodescripció")
-    def login_form():
-        st.subheader("Inici de sessió")
-        username = st.text_input("Usuari")
-        password = st.text_input("Contrasenya", type="password")
-        if st.button("Entrar", type="primary"):
-            row = get_user(username)
-            if row and verify_password(password, row["pw_hash"]):
-                st.session_state.user = {"id": row["id"], "username": row["username"], "role": row["role"]}
-                st.success(f"Benvingut/da, {row['username']}")
-                st.rerun()
-            else:
-                st.error("Credencials invàlides")
-    login_form()
-    st.stop()
-
-# --- Pages ---
-if page == "Processar vídeo nou":
-    require_login()
-    if role != "verd":
-        st.error("No tens permisos per processar nous vídeos. Canvia d'usuari o sol·licita permisos.")
-        st.stop()
-
-    st.header("Processar un nou clip de vídeo")
-
-    # Inicializar el estado de la página si no existe
-    if 'video_uploaded' not in st.session_state:
-        st.session_state.video_uploaded = None
-    if 'characters_detected' not in st.session_state:
-        st.session_state.characters_detected = None
-    if 'characters_saved' not in st.session_state:
-        st.session_state.characters_saved = False
-
-    # --- 1. Subida del vídeo ---
-    MAX_SIZE_MB = 20
-    MAX_DURATION_S = 240 # 4 minutos
-
-    uploaded_file = st.file_uploader("Puja un clip de vídeo (MP4, < 20MB, < 4 minuts)", type=["mp4"], key="video_uploader")
-
-    if uploaded_file is not None:
-        # Resetear el estado si se sube un nuevo archivo
-        if st.session_state.video_uploaded is None or uploaded_file.name != st.session_state.video_uploaded.get('original_name'):
-            st.session_state.video_uploaded = {'original_name': uploaded_file.name, 'status': 'validating'}
-            st.session_state.characters_detected = None
-            st.session_state.characters_saved = False
-
-        # --- Validación y Procesamiento ---
-        if st.session_state.video_uploaded['status'] == 'validating':
-            is_valid = True
-            # 1. Validar tamaño
-            if uploaded_file.size > MAX_SIZE_MB * 1024 * 1024:
-                st.error(f"El vídeo supera el límit de {MAX_SIZE_MB}MB.")
-                is_valid = False
-
-            if is_valid:
-                with st.spinner("Processant el vídeo..."):
-                    # Guardar temporalmente para analizarlo
-                    with open("temp_video.mp4", "wb") as f:
+import os
+import io
+import json
+import yaml
+import shutil
+import subprocess
+from pathlib import Path
+from passlib.hash import bcrypt
+try:
+    import tomllib
+except ModuleNotFoundError:  # Py<3.11
+    import tomli as tomllib
+import streamlit as st
+
+from database import set_db_path, init_schema, get_user, create_video, update_video_status, list_videos, get_video, get_all_users, upsert_result, get_results, add_feedback, get_feedback_for_video, get_feedback_stats
+from api_client import APIClient
+from utils import ensure_dirs, save_bytes, save_text, human_size
+
+
+def _get_video_duration(path: str) -> float:
+    """Return video duration in seconds using ffprobe."""
+    cmd = [
+        "ffprobe",
+        "-v",
+        "error",
+        "-show_entries",
+        "format=duration",
+        "-of",
+        "default=noprint_wrappers=1:nokey=1",
+        path,
+    ]
+    try:
+        result = subprocess.run(cmd, capture_output=True, text=True, check=True)
+        return float(result.stdout.strip())
+    except (subprocess.CalledProcessError, ValueError):
+        return 0.0
+
+
+def _transcode_video(input_path: str, output_path: str, max_duration: int | None = None) -> None:
+    cmd = ["ffmpeg", "-y", "-i", input_path]
+    if max_duration is not None:
+        cmd += ["-t", str(max_duration)]
+    cmd += [
+        "-c:v",
+        "libx264",
+        "-preset",
+        "veryfast",
+        "-crf",
+        "23",
+        "-c:a",
+        "aac",
+        "-movflags",
+        "+faststart",
+        output_path,
+    ]
+    result = subprocess.run(cmd, capture_output=True, text=True)
+    if result.returncode != 0:
+        raise RuntimeError(result.stderr.strip() or "ffmpeg failed")
+
+
+# -- Move DB ---
+os.environ["STREAMLIT_DATA_DIRECTORY"] = "/tmp/.streamlit"
+Path("/tmp/.streamlit").mkdir(parents=True, exist_ok=True)
+Path("/tmp/data").mkdir(parents=True, exist_ok=True)
+source_db = "init_data/veureu.db"
+target_db = "/tmp/data/app.db"
+if not os.path.exists(target_db) and os.path.exists(source_db):
+    shutil.copy(source_db, target_db)
+
+static_videos = Path(__file__).parent / "videos"
+runtime_videos = Path("/tmp/data/videos")
+if not runtime_videos.exists():
+    shutil.copytree(static_videos, runtime_videos, dirs_exist_ok=True)
+
+
+# --- Config ---
+def _load_yaml(path="config.yaml") -> dict:
+    with open(path, "r", encoding="utf-8") as f:
+        cfg = yaml.safe_load(f) or {}
+    # interpolación sencilla de ${VARS} si las usas en el YAML
+    def _subst(s: str) -> str:
+        return os.path.expandvars(s) if isinstance(s, str) else s
+
+    # aplica sustitución en los campos que te interesan
+    if "api" in cfg:
+        cfg["api"]["base_url"] = _subst(cfg["api"].get("base_url", ""))
+        cfg["api"]["token"]    = _subst(cfg["api"].get("token", ""))
+
+    if "storage" in cfg and "root_dir" in cfg["storage"]:
+        cfg["storage"]["root_dir"] = _subst(cfg["storage"]["root_dir"])
+
+    if "sqlite" in cfg and "path" in cfg["sqlite"]:
+        cfg["sqlite"]["path"] = _subst(cfg["sqlite"]["path"])
+
+    return cfg
+
+CFG = _load_yaml("config.yaml")
+
+# Ajuste de variables según tu esquema YAML
+DATA_DIR = CFG.get("storage", {}).get("root_dir", "data")
+BACKEND_BASE_URL = CFG.get("api", {}).get("base_url", "http://localhost:8000")
+USE_MOCK = bool(CFG.get("app", {}).get("use_mock", False))  # si no la tienes en el yaml, queda False
+API_TOKEN = CFG.get("api", {}).get("token") or os.getenv("API_SHARED_TOKEN")
+
+os.makedirs(DATA_DIR, exist_ok=True)
+ensure_dirs(DATA_DIR)
+DB_PATH = os.path.join(DATA_DIR, "app.db")
+set_db_path(DB_PATH)
+init_schema()
+
+api = APIClient(BACKEND_BASE_URL, use_mock=USE_MOCK, data_dir=DATA_DIR, token=API_TOKEN)
+
+st.set_page_config(page_title="Veureu — Audiodescripció", page_icon="🎬", layout="wide")
+
+# --- Session: auth ---
+# print("Usuarios disponibles:", get_all_users()) # Descomentar para depurar
+if "user" not in st.session_state:
+    st.session_state.user = None  # dict with {username, role, id(optional)}
+
+def require_login():
+    if not st.session_state.user:
+        st.info("Por favor, inicia sesión para continuar.")
+        login_form()
+        st.stop()
+
+def verify_password(password: str, pw_hash: str) -> bool:
+    try:
+        return bcrypt.verify(password, pw_hash)
+    except Exception:
+        return False
+        
+# --- Sidebar (only after login) ---
+role = st.session_state.user["role"] if st.session_state.user else None
+with st.sidebar:
+    st.title("Veureu")
+    if st.session_state.user:
+        st.write(f"Usuari: **{st.session_state.user['username']}** (rol: {st.session_state.user['role']})")
+        if st.button("Tancar sessió"):
+            st.session_state.user = None
+            st.rerun()
+    if st.session_state.user:
+        page = st.radio("Navegació", ["Analitzar video-transcripcions","Processar vídeo nou","Estadístiques"], index=0)
+    else:
+        page = None
+
+# --- Pre-login screen ---
+if not st.session_state.user:
+    st.title("Veureu — Audiodescripció")
+    def login_form():
+        st.subheader("Inici de sessió")
+        username = st.text_input("Usuari")
+        password = st.text_input("Contrasenya", type="password")
+        if st.button("Entrar", type="primary"):
+            row = get_user(username)
+            if row and verify_password(password, row["pw_hash"]):
+                st.session_state.user = {"id": row["id"], "username": row["username"], "role": row["role"]}
+                st.success(f"Benvingut/da, {row['username']}")
+                st.rerun()
+            else:
+                st.error("Credencials invàlides")
+    login_form()
+    st.stop()
+
+# --- Pages ---
+if page == "Processar vídeo nou":
+    require_login()
+    if role != "verd":
+        st.error("No tens permisos per processar nous vídeos. Canvia d'usuari o sol·licita permisos.")
+        st.stop()
+
+    st.header("Processar un nou clip de vídeo")
+
+    # Inicializar el estado de la página si no existe
+    if 'video_uploaded' not in st.session_state:
+        st.session_state.video_uploaded = None
+    if 'characters_detected' not in st.session_state:
+        st.session_state.characters_detected = None
+    if 'characters_saved' not in st.session_state:
+        st.session_state.characters_saved = False
+
+    # --- 1. Subida del vídeo ---
+    MAX_SIZE_MB = 20
+    MAX_DURATION_S = 240 # 4 minutos
+
+    uploaded_file = st.file_uploader("Puja un clip de vídeo (MP4, < 20MB, < 4 minuts)", type=["mp4"], key="video_uploader")
+
+    if uploaded_file is not None:
+        # Resetear el estado si se sube un nuevo archivo
+        if st.session_state.video_uploaded is None or uploaded_file.name != st.session_state.video_uploaded.get('original_name'):
+            st.session_state.video_uploaded = {'original_name': uploaded_file.name, 'status': 'validating'}
+            st.session_state.characters_detected = None
+            st.session_state.characters_saved = False
+
+        # --- Validación y Procesamiento ---
+        if st.session_state.video_uploaded['status'] == 'validating':
+            is_valid = True
+            # 1. Validar tamaño
+            if uploaded_file.size > MAX_SIZE_MB * 1024 * 1024:
+                st.error(f"El vídeo supera el límit de {MAX_SIZE_MB}MB.")
+                is_valid = False
+
+            if is_valid:
+                with st.spinner("Processant el vídeo..."):
+                    # Guardar temporalmente para analizarlo
+                    temp_path = Path("temp_video.mp4")
+                    with temp_path.open("wb") as f:
                         f.write(uploaded_file.getbuffer())
-                    
-                    clip = VideoFileClip("temp_video.mp4")
-                    duration = clip.duration
 
-                    # 2. Validar y truncar duración
                     was_truncated = False
-                    if duration > MAX_DURATION_S:
-                        clip = clip.subclip(0, MAX_DURATION_S)
-                        was_truncated = True
-
-                    # Crear carpeta y guardar el vídeo final
-                    video_name = Path(uploaded_file.name).stem
-                    video_dir = Path("/tmp/data/videos") / video_name
-                    video_dir.mkdir(parents=True, exist_ok=True)
-                    final_video_path = video_dir / f"{video_name}.mp4"
-                    clip.write_videofile(str(final_video_path), codec="libx264", audio_codec="aac")
-
-                    clip.close()
-                    os.remove("temp_video.mp4")
-
-                    # Actualizar estado
-                    st.session_state.video_uploaded.update({
-                        'status': 'processed',
-                        'path': str(final_video_path),
-                        'was_truncated': was_truncated
-                    })
-                    st.rerun()
+                    final_video_path = None
+                    try:
+                        duration = _get_video_duration(str(temp_path))
+                        if not duration:
+                            st.error("No s'ha pogut obtenir la durada del vídeo.")
+                            is_valid = False
+
+                        if is_valid:
+                            if duration > MAX_DURATION_S:
+                                was_truncated = True
+
+                            video_name = Path(uploaded_file.name).stem
+                            video_dir = Path("/tmp/data/videos") / video_name
+                            video_dir.mkdir(parents=True, exist_ok=True)
+                            final_video_path = video_dir / f"{video_name}.mp4"
+
+                            try:
+                                _transcode_video(
+                                    str(temp_path),
+                                    str(final_video_path),
+                                    MAX_DURATION_S if was_truncated else None,
+                                )
+                            except RuntimeError as exc:
+                                st.error(f"No s'ha pogut processar el vídeo: {exc}")
+                                is_valid = False
+
+                        if is_valid and final_video_path is not None:
+                            st.session_state.video_uploaded.update({
+                                'status': 'processed',
+                                'path': str(final_video_path),
+                                'was_truncated': was_truncated
+                            })
+                            st.rerun()
+                    finally:
+                        if temp_path.exists():
+                            temp_path.unlink()
 
     # --- Mensajes de estado ---
     if st.session_state.video_uploaded and st.session_state.video_uploaded['status'] == 'processed':

requirements.txt

@@ -8,9 +8,6 @@ python-dotenv
 gradio_client  # Para llamar al space svision
 Pillow         # Para procesar imágenes antes de enviar a svision
 passlib[bcrypt]
-moviepy
-imageio
-imageio-ffmpeg
 streamlit-authenticator>=0.2.3
 web3>=6.0.0   # Para integración con Polygon blockchain
 # Forzar rebuild 2025-11-03