Update app.py

app.py

@@ -5,32 +5,12 @@ import json
 import os
 import requests
 import torch
-import numpy as np
+import torchaudio
 from datetime import timedelta
 from pyannote.audio import Pipeline
-from huggingface_hub import login, hf_hub_download
+from huggingface_hub import login
 from pydub import AudioSegment
 
-# --- Safe Globals for PyTorch 2.6+ ---
-try:
-    from pyannote.audio.core.task import Specifications, Problem, Resolution
-    from pyannote.audio.core.model import Model
-    from pyannote.audio.pipelines.speaker_diarization import SpeakerDiarization
-    
-    torch.serialization.add_safe_globals([
-        torch.torch_version.TorchVersion,
-        Specifications,
-        Problem,
-        Resolution,
-        Model,
-        SpeakerDiarization,
-        np.dtype,
-        torch.nn.modules.container.ModuleList,
-        np.core.multiarray.scalar
-    ])
-except Exception as e:
-    print(f"Safe Globals Warning: {e}")
-
 # --- Configuration & Tokens ---
 HARDCODED_HF_TOKEN = "PASTE_YOUR_HF_TOKEN_HERE"
 HARDCODED_GEMINI_KEY = ""
@@ -71,7 +51,7 @@ def generate_cmx_edl(edl_title, segments, source_name, fps=25):
 def call_gemini_for_edl(transcript_data, story_prompt, api_key):
     """Sends diarized, word-level transcript to Gemini Senior Editor."""
     if not api_key:
-        st.error("Gemini API Key is missing. Set it in Space Secrets or app.py.")
+        st.error("Gemini API Key is missing.")
         return None
         
     url = f"https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash-preview-09-2025:generateContent?key={api_key}"
@@ -106,7 +86,7 @@ def call_gemini_for_edl(transcript_data, story_prompt, api_key):
 
 # --- Streamlit UI ---
 st.set_page_config(page_title="DocAI Editor", layout="wide")
-st.title("Documentary AI: Pipeline (Stable v2.1)")
+st.title("Documentary AI: Pipeline (Stable 2.1)")
 
 with st.sidebar:
     st.header("Project Settings")
@@ -127,14 +107,13 @@ if uploaded_file:
             if not ACTIVE_HF_TOKEN or "PASTE_YOUR_HF_TOKEN" in ACTIVE_HF_TOKEN:
                 st.error("Please provide a valid Hugging Face Token.")
             else:
-                with st.spinner("Processing... This may take a moment."):
-                    # Save local temp file
+                with st.spinner("Processing..."):
                     with open("temp_input", "wb") as f:
                         f.write(uploaded_file.getbuffer())
                     
                     st.write("🎵 **Preprocessing Audio...**")
                     try:
-                        # Use PyDub to convert to WAV (Mono, 16kHz)
+                        # Use PyDub for robust WAV conversion
                         audio = AudioSegment.from_file("temp_input")
                         audio = audio.set_channels(1)
                         audio = audio.set_frame_rate(16000)
@@ -147,15 +126,15 @@ if uploaded_file:
                     st.write("🗣️ **Running Speaker Diarization...**")
                     diarization = None
                     try:
-                        # FIX: Pass revision separately, remove @ from repo_id
-                        config_path = hf_hub_download(
-                            repo_id="pyannote/speaker-diarization", 
-                            revision="2.1",  # Explicit revision
-                            filename="config.yaml", 
-                            token=ACTIVE_HF_TOKEN
-                        )
+                        # Login with token first
+                        login(token=ACTIVE_HF_TOKEN)
                         
-                        pipeline = Pipeline.from_pretrained(config_path)
+                        # Load pipeline using legacy API (use_auth_token is valid here)
+                        # NOTE: Using the older model ID for 2.1 compatibility
+                        pipeline = Pipeline.from_pretrained(
+                            "pyannote/speaker-diarization@2.1",
+                            use_auth_token=ACTIVE_HF_TOKEN
+                        )
                         
                         if torch.cuda.is_available():
                             st.write("🚀 Using GPU for Diarization")
@@ -181,13 +160,16 @@ if uploaded_file:
                     
                     if diarization:
                         # 2.1.1 returns a proper Annotation object directly
-                        for turn, _, speaker_id in diarization.itertracks(yield_label=True):
-                            speaker_turns.append({"start": turn.start, "end": turn.end, "speaker": speaker_id})
-                        
-                        if len(speaker_turns) > 0:
-                            st.write(f"✅ Found {len(speaker_turns)} speaker turns.")
-                        else:
-                            st.warning("⚠️ Pipeline ran but returned no tracks.")
+                        try:
+                            for turn, _, speaker_id in diarization.itertracks(yield_label=True):
+                                speaker_turns.append({"start": turn.start, "end": turn.end, "speaker": speaker_id})
+                            
+                            if len(speaker_turns) > 0:
+                                st.write(f"✅ Found {len(speaker_turns)} speaker turns.")
+                            else:
+                                st.warning("⚠️ Pipeline ran but returned no tracks.")
+                        except AttributeError:
+                             st.error("Could not iterate tracks.")
 
                     for segment in result['segments']:
                         mid_time = (segment['start'] + segment['end']) / 2