Ajout Docker

Dockerfile

@@ -0,0 +1,32 @@
+# Use an official Python runtime as a parent image
+FROM python:3.12-slim
+
+# Install system dependencies
+RUN apt-get update && apt-get install -y \
+    ffmpeg \
+    libsndfile1 \
+    && rm -rf /var/lib/apt/lists/*
+
+# Create a volume for the Whisper model
+VOLUME /root/.cache/huggingface
+
+# Make port 8501 available to the world outside this container
+EXPOSE 8501
+
+# Set the working directory in the container
+WORKDIR /app
+
+# Copy requirements.txt
+COPY requirements.txt /app
+
+# Install any needed packages specified in requirements.txt
+RUN pip install --no-cache-dir -r requirements.txt
+
+# Define environment variable
+ENV NAME=ReelCaptionMaker
+
+# Copy the current directory contents into the container at /app
+COPY . /app
+
+# Run app.py when the container launches
+CMD ["streamlit", "run", "app.py"]

README.md

@@ -13,13 +13,15 @@ Before you begin, make sure you have the following installed on your computer:
 
 1. Open your computer's terminal or command prompt.
 
-2. Clone the repository:# reel-caption-maker
+2. Clone the repository:
 
 ```
 git clone https://github.com/yourusername/reel-caption-maker.git
 cd reel-caption-maker
 ```
 
+### Using Python locally
+
 3. Create a virtual environment:
 - On Windows:
   ```
@@ -50,6 +52,33 @@ streamlit run app.py
 
 3. Your default web browser should open automatically. If it doesn't, copy the URL shown in the terminal (usually http://localhost:8501) and paste it into your browser.
 
+### Using Docker
+
+To run the Reel Caption Maker using Docker, follow these steps:
+
+1. Ensure you have Docker installed on your system. 
+2. If not, you can download and install it from [Docker's official website](https://www.docker.com/get-started).
+3. Build the Docker image:
+   ```
+   docker build -t reel-caption-maker .
+   ```
+4. Run the Docker container:
+   ```
+   docker run -p 8501:8501 -v whisper_model:/root/.cache/huggingface reel-caption-maker
+   ```
+   This command does the following:
+    - Maps port 8501 from the container to port 8501 on your host machine.
+    - Creates a volume named `whisper_model` to persist the Whisper model data.
+
+5. Open your web browser and navigate to `http://localhost:8501` to use the Reel Caption Maker.
+   - Note: The first time you run the container, it may take a few minutes to download the Whisper model. Subsequent runs will be faster as the model will be cached in the Docker volume.
+   - To stop the container, use `Ctrl+C` in the terminal where it's running.
+   - For convenience, you can use the provided `docker-run.sh` script to start the container:
+   ```
+   chmod +x docker-run.sh
+   ./docker-run.sh
+   ```
+
 ## Using Reel Caption Maker
 
 1. Click "Browse files" to upload your video.

app.py

@@ -3,7 +3,7 @@ import os
 import tempfile
 from moviepy.editor import VideoFileClip
 import ffmpeg
-from transformers import WhisperProcessor, WhisperForConditionalGeneration
+from transformers import WhisperProcessor, WhisperForConditionalGeneration, WhisperTokenizer
 import tqdm
 
 # Load Whisper model
@@ -19,39 +19,43 @@ def load_whisper_model():
 
 processor, model = load_whisper_model()
 
-def transcribe_audio(audio_file, language, chunk_length=3):  # Changed default to 3 seconds
+def transcribe_audio(audio_file, language, chunk_length=3):
     if model is None or processor is None:
         st.error("Whisper model is not loaded. Cannot transcribe audio.")
         return []
-    
+
     # Load audio
     audio_input, sr = AudioLoader.load_audio(audio_file)
-    
+
     # Calculate number of samples per chunk
     samples_per_chunk = int(chunk_length * sr)
-    
+
+    # Get the tokenizer
+    tokenizer = WhisperTokenizer.from_pretrained(model.config._name_or_path, language=language)
+
     segments = []
     for i in tqdm.tqdm(range(0, len(audio_input), samples_per_chunk)):
         chunk = audio_input[i:i+samples_per_chunk]
-        
+
         # Pad/trim audio chunk
         chunk_input = processor.feature_extractor(chunk, sampling_rate=sr, return_tensors="pt").input_features
 
         # Generate token ids
-        predicted_ids = model.generate(chunk_input, language=language)
-        
+        forced_decoder_ids = tokenizer.get_decoder_prompt_ids(language=language, task="transcribe")
+        predicted_ids = model.generate(chunk_input, forced_decoder_ids=forced_decoder_ids)
+
         # Decode token ids to text
         transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True)
-        
+
         start_time = i / sr
         end_time = min((i + samples_per_chunk) / sr, len(audio_input) / sr)
-        
+
         segments.append({
             "start": start_time,
             "end": end_time,
             "text": transcription[0].strip()
         })
-    
+
     return segments
 
 def format_srt(segments):

build-local.sh

@@ -0,0 +1 @@
+docker build -t reel-caption-maker .

docker-run.sh

@@ -0,0 +1,18 @@
+#!/bin/bash
+
+# Create a Docker network if it doesn't exist
+docker network create reel-caption-network 2>/dev/null || true
+
+# Remove existing container if it exists
+docker rm -f reel-caption-maker 2>/dev/null || true
+
+# Run the Docker container
+docker run -d \
+  --name reel-caption-maker \
+  --network reel-caption-network \
+  -p 8501:8501 \
+  -v whisper_model:/root/.cache/huggingface \
+  reel-caption-maker
+
+# Print the container logs
+docker logs -f reel-caption-maker