REFAC: Update model loading to use staged approach and enhance audio analysis in Streamlit app

REFAC: Enhance emotion vector creation and add ECI timeline calculation in Streamlit app

REFAC: Improve emotion vector creation and add ECI timeline calculation in Streamlit app

REFAC: Enhance UI layout for video analysis results and improve logging configuration in Streamlit app

REFAC: Update model loading to use 'tiny.en' for improved performance in Streamlit app

REFAC: Finalize cache directory setup and improve resource management in Streamlit app

REFAC: Simplify emotion vector creation and enhance video processing logic in Streamlit app

REFAC: Update Streamlit app to enhance emotion analysis and visualization features

REFAC: Improve emotion mapping and display logic in Streamlit app

REFAC: Update requirements and enhance Streamlit app for multimodal emotion analysis

REFAC: Remove unused streamlit app files and clean up imports in streamlit_app.py

FEAT: Preload deepface model from repository

Set DEEPFACE_HOME environment variable to a writable /tmp directory for cache

Refactor audio processing logic to improve handling of stereo audio and ensure minimum audio length for analysis in Streamlit app

Refactor audio processing logic to handle stereo audio conversion and improve temporary file cleanup in Streamlit app

Refactor video processing logic to improve error handling and resource cleanup in Streamlit app

Fix caching directory setup and improve video/audio processing error handling in Streamlit app

Refactor Streamlit app to support video uploads for emotion analysis and enhance audio processing capabilities

Refactor Streamlit app for audio transcription and remove facial emotion recognition logic

Remove redundant packages.txt, rely on SDK defaults

Specify app_file path for Streamlit SDK

Change SDK from docker to streamlit

FIX: Ensure DEEPFACE_HOME is set before import

FIX: Add DEEPFACE_HOME env var to minimal FER test

FEAT: Set Hugging Face cache directory to /tmp for improved compatibility

FEAT: Add cv2 import and set DeepFace cache directory to /tmp for improved compatibility

FEAT: Update image upload functionality and integrate emotion analysis with DeepFace

FIX: Disable XSRF protection for file uploads

FEAT: Add minimal video upload and playback

fix: specify moviepy version in requirements for compatibility

TEST: Deploy minimal app to isolate moviepy installation issue"

refactor: update requirements and set environment variables for model caching

FEAT: Refactor app for post-hoc file upload analysis

refactor: simplify Streamlit app for webcam testing and remove unused components

Switch to full python image for webrtc stability

Pin all major dependencies to stable versions

Allow pip to resolve latest stable dependencies

Add Streamlit config and robust WebRTC settings

fix: update torch and torchaudio versions for compatibility

fix: remove version specifications from requirements.txt for flexibility

fix: update streamlit-webrtc version and adjust numpy requirement

fix: specify cache directory for Whisper model loading

fix: update model caching directories to use /tmp for guaranteed write access

fix: set environment variables for model caching directories

fix: consolidate system dependency installation and remove redundant command

fix: move libgl1-mesa-glx installation to a separate RUN command

fix: correct package name from libgl1-mesa-glxst to libgl1-mesa-glx

"Use --extra-index-url for torch CPU packages"

update base image to Python 3.11 for improved performance and compatibility

Pin dependencies for faster build