deploy: switch to chatterbox requirements @ a2451a8

.dockerignore

@@ -1,10 +0,0 @@
-frontend/
-tmp/
-data/
-batch_outputs/
-uploads/
-outputs/
-.venv/
-.git/
-*.mp4
-wiki/

.env.example

@@ -4,6 +4,12 @@
 # Server port (default 8000)
 PORT=8000
 
+# Where per-job artifact folders get written. On HF Spaces this is resolved
+# automatically (/data/jobs with persistent storage, /tmp/videovoice_jobs
+# without). For local dev, set this to ./data so jobs land next to the repo
+# — same layout the old `main` used.
+ARTIFACTS_ROOT=./data
+
 # OpenAI API key (for translation step)
 OPENAI_API_KEY=sk-...
 

.gitattributes

@@ -0,0 +1,10 @@
+# Files in this repo that are dev-only and must NOT ship to the HF Spaces.
+# `deploy.sh` honors this via `git archive --worktree-attributes`.
+# Rule of thumb: if HF Spaces would never import/execute it, export-ignore it.
+# Do NOT export-ignore server.py — app.py imports from it at runtime on HF.
+
+.github/          export-ignore
+SPLIT_STRATEGY.md export-ignore
+deploy.sh         export-ignore
+Dockerfile        export-ignore
+.dockerignore     export-ignore

Dockerfile

@@ -1,32 +0,0 @@
-FROM python:3.11-slim-bookworm
-
-RUN apt-get update && apt-get install -y --no-install-recommends \
-    ffmpeg \
-    git \
-    curl \
-    build-essential \
-    libsndfile1 \
-    && rm -rf /var/lib/apt/lists/*
-
-RUN curl -LsSf https://astral.sh/uv/install.sh | sh && ln -s /root/.local/bin/uv /usr/local/bin/uv
-
-WORKDIR /app
-
-COPY pyproject.toml uv.lock ./
-RUN uv sync --frozen --no-dev
-
-COPY scripts/prefetch_models.py /app/scripts/prefetch_models.py
-ARG PREFETCH_MODELS=1
-RUN if [ "$PREFETCH_MODELS" = "1" ]; then HF_HOME=/data/hf-cache uv run python scripts/prefetch_models.py; fi
-
-COPY . /app
-
-ENV HF_HOME=/data/hf-cache \
-    PORT=7860 \
-    ARTIFACTS_ROOT=/data/jobs
-
-RUN mkdir -p /data/hf-cache /data/jobs
-
-EXPOSE 7860
-
-CMD ["uv", "run", "uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]

README.md

@@ -77,6 +77,8 @@ uv run python server.py
 
 The app will be available at [http://localhost:8000](http://localhost:8000).
 
+Per-job artifacts (uploads, intermediate audio, outputs) land in `ARTIFACTS_ROOT`. Set `ARTIFACTS_ROOT=./data` in your `.env` to match the layout the repo used historically — each job gets its own `data/<job_id>/` folder with every pipeline file.
+
 ### CLI Usage
 
 You can also run the pipeline directly:
@@ -268,9 +270,40 @@ VideoVoice/
 
 ---
 
+## Entrypoints
+
+Two files intentionally exist, run in different contexts, but **ship the same code**:
+
+| File | When it runs | What it does |
+|------|-------------|--------------|
+| `server.py` | Local dev (`uv run python server.py`) | Plain FastAPI app — defines every `/api/*` route. |
+| `app.py`    | Hugging Face Spaces               | Gradio Server that imports `server.py`'s router and wraps it with `@spaces.GPU` for ZeroGPU. |
+
+`app.py` depends on `server.py`, so server.py must ship to HF. Do not strip it.
+
 ## Deployment
 
-### AWS (Recommended for GPU)
+### Hugging Face Spaces (production)
+
+Push to `main` → GitHub Actions runs `.github/workflows/deploy-hf.yml` → both Spaces (`Rafii/videovoice` and `Rafii/videovoice-omni`) redeploy automatically. No manual step.
+
+One-time CI setup:
+1. Create an HF access token with write access to both Spaces: https://huggingface.co/settings/tokens
+2. Add it as `HF_TOKEN` under **Settings → Secrets and variables → Actions** in the GitHub repo.
+
+Manual fallback (from a local clean checkout with `space` and `space-omni` remotes configured):
+```bash
+./deploy.sh          # skips if remote is already at HEAD
+./deploy.sh --force  # always redeploy
+```
+
+Files filtered out of every Space deploy are listed in `.gitattributes` (`export-ignore`).
+
+### Branching
+
+`main` is canonical. Use short-lived `feat/<thing>` branches, open a PR, merge, delete. Never maintain a parallel deploy branch — every change on main reaches both Spaces via CI.
+
+### AWS (alternative GPU host)
 
 ```bash
 # On a g4dn.xlarge instance

SPLIT_STRATEGY.md

@@ -1,113 +0,0 @@
-# TTS Backend Split Implementation
-
-## Summary
-
-Successfully implemented the TTS backend split strategy to resolve the dependency conflict between `chatterbox-tts` (requires `transformers==5.2.0`) and `omnivoice` (requires `transformers>=4.57.3`).
-
-## Changes Made
-
-### 1. Split Requirements Files
-- **`requirements-cbox.txt`** — All dependencies + `chatterbox-tts==0.1.7 --no-deps`
-- **`requirements-omni.txt`** — All dependencies + `omnivoice>=0.1.4`
-- **`deploy.sh`** — Deploy script that swaps requirements.txt before pushing to each HF Space
-
-### 2. Updated `pyproject.toml`
-- Removed `chatterbox-tts` and `omnivoice` from base dependencies
-- Added `[project.optional-dependencies]` with `chatterbox`, `omnivoice`, and `all` groups
-- Removed the transformers override (now directly in base dependencies)
-
-### 3. Updated `app.py`
-- Removed the hacky `--no-deps` pip install at startup
-- Added `TTS_ENGINE` env var validation at import time
-- Imports are now guarded based on `TTS_ENGINE` value
-- Space landing page shows which engine is configured
-
-### 4. Updated `server.py`
-- Added `TTS_ENGINE` env var handling
-- `/api/config` now returns only the configured TTS engine
-- Validation endpoints check against `TTS_ENGINE` instead of hardcoded list
-- Removed `preview_both` mode handling (single-engine Spaces don't need it)
-
-### 5. Updated `steps/s4_tts.py`
-- Added `TTS_ENGINE` env var handling
-- Conditional imports: only imports `chatterbox` if `TTS_ENGINE==chatterbox`
-- `synthesise_segments()` validates that requested model matches configured engine
-
-### 6. Updated `steps/s4_preview.py`
-- Added `TTS_ENGINE` env var handling
-- Conditional imports based on `TTS_ENGINE`
-- `generate_previews()` only generates preview for configured engine
-
-### 7. Updated `pipeline.py`
-- Removed `preview` voice mode (no more preview_both flow)
-- Pipeline now respects `TTS_ENGINE` env var in Space deployments
-- Simplified step counting (no conditional preview steps)
-
-## Deployment
-
-### First-Time Setup
-
-1. Add HF Space remotes (if not already configured):
-```bash
-git remote add space https://huggingface.co/spaces/YOUR_USERNAME/videovoice
-git remote add space-omni https://huggingface.co/spaces/YOUR_USERNAME/videovoice-omni
-```
-
-2. Set up the two HF Spaces:
-   - `videovoice` — Set `TTS_ENGINE=chatterbox` in Space Secrets
-   - `videovoice-omni` — Set `TTS_ENGINE=omnivoice` in Space Secrets
-
-### Deploying
-
-```bash
-./deploy.sh
-```
-
-This script will:
-1. Copy `requirements-cbox.txt` → `requirements.txt` and push to `space` remote
-2. Copy `requirements-omni.txt` → `requirements.txt` and push to `space-omni` remote
-3. Restore the original `requirements.txt`
-
-## Frontend Changes Needed
-
-The React frontend should check `/api/config` to determine which TTS engines are available:
-
-```javascript
-const config = await fetch('/api/config').then(r => r.json());
-const availableEngine = config.tts_engine; // "chatterbox" or "omnivoice"
-const availableEngines = config.tts_models; // Array with single engine
-```
-
-The frontend can then route API calls to the appropriate Space based on user selection:
-- Chatterbox: `https://videovoice.hf.space`
-- OmniVoice: `https://videovoice-omni.hf.space`
-
-## Local Development
-
-For local development with both engines:
-
-```bash
-# Install with both TTS engines
-pip install -e ".[all]"
-
-# Or just one
-pip install -e ".[chatterbox]"
-```
-
-The CLI still accepts `--voice-mode chatterbox|omnivoice` for local testing.
-
-## Files Changed
-
-- `pyproject.toml`
-- `app.py`
-- `server.py`
-- `steps/s4_tts.py`
-- `steps/s4_preview.py`
-- `pipeline.py`
-
-## Files Added
-
-- `requirements-cbox.txt`
-- `requirements-omni.txt`
-- `deploy.sh`
-- `SPLIT_STRATEGY.md` (this file)

deploy.sh

@@ -1,178 +0,0 @@
-#!/bin/bash
-# deploy.sh — push to both HuggingFace Spaces with the correct requirements.txt
-# Usage: ./deploy.sh [--force]
-#
-# This script exports the current commit while excluding large media folders,
-# then updates each Space repo on top of its current main branch so commit
-# history is preserved.
-
-set -e
-
-# Colors for output
-RED='\033[0;31m'
-GREEN='\033[0;32m'
-YELLOW='\033[1;33m'
-BLUE='\033[0;34m'
-NC='\033[0m' # No Color
-
-FORCE_DEPLOY=false
-if [ "$1" == "--force" ]; then
-    FORCE_DEPLOY=true
-fi
-
-echo -e "${YELLOW}VideoVoice Dual-Space Deploy Script${NC}"
-echo ""
-
-# Ensure we're in the right directory
-if [ ! -f "app.py" ] || [ ! -d "steps" ]; then
-    echo -e "${RED}Error: Must run from VideoVoice-be root directory${NC}"
-    exit 1
-fi
-
-# Check for uncommitted changes
-if [ -n "$(git status --porcelain)" ]; then
-    echo -e "${YELLOW}Warning: You have uncommitted changes.${NC}"
-    echo "Commit to origin first, then run this deploy script."
-    git status --short
-    echo ""
-    read -p "Continue anyway? (y/N) " -n 1 -r
-    echo
-    if [[ ! $REPLY =~ ^[Yy]$ ]]; then
-        exit 1
-    fi
-fi
-
-REPO_ROOT=$(git rev-parse --show-toplevel)
-LOCAL_SHA=$(git rev-parse HEAD)
-
-# Verify remotes exist
-if ! git remote | grep -q "^space$"; then
-    echo -e "${RED}Error: Missing 'space' remote for Chatterbox Space${NC}"
-    echo "Add it with: git remote add space https://huggingface.co/spaces/Rafii/videovoice"
-    exit 1
-fi
-
-if ! git remote | grep -q "^space-omni$"; then
-    echo -e "${RED}Error: Missing 'space-omni' remote for OmniVoice Space${NC}"
-    echo "Add it with: git remote add space-omni https://huggingface.co/spaces/Rafii/videovoice-omni"
-    exit 1
-fi
-
-echo -e "${GREEN}Remotes configured:${NC}"
-git remote -v | grep -E "^(space|space-omni)"
-echo ""
-
-# Get space URLs for pushing
-SPACE_URL=$(git remote get-url space)
-SPACE_OMNI_URL=$(git remote get-url space-omni)
-
-# Function to check if Space needs update
-_needs_update() {
-    local remote_url=$1
-
-    # Try to get the remote HEAD SHA
-    local remote_sha
-    remote_sha=$(git ls-remote "$remote_url" HEAD 2>/dev/null | awk '{print $1}')
-
-    if [ -z "$remote_sha" ]; then
-        # Can't determine, assume needs update
-        return 0
-    fi
-
-    # Compare with local SHA (if equal, no update needed)
-    if [ "$remote_sha" == "$LOCAL_SHA" ]; then
-        return 1
-    fi
-
-    return 0
-}
-
-# Deploy to one Space while preserving commit history
-_deploy_space() {
-    local target_name=$1
-    local target_url=$2
-    local export_dir_name=$3
-    local source_requirements=$4
-    local commit_message=$5
-    local strip_path=$6
-
-    echo "Creating clean export excluding media files..."
-
-    local temp_dir
-    temp_dir=$(mktemp -d)
-    local export_dir="$temp_dir/$export_dir_name"
-    mkdir -p "$export_dir"
-
-    # Export current HEAD
-    git archive --format=tar HEAD | tar -C "$export_dir" -xf -
-
-    # Remove any lingering media directories
-    rm -rf "$export_dir/data" \
-           "$export_dir/uploads" \
-           "$export_dir/outputs" \
-           "$export_dir/tmp" 2>/dev/null || true
-
-    if [ -n "$strip_path" ]; then
-        rm -rf "$export_dir/$strip_path"
-    fi
-
-    local repo_dir="$temp_dir/${export_dir_name}-repo"
-
-    # Clone remote so we keep full history instead of force-replacing it
-    git clone "$target_url" "$repo_dir"
-    cd "$repo_dir"
-
-    # Ensure we are on main and up to date
-    git checkout -B main origin/main 2>/dev/null || git checkout -b main
-    git pull --ff-only origin main 2>/dev/null || true
-
-    # Replace tracked files with clean export content
-    find . -mindepth 1 -maxdepth 1 \
-        ! -name .git \
-        ! -name .gitattributes \
-        -exec rm -rf {} +
-
-    cp -R "$export_dir"/. .
-
-    # Apply target requirements variant
-    cp "$source_requirements" requirements.txt
-
-    git add -A
-
-    if git diff --cached --quiet; then
-        echo -e "${BLUE}  No file changes for ${target_name}, skipping push.${NC}"
-    else
-        git commit -m "$commit_message @ ${LOCAL_SHA:0:7}" --no-verify
-        git push "$target_url" HEAD:main
-        echo -e "${GREEN}✓ ${target_name} deployed${NC}"
-    fi
-
-    cd "$REPO_ROOT"
-    rm -rf "$temp_dir"
-}
-
-# Deploy to Chatterbox Space
-echo -e "${YELLOW}>>> Checking Chatterbox Space (videovoice)...${NC}"
-if ! $FORCE_DEPLOY && ! _needs_update "$SPACE_URL"; then
-    echo -e "${BLUE}  Chatterbox Space is already at commit ${LOCAL_SHA:0:7}, skipping.${NC}"
-    echo -e "${BLUE}  Use --force to redeploy anyway.${NC}"
-else
-    _deploy_space "Chatterbox Space" "$SPACE_URL" "videovoice-deploy" "requirements-cbox.txt" "deploy: switch to chatterbox requirements" ""
-fi
-echo ""
-
-# Deploy to OmniVoice Space
-echo -e "${YELLOW}>>> Checking OmniVoice Space (videovoice-omni)...${NC}"
-if ! $FORCE_DEPLOY && ! _needs_update "$SPACE_OMNI_URL"; then
-    echo -e "${BLUE}  OmniVoice Space is already at commit ${LOCAL_SHA:0:7}, skipping.${NC}"
-    echo -e "${BLUE}  Use --force to redeploy anyway.${NC}"
-else
-    _deploy_space "OmniVoice Space" "$SPACE_OMNI_URL" "videovoice-omni-deploy" "requirements-omni.txt" "deploy: switch to omnivoice requirements" "chatterbox"
-fi
-echo ""
-
-echo -e "${GREEN}✓ Deploy check complete!${NC}"
-echo ""
-echo "Space URLs:"
-echo "  - Chatterbox: https://huggingface.co/spaces/Rafii/videovoice"
-echo "  - OmniVoice:  https://huggingface.co/spaces/Rafii/videovoice-omni"