start.sh

#!/bin/bash

# Hugging Face Spaces startup script for OpenClaw

echo "🦞 Starting OpenClaw on Hugging Face Spaces..."

# Create config directory
mkdir -p /root/.openclaw/workspace

# Check required env vars
if [ -z "$API_KEY" ] || [ -z "$BASE_URL" ]; then
    echo "⚠️ Warning: API_KEY or BASE_URL not set. AI features won't work."
fi

# ==========================================
# GITHUB STATE SYNC (STARTUP PULL)
# ==========================================
OPENCLAW_DIR="/root/.openclaw"

if [ -n "$MEMORY_REPO_URL" ] && [ -n "$MEMORY_REPO_TOKEN" ]; then
    echo "🔄 Memory Sync Enabled! Attempting to restore state from GitHub..."
    
    # Extract domain and path from repo URL to format it with the token
    # Supports formats like: https://github.com/user/repo or github.com/user/repo
    REPO_CLEAN=$(echo "$MEMORY_REPO_URL" | sed -e 's|^[^/]*//||' -e 's|^.*@||')
    AUTH_REPO_URL="https://${MEMORY_REPO_TOKEN}@${REPO_CLEAN}"
    
    # Use a temporary directory for cloning to avoid conflict with existing config
    SYNC_TMP="/tmp/openclaw_sync"
    rm -rf "$SYNC_TMP"
    
    # Try cloning the repository
    if git clone "$AUTH_REPO_URL" "$SYNC_TMP" 2>/dev/null; then
        echo "✅ Successfully downloaded state from GitHub repository."
        
        # Copy over the state, ensuring the destination directory exists
        mkdir -p "$OPENCLAW_DIR"
        
        # We use rsync or standard cp to move contents (excluding .git from remote)
        # Note: copying everything EXCEPT the remote's .git to our local working dir.
        if [ "$(ls -A $SYNC_TMP)" ]; then
            cp -r "$SYNC_TMP/." "$OPENCLAW_DIR/"
            rm -rf "$OPENCLAW_DIR/.git"
        fi
        
        # Initialize Git in the actual OpenClaw directory so we can push back later
        cd "$OPENCLAW_DIR"
        git init
        git config user.email "openclaw-bot@huggingface.co"
        git config user.name "OpenClaw Auto-Sync"
        git remote add origin "$AUTH_REPO_URL"
        
        # Force the branch name if needed (assumes main)
        git checkout -B main
        # Reset local to match remote exactly
        git fetch origin main 2>/dev/null || true
        git reset --soft origin/main 2>/dev/null || true
        
        echo "✅ State restoration complete!"
    else
        echo "❌ Failed to clone repository. Is the token valid and URL format correct? Starting fresh..."
    fi
else
    echo "ℹ️ MEMORY_REPO_URL or MEMORY_REPO_TOKEN not provided. State will NOT be persisted across restarts."
fi

# ==========================================
# CLEAN STALE SESSION LOCKS
# ==========================================
# When Hugging Face spaces restart unexpectedly, lock files might be left behind.
# Because Docker PIDs start at 1 again, OpenClaw might think the old lock is still active
# and time out trying to access the session. We wipe them cleanly on boot.
if [ -d "/root/.openclaw/agents" ]; then
    echo "🧹 Cleaning up any stale session lock files on boot..."
    find /root/.openclaw/agents -type f -name "*.jsonl.lock" -exec rm -f {} +
fi

# Set defaults
BASE_URL=${BASE_URL:-"http://127.0.0.1:20128/v1"}
MODEL_ID=${MODEL_ID:-"Multiple-models"}
PROVIDER_NAME=${PROVIDER_NAME:-"custom"}

# Create necessary OpenClaw memory files to prevent ENOENT errors
mkdir -p /root/.openclaw/workspace/memory
touch /root/.openclaw/workspace/MEMORY.md
touch "/root/.openclaw/workspace/memory/$(date -u +%Y-%m-%d).md"
touch "/root/.openclaw/workspace/memory/$(date -u -d "yesterday" +%Y-%m-%d 2>/dev/null || date -u -v -1d +%Y-%m-%d 2>/dev/null || echo "").md" 2>/dev/null || true

# Create OpenClaw config
cat > /root/.openclaw/openclaw.json << EOF
{
  "meta": {
    "lastTouchedVersion": "2026.2.19-2"
  },
  "models": {
    "providers": {
      "${PROVIDER_NAME}": {
        "baseUrl": "${BASE_URL}",
        "apiKey": "${API_KEY}",
        "api": "openai-completions",
        "models": [
          { "id": "${MODEL_ID}", "name": "${MODEL_ID}" }
        ]
      }
    }
  },
  "agents": {
    "defaults": {
      "model": { "primary": "${PROVIDER_NAME}/${MODEL_ID}" },
      "workspace": "/root/.openclaw/workspace",
      "timeoutSeconds": 300,
      "lockTimeoutMs": 120000
    }
  },
  "commands": { "native": "auto", "nativeSkills": "auto", "restart": false },
  "hooks": {
    "internal": {
      "enabled": true,
      "entries": {
        "boot-md": { "enabled": true },
        "command-logger": { "enabled": true },
        "session-memory": { "enabled": true }
      }
    }
  },
  "session": { "dmScope": "per-channel-peer" },
  "gateway": {
    "port": 7860,
    "mode": "local",
    "bind": "lan",
    "controlUi": {
      "dangerouslyDisableDeviceAuth": true
    },
    "trustedProxies": ["0.0.0.0/0"],
    "tailscale": { "mode": "off" }
  },
  "skills": { "install": { "nodeManager": "npm" } },
  "plugins": { "entries": {} }
}
EOF

echo "✅ Config created"
echo "📡 API: ${BASE_URL}"
echo "🤖 Model: ${MODEL_ID}"
echo ""
echo "🌐 Starting OpenClaw Gateway..."
echo "📱 Open your Space URL to access the dashboard"
echo "🔑 Note: OpenClaw will natively generate a Gateway Token on its first run. Check these initialization logs below to find it!"
echo ""

# Start OpenClaw gateway directly in the background
openclaw gateway &
GATEWAY_PID=$!

sleep 4

# Run a background task to display the token right after initialization
(
  echo ""
  echo "=========================================================================="
  # Try to grab the token from the updated config file
  TOKEN=$(grep -E '"token"\s*:' /root/.openclaw/openclaw.json | head -1 | awk -F '"' '{print $4}')
  
  if [ -z "$TOKEN" ] || [ "$TOKEN" == "null" ]; then
      # Fallback to CLI command
      TOKEN=$(openclaw config get gateway.auth.token 2>/dev/null)
  fi
  
  echo "🔑 GENERATED GATEWAY TOKEN: $TOKEN"
  echo "👉 Use this token in the OpenClaw pairing screen!"
  echo "=========================================================================="
  echo ""
)

# ==========================================
# GITHUB STATE SYNC (CONTINUOUS PUSH)
# ==========================================
if [ -n "$MEMORY_REPO_URL" ] && [ -n "$MEMORY_REPO_TOKEN" ]; then
    echo "🔄 Starting background memory synchronization..."
    
    # Extract domain and path again just in case for the fallback initialization
    REPO_CLEAN=$(echo "$MEMORY_REPO_URL" | sed -e 's|^[^/]*//||' -e 's|^.*@||')
    AUTH_REPO_URL="https://${MEMORY_REPO_TOKEN}@${REPO_CLEAN}"

    (
      # Ensure working directory exists before jumping in
      mkdir -p "$OPENCLAW_DIR"
      cd "$OPENCLAW_DIR" || exit
      
      # If for some reason git wasn't initialized during the startup step (e.g., clone failed because repo was empty),
      # we initialize it here as a fallback before entering the infinite loop.
      if [ ! -d ".git" ]; then
          echo "⚠️ Git not initialized in $OPENCLAW_DIR. Initializing fallback..."
          git init
          git config user.email "openclaw-bot@huggingface.co"
          git config user.name "OpenClaw Auto-Sync"
          git remote add origin "$AUTH_REPO_URL"
          git checkout -B main
      fi
      
      # Ensure .gitignore prevents pushing bad/heavy files
      cat > .gitignore << 'IGNORE'
node_modules/
.env
*.jsonl.lock
IGNORE

      # Untrack any lock files that may have accidentally been added previously
      find "$OPENCLAW_DIR" -type f -name "*.jsonl.lock" -exec git rm --cached {} + 2>/dev/null || true

      while true; do
          # Sync every 30 seconds
          sleep 30
          
          # Remove any nested .git directories that confuse our top-level git repo
          # This prevents "error: 'workspace/' does not have a commit checked out"
          find "$OPENCLAW_DIR" -mindepth 2 -type d -name ".git" -exec rm -rf {} +
          
          # Add all changes
          git add .
          
          # Commit will fail cleanly if there are no changes, acting as our checker
          if git commit -m "Auto-sync OpenClaw state [$(date -u)]" >/dev/null 2>&1; then
              echo "[$(date -u)] 🔄 Syncing OpenClaw memory state to GitHub..."
              
              # Try to pull in case of external changes to avoid conflicts, then push
              git pull --rebase origin main || echo "⚠️ Git pull failed"
              git push -u origin main || echo "❌ Git push failed! Please check token permissions."
              # echo "[$(date -u)] ✅ Memory state saved."
          fi
      done
    ) &
fi

# Wait for the main gateway process
wait $GATEWAY_PID