scripts/sync_dataset.sh

#!/bin/bash
set -e

# Configuration
DATASET_REPO="${DATASET_REPO:-https://github.com/personalbotai/picoclaw-memory.git}" # Default: personalbotai/picoclaw-memory
SYNC_INTERVAL="${SYNC_INTERVAL:-300}" # Default: 5 minutes

# Determine home directory
if [ -z "$HOME" ]; then
    echo "HOME not set, defaulting to /root"
    export HOME="/root"
fi

PICOCLAW_HOME="${PICOCLAW_HOME:-$HOME/.picoclaw}"
WORKSPACE_DIR="$PICOCLAW_HOME/workspace"
CONFIG_FILE="$PICOCLAW_HOME/config.json"
BACKUP_DIR="$PICOCLAW_HOME/backup"

echo "Using PICOCLAW_HOME: $PICOCLAW_HOME"
echo "Backup Dir: $BACKUP_DIR"
echo "Workspace Dir: $WORKSPACE_DIR"

# Prevent git from asking for credentials interactively
export GIT_TERMINAL_PROMPT=0

if [ -z "$DATASET_REPO" ]; then
    echo "DATASET_REPO environment variable not set. Skipping dataset sync."
    exit 0
fi

# Check for token or SSH key (implied by absence of error)
if [ -z "$GITHUB_TOKEN" ] && [ ! -f "$HOME/.ssh/id_rsa" ] && [ ! -f "$HOME/.ssh/id_ed25519" ]; then
    echo "Warning: GITHUB_TOKEN not set and no SSH keys found. Sync might fail if repo is private or requires auth."
fi

# Setup Git for Dataset
setup_git() {
    echo "Setting up git for dataset sync..."
    git config --global user.name "${GIT_AUTHOR_NAME:-picoclaw}"
    git config --global user.email "${GIT_AUTHOR_EMAIL:-picoclaw@example.com}"
    
    # Configure credential helper for GitHub if token is present
    if [ -n "$GITHUB_TOKEN" ]; then
        git config --global credential.helper store
        echo "https://${GIT_AUTHOR_NAME}:${GITHUB_TOKEN}@github.com" > ~/.git-credentials
    fi
}

# Initial Clone/Pull
initial_sync() {
    mkdir -p "$BACKUP_DIR"
    
    if [ ! -d "$BACKUP_DIR/.git" ]; then
        echo "Cloning dataset $DATASET_REPO..."
        git clone "$DATASET_REPO" "$BACKUP_DIR" || echo "Clone failed, continuing..."
    else
        echo "Pulling latest changes from dataset..."
        cd "$BACKUP_DIR" && git pull origin main || echo "Pull failed, continuing..."
    fi

    # Restore to workspace if backup has data
    if [ -d "$BACKUP_DIR/workspace" ]; then
        echo "Restoring workspace from backup..."
        mkdir -p "$WORKSPACE_DIR"
        # Use rsync for better synchronization if available, otherwise cp
        if command -v rsync >/dev/null 2>&1; then
            rsync -av --update "$BACKUP_DIR/workspace/" "$WORKSPACE_DIR/"
        else
            cp -r "$BACKUP_DIR/workspace/"* "$WORKSPACE_DIR/" 2>/dev/null || true
        fi
    fi
    
    if [ -f "$BACKUP_DIR/config.json" ]; then
        echo "Restoring config from backup..."
        cp "$BACKUP_DIR/config.json" "$CONFIG_FILE" 2>/dev/null || true
    fi
}

# Sync Loop
sync_loop() {
    echo "Starting sync loop (interval: ${SYNC_INTERVAL}s)..."
    
    # Run sync immediately on start
    sync_now
    
    while true; do
        sleep "$SYNC_INTERVAL"
        sync_now
    done
}

sync_now() {
    echo "Syncing data to dataset..."
    
    # Copy current state to backup dir
    mkdir -p "$BACKUP_DIR/workspace"
    if [ -d "$WORKSPACE_DIR" ]; then
        if command -v rsync >/dev/null 2>&1; then
             rsync -av --update --exclude='.git' "$WORKSPACE_DIR/" "$BACKUP_DIR/workspace/"
        else
             cp -r "$WORKSPACE_DIR/"* "$BACKUP_DIR/workspace/" 2>/dev/null || true
        fi
    fi
    
    if [ -f "$CONFIG_FILE" ]; then
        cp "$CONFIG_FILE" "$BACKUP_DIR/config.json" 2>/dev/null || true
    fi
    
    # Commit and Push
    if [ -d "$BACKUP_DIR/.git" ]; then
        cd "$BACKUP_DIR"
        
        # Security check: Prevent pushing to upstream repository
        REMOTE_URL=$(git remote get-url origin 2>/dev/null || echo "")
        if echo "$REMOTE_URL" | grep -q "sipeed/picoclaw"; then
            echo "SECURITY WARNING: Detected attempt to push to upstream (sipeed/picoclaw). Aborting push."
            return
        fi

        if [[ -n $(git status -s) ]]; then
            git add .
            git commit -m "Auto-sync: $(date '+%Y-%m-%d %H:%M:%S')"
            git push origin main || echo "Push failed, will retry next time..."
            echo "Sync completed successfully."
        else
            echo "No changes to sync."
        fi
    else
        echo "Backup directory is not a git repository. Attempting initial sync..."
        initial_sync
    fi
}

# Main execution
setup_git
initial_sync
sync_loop