start.sh

#!/bin/bash
set -e

echo "🚀 Starting Vane on Hugging Face Spaces..."

CONFIG_FILE="/app/config.toml"

# ── Inject into config.toml ───────────────────────────────────────────────────
if [ -n "$SEARXNG_URL" ]; then
    echo "✅ SearXNG URL found — injecting into config..."
    sed -i "s|SEARXNG = \"\"|SEARXNG = \"$SEARXNG_URL\"|g" "$CONFIG_FILE"
else
    echo "❌ ERROR: SEARXNG_URL secret is not set!"
    exit 1
fi

if [ -n "$GROQ_API_KEY" ]; then
    sed -i "s|GROQ = \"\"|GROQ = \"$GROQ_API_KEY\"|g" "$CONFIG_FILE"
fi
if [ -n "$OPENAI_API_KEY" ]; then
    sed -i "s|OPENAI = \"\"|OPENAI = \"$OPENAI_API_KEY\"|g" "$CONFIG_FILE"
fi
if [ -n "$ANTHROPIC_API_KEY" ]; then
    sed -i "s|ANTHROPIC = \"\"|ANTHROPIC = \"$ANTHROPIC_API_KEY\"|g" "$CONFIG_FILE"
fi
if [ -n "$OLLAMA_URL" ]; then
    sed -i "s|OLLAMA = \"\"|OLLAMA = \"$OLLAMA_URL\"|g" "$CONFIG_FILE"
fi

if [ -z "$OPENAI_API_KEY" ] && [ -z "$GROQ_API_KEY" ] && [ -z "$ANTHROPIC_API_KEY" ] && [ -z "$OLLAMA_URL" ]; then
    echo "❌ ERROR: No LLM configured!"
    exit 1
fi

# ── Determine provider + model ────────────────────────────────────────────────
if [ -n "$GROQ_API_KEY" ]; then
    PROVIDER="groq"
    API_KEY="$GROQ_API_KEY"
    MODEL="llama-3.1-8b-instant"
elif [ -n "$OPENAI_API_KEY" ]; then
    PROVIDER="openai"
    API_KEY="$OPENAI_API_KEY"
    MODEL="gpt-4o-mini"
elif [ -n "$ANTHROPIC_API_KEY" ]; then
    PROVIDER="anthropic"
    API_KEY="$ANTHROPIC_API_KEY"
    MODEL="claude-3-5-haiku-20241022"
elif [ -n "$OLLAMA_URL" ]; then
    PROVIDER="ollama"
    API_KEY="$OLLAMA_URL"
    MODEL="qwen2.5:3b"
fi

# ── Restore DB from HF Dataset repo ──────────────────────────────────────────
mkdir -p /app/data

if [ -n "$HF_TOKEN" ] && [ -n "$HF_DATASET_REPO" ]; then
    echo "📥 Restoring DB from HF Dataset repo..."
    python3 -c "
from huggingface_hub import hf_hub_download
import shutil, os
try:
    path = hf_hub_download(
        repo_id='$HF_DATASET_REPO',
        filename='vane.db',
        repo_type='dataset',
        token='$HF_TOKEN'
    )
    os.makedirs('/app/data', exist_ok=True)
    shutil.copy(path, '/app/data/vane.db')
    print('✅ DB restored from HF Dataset!')
except Exception as e:
    print(f'⚠️ No existing DB found, starting fresh: {e}')
"
else
    echo "⚠️ HF_TOKEN or HF_DATASET_REPO not set — storage will be ephemeral!"
fi

# ── Run DB migrations ─────────────────────────────────────────────────────────
echo "🗄️ Running database migrations..."
cd /app
npx drizzle-kit migrate 2>/dev/null || echo "⚠️ Migration skipped"

# ── Seed DB with correct key names ───────────────────────────────────────────
echo "🌱 Seeding database settings..."
python3 -c "
import sqlite3, os

db_path = '/app/data/vane.db'
if not os.path.exists(db_path):
    print('⚠️ DB not found, creating...')
    os.makedirs('/app/data', exist_ok=True)

conn = sqlite3.connect(db_path)
cur = conn.cursor()

cur.execute('''CREATE TABLE IF NOT EXISTS settings (
    key TEXT PRIMARY KEY,
    value TEXT
)''')

provider = '$PROVIDER'
api_key  = '$API_KEY'
model    = '$MODEL'
searxng  = '$SEARXNG_URL'
ollama   = os.environ.get('OLLAMA_URL', '')

settings = [
    ('setup_complete',              'true'),
    ('chatModelProvider',           provider),
    ('chatModel',                   model),
    ('embeddingModelProvider',      'local'),
    ('embeddingModel',              'xenova/all-MiniLM-L6-v2'),
    ('defaultChatModelProvider',    provider),
    ('defaultChatModel',            model),
    ('defaultEmbeddingModelProvider', 'local'),
    ('defaultEmbeddingModel',       'xenova/all-MiniLM-L6-v2'),
    ('groqApiKey',                  os.environ.get('GROQ_API_KEY', '')),
    ('openaiApiKey',                os.environ.get('OPENAI_API_KEY', '')),
    ('anthropicApiKey',             os.environ.get('ANTHROPIC_API_KEY', '')),
    ('ollamaApiUrl',                ollama),
    ('searxngUrl',                  searxng),
]

for key, value in settings:
    cur.execute('INSERT OR REPLACE INTO settings (key, value) VALUES (?, ?)', (key, value))

conn.commit()
print('✅ DB seeded with correct keys!')

# Verify
cur.execute('SELECT key, value FROM settings')
for row in cur.fetchall():
    k, v = row
    if 'key' in k.lower() or 'Key' in k:
        v = v[:8] + '...' if len(v) > 8 else v  # mask keys
    print(f'  {k} = {v}')

conn.close()
"

# ── Write backup script ───────────────────────────────────────────────────────
cat > /backup.sh << 'BACKUPSCRIPT'
#!/bin/bash
echo "🔄 DB backup loop started..."
while true; do
    sleep 600
    if [ -f "/app/data/vane.db" ] && [ -n "$HF_TOKEN" ] && [ -n "$HF_DATASET_REPO" ]; then
        echo "💾 Backing up DB to HF Dataset..."
        python3 -c "
from huggingface_hub import HfApi
import os
api = HfApi()
try:
    api.upload_file(
        path_or_fileobj='/app/data/vane.db',
        path_in_repo='vane.db',
        repo_id=os.environ['HF_DATASET_REPO'],
        repo_type='dataset',
        token=os.environ['HF_TOKEN']
    )
    print('✅ DB backed up!')
except Exception as e:
    print(f'❌ Backup failed: {e}')
"
    fi
done
BACKUPSCRIPT
chmod +x /backup.sh

mkdir -p /var/log/supervisor
echo "✅ All config done"
echo "🔧 Starting services..."

exec /usr/bin/supervisord -c /etc/supervisor/conf.d/supervisord.conf