Spaces:

Text-to-Document-Generation
/

Docgenie-API

Paused

Ahadhassan-2003
deploy: update HF Space
dc4e6da 6 months ago
25 kB
# Testing Guide: DocGenie API

Complete guide for testing the document generation API endpoints with Google Drive integration.

## Table of Contents

1. [Prerequisites](#prerequisites)
2. [Quick Start](#quick-start)
3. [Getting Google Drive Token](#getting-google-drive-token)
4. [Testing Async API](#testing-async-api)
5. [Testing Sync PDF API](#testing-sync-pdf-api)
6. [Manual Testing with cURL](#manual-testing-with-curl)
7. [Frontend Integration Example](#frontend-integration-example)
8. [Troubleshooting](#troubleshooting)

---

## Prerequisites

### 1. Start Required Services

```bash
# Terminal 1: Start Redis
## Option A: Local Redis (Recommended for Development)
# Install Redis (Ubuntu/Debian)
sudo apt-get update && sudo apt-get install redis-server -y
sudo systemctl start redis-server
sudo systemctl enable redis-server
# Verify Redis is running
redis-cli ping  # Should return "PONG"
## Option B: Docker (if Docker is installed)
# docker run -d -p 6379:6379 --name redis redis:7-alpine
# Terminal 2: Start FastAPI Server
cd docgenie/api
python main.py
# Terminal 3: Start RQ Worker
cd docgenie/api
./start_worker.sh
```

### 2. Configure Environment

Make sure your `api/.env` file has:

```bash
# Required
ANTHROPIC_API_KEY=your_claude_api_key
SUPABASE_URL=https://your-project.supabase.co
SUPABASE_KEY=your_supabase_key
REDIS_URL=redis://localhost:6379/0
# Optional (for token refresh)
GOOGLE_CLIENT_ID=your_client_id.apps.googleusercontent.com
GOOGLE_CLIENT_SECRET=your_client_secret
```

### 3. Create Supabase Tables

Run the SQL from [DEPLOYMENT.md](DEPLOYMENT.md#32-create-database-schema) in your Supabase SQL Editor.

---

## Quick Start

### Option 1: Using Test Script (Easiest)

```bash
# Get Google Drive token first (one-time setup)
python api/test_get_google_token.py \
  --client-id YOUR_CLIENT_ID \
  --client-secret YOUR_CLIENT_SECRET
# Copy the access token, then run test
python api/test_async_api.py --google-token YOUR_ACCESS_TOKEN
```

### Option 2: Using OAuth Playground (Quick Test)

1. Go to [OAuth Playground](https://developers.google.com/oauthplayground/)
2. Configure with your credentials
3. Get access token
4. Run test script with token

---

## Getting Google Drive Token

### Method 1: Using Helper Script (Recommended)

Our helper script automates the OAuth flow:

```bash
cd docgenie/api
python test_get_google_token.py \
  --client-id YOUR_GOOGLE_CLIENT_ID \
  --client-secret YOUR_GOOGLE_CLIENT_SECRET
```

**What it does:**
1. Opens browser for Google authorization
2. Starts local server on port 8080 for callback
3. Exchanges authorization code for tokens
4. Displays access token and refresh token

**Output:**
```
Access Token: ya29.a0AfH6SMBx...
Refresh Token: 1//0gw...
```

### Method 2: OAuth Playground (No Code)

1. **Go to**: https://developers.google.com/oauthplayground/

2. **Configure Credentials**:
   - Click gear icon (⚙) in top right
   - Check "Use your own OAuth credentials"
   - Enter your Client ID and Client Secret

3. **Authorize API**:
   - In left panel, scroll to "Drive API v3"
   - Select: `https://www.googleapis.com/auth/drive.file`
   - Click "Authorize APIs"
   - Sign in with your Google account

4. **Get Token**:
   - Click "Exchange authorization code for tokens"
   - Copy the "Access token" value

### Method 3: Manual cURL (For Advanced Users)

**Step 1: Get Authorization Code**

Open this URL in browser (replace YOUR_CLIENT_ID):

```
https://accounts.google.com/o/oauth2/v2/auth?client_id=YOUR_CLIENT_ID&redirect_uri=http://localhost:8080&response_type=code&scope=https://www.googleapis.com/auth/drive.file&access_type=offline&prompt=consent
```

**Step 2: Exchange Code for Token**

After authorization, you'll be redirected to:
```
http://localhost:8080/?code=AUTHORIZATION_CODE
```

Exchange the code:

```bash
curl -X POST https://oauth2.googleapis.com/token \
  -d "code=AUTHORIZATION_CODE" \
  -d "client_id=YOUR_CLIENT_ID" \
  -d "client_secret=YOUR_CLIENT_SECRET" \
  -d "redirect_uri=http://localhost:8080" \
  -d "grant_type=authorization_code"
```

Response:
```json
{
  "access_token": "ya29.a0AfH6SMBx...",
  "refresh_token": "1//0gw...",
  "expires_in": 3600,
  "token_type": "Bearer"
}
```

---

## Testing Async API

The async API (`/generate/async`) is optimized for batch processing with 50% cost savings. Jobs are queued and processed in the background, with status polling.

### Full Automated Test

```bash
cd docgenie/api
# Set token as environment variable
export GOOGLE_DRIVE_TOKEN="ya29.a0AfH6SMBx..."
# Run test (generates 2 documents by default)
python test_async_api.py
# Or pass token directly
python test_async_api.py --google-token "ya29.a0AfH6SMBx..."
```

**Test Flow:**
1. ✓ Health check
2. ✓ Submit async job
3. ✓ Poll status (every 30 seconds)
4. ✓ List user jobs
5. ✓ Display Google Drive link

**Expected Output:**
```
================================================================================
                        ASYNC API TEST SUITE
================================================================================
Base URL: http://localhost:8000
User ID: 1
Documents to Generate: 2
================================================================================

============================================================
1. Testing API Health
============================================================
✓ API is healthy: {'status': 'healthy', 'version': '1.0.0'}

============================================================
2. Submitting Async Job
============================================================
Payload:
  User ID: 1
  Seed Images: 1
  Num Solutions: 2
  Google Token: ya29.a0AfH6SMBx...

✓ Job submitted successfully!
  Request ID: 550e8400-e29b-41d4-a716-446655440000
  Status: queued
  Estimated Time: 10 minutes
  Poll URL: /jobs/550e8400-e29b-41d4-a716-446655440000/status

============================================================
3. Polling Job Status
============================================================
Polling every 30 seconds (max 60 attempts)
Status flow: queued → processing → generating → completed/failed

[12:00:00] Poll 1/60: QUEUED
[12:00:30] Poll 2/60: PROCESSING - Creating batch request...
[12:01:00] Poll 3/60: GENERATING - Batch submitted to Claude...
[12:08:30] Poll 17/60: GENERATING - Polling batch status...
[12:15:00] Poll 30/60: COMPLETED

============================================================
✓ JOB COMPLETED!
============================================================
  Download URL: https://drive.google.com/file/d/abc123xyz/view?usp=sharing
  File Size: 15.4 MB
  Document Count: 2
  Created: 2026-02-28T12:00:00Z
  Completed: 2026-02-28T12:15:00Z

================================================================================
                           TEST SUMMARY
================================================================================
✓ ALL TESTS PASSED!

Your documents are available at:
  https://drive.google.com/file/d/abc123xyz/view?usp=sharing

Next steps:
  1. Open the Google Drive link in your browser
  2. Download the ZIP file
  3. Extract and verify generated documents
```
### Test Options
```bash
# Custom number of documents
python test_async_api.py --google-token TOKEN --num-solutions 5

# Custom API URL (if deployed)
python test_async_api.py --google-token TOKEN --base-url https://api.yourdomain.com

# Different user ID
python test_async_api.py --google-token TOKEN --user-id 42

# With refresh token
python test_async_api.py \
  --google-token ACCESS_TOKEN \
  --google-refresh-token REFRESH_TOKEN

# Show help for getting token
python test_async_api.py --help-token
```
---Testing Sync PDF API
The sync PDF API (`/generate/pdf`) returns results immediately (20-60s) and supports three modes of operation. Perfect for smaller batch sizes and real-time workflows.
### Three Operating Modes
**Mode 1: Quick Demo (No Tracking)**
- Returns ZIP immediately
- No Supabase records created
- Perfect for quick testing and demos
- No user_id required
**Mode 2: Demo with Tracking**
- Returns ZIP immediately
- Creates Supabase record for tracking
- Can poll status during generation
- Requires user_id
**Mode 3: Full Production**
- Returns ZIP immediately
- Creates Supabase record
- Uploads to Google Drive in background
- Requires user_id + google_drive_token
- Best for production use
### Full Automated Test
```bash
cd docgenie/api

# Mode 1: Quick demo (no tracking)
python test_sync_pdf_api.py
# Mode 2: Demo with tracking
python test_sync_pdf_api.py --user-id 123

# Mode 3: Full production (tracking + GDrive)
python test_sync_pdf_api.py \
  --user-id 123 \
  --google-token "ya29.a0AfH6SMBx..." \
  --google-refresh-token "1//0gw..."
```
**Test Flow for All Modes:**
1. ✓ Health check
2. ✓ Test Mode 1: Quick demo (always runs)
3. ✓ Test Mode 2: With tracking (if user_id provided)
4. ✓ Test Mode 3: Full production (if user_id + token provided)
5. ✓ Validate ZIP contents
6. ✓ Test status polling (Modes 2 & 3)
7. ✓ Verify GDrive upload (Mode 3)
**Expected Output:**
```
================================================================================
DocGenie /generate/pdf Endpoint Test Suite
================================================================================
================================================================================
1. Testing API Health
================================================================================
✓ API is healthy: {'status': 'healthy', 'version': '1.0.0'}
================================================================================
2. Testing Mode 1: Quick Demo (No Tracking)
================================================================================
This mode returns ZIP immediately without creating Supabase records.
Use for quick testing and demos.
Payload:
  Seed Images: 1
  Num Solutions: 1
  User ID: None (no tracking)
  Google Token: None
⏳ Calling /generate/pdf (expect 20-60 seconds)...
✓ Response received in 42.3 seconds
Response Headers:
  Content-Type: application/zip
  Content-Disposition: attachment; filename=docgenie_documents.zip
  X-Request-ID: NOT SET (expected in mode 1)
  X-Status-URL: NOT SET (expected in mode 1)

✓ ZIP file size: 145.2 KB
✓ ZIP contains 18 files:
    - README.md
    - metadata.json
    - analysis/document_1.json
    - annotations/gt/document_1.json
    - bbox/bbox_pdf/word/document_1.json
    - html/document_1.css
    - html/document_1.html
    - img/document_1.png
    - pdf/pdf_final/document_1.pdf
    - pdf/pdf_initial/document_1.pdf
  ✓ Contains metadata.json
  ✓ Contains README.md
✅ Mode 1 (Quick Demo) Test PASSED
   ⚡ Fast response: 42.3s
   📦 Valid ZIP file
   ✓ No tracking overhead
================================================================================
3. Testing Mode 2: Demo with Progress Tracking
================================================================================
This mode returns ZIP immediately AND creates Supabase record.
Client can poll /jobs/{request_id}/status during generation.

Payload:
  User ID: 123 (tracking enabled)
  Seed Images: 1
  Num Solutions: 2
  Google Token: None

⏳ Calling /generate/pdf (expect 20-60 seconds)...

✓ Response received in 58.7 seconds

Response Headers:
  Content-Type: application/zip
  ✓ X-Request-ID: 550e8400-e29b-41d4-a716-446655440000
  ✓ X-Status-URL: /jobs/550e8400-e29b-41d4-a716-446655440000/status

✓ ZIP file size: 287.4 KB
✓ ZIP contains 32 files
✓ Found 4 PDF files

⏳ Testing status polling endpoint...
✓ Status endpoint working:
  Request ID: 550e8400-e29b-41d4-a716-446655440000
  Status: completed
  Created: 2026-03-01T10:15:00Z
  Updated: 2026-03-01T10:15:58Z
  ✓ Job marked as completed

✅ Mode 2 (Tracking) Test PASSED
   ⚡ Fast response: 58.7s
   📦 Valid ZIP file
   📊 Progress tracking enabled
   ✓ Can poll status during generation

================================================================================
4. Testing Mode 3: Full Production (Tracking + GDrive Upload)
================================================================================
This mode returns ZIP immediately AND uploads to Google Drive in background.
Best for production use with full tracking and backup.

Payload:
  User ID: 123
  Google Token: ya29.a0AfH6SMBx...
  Google Refresh: Yes
  Seed Images: 1
  Num Solutions: 1

⏳ Calling /generate/pdf (expect 20-60 seconds)...

✓ Response received in 45.1 seconds

Response Headers:
  ✓ X-Request-ID: 660f9511-f3ac-52e5-b827-557766551111
  ✓ X-Status-URL: /jobs/660f9511-f3ac-52e5-b827-557766551111/status

✓ ZIP file size: 151.8 KB
✓ ZIP contains 18 files

⏳ ZIP returned immediately, GDrive upload happening in background...
   (This doesn't block the response)

⏳ Waiting 10 seconds for background GDrive upload...
✓ Status after background upload:
  Status: completed
  ✓ GDrive URL: https://drive.google.com/file/d/abc123xyz/view?usp=...
  ✓ Background upload completed!

✅ Mode 3 (Full Production) Test PASSED
   ⚡ Fast response: 45.1s (GDrive doesn't block)
   📦 Valid ZIP file delivered immediately
   📊 Progress tracking enabled
   ☁️  Google Drive backup scheduled
   ✓ Production-ready configuration

================================================================================
TEST SUMMARY
================================================================================
  ✅ health: PASSED
  ✅ mode_1: PASSED
  ✅ mode_2: PASSED
  ✅ mode_3: PASSED
4/4 tests passed
🎉 All tests passed!
================================================================================
```
### Test Options
```bash
# Mode 1 only (default)
python test_sync_pdf_api.py

# Mode 2 with custom user ID
python test_sync_pdf_api.py --user-id 456
# Mode 3 with custom API URL
python test_sync_pdf_api.py \
  --base-url https://api.yourdomain.com \
  --user-id 123 \
  --google-token TOKEN \
  --google-refresh-token REFRESH_TOKEN
```
### Comparing Sync vs Async
| Feature | Sync (`/generate/pdf`) | Async (`/generate/async`) |
|---------|------------------------|---------------------------|
| **Response Time** | 20-60 seconds | 5-30 minutes |
| **Best For** | 1-3 documents | 5-50+ documents |
| **Cost** | Standard API pricing | 50% cheaper (Batch API) |
| **Result Delivery** | Direct ZIP download | Google Drive upload |
| **Progress Tracking** | Optional (Modes 2 & 3) | Always enabled |
| **Use Case** | Real-time workflows, demos | Bulk generation, scheduled jobs |
**When to use Sync:**
- Generating 1-3 documents
- Need immediate results
- Real-time user interactions
- Quick testing and demos
**When to use Async:**
- Generating 5+ documents
- Cost optimization (50% savings)
- Background/scheduled processing
- Large batch jobs
---
## Manual Testing with cURL
### Async API (`/generate/async`)
#### 1. Submit Async Job
```bash
curl -X POST http://localhost:8000/generate/async \
  -H "Content-Type: application/json" \
  -d '{
    "user_id": 1,
    "google_drive_token": "ya29.a0AfH6SMBx...",
    "seed_images": ["https://ocr.space/Content/Images/receipt-ocr-original.webp"],
    "prompt_params": {
      "language": "English",
      "doc_type": "receipts",
      "num_solutions": 2,
      "enable_handwriting": false,
      "enable_visual_elements": false,
      "output_detail": "minimal"
    }
  }'
```
**Response:**
```json
{
  "request_id": "550e8400-e29b-41d4-a716-446655440000",
  "status": "queued",
  "estimated_time_minutes": 10,
  "poll_url": "/jobs/550e8400-e29b-41d4-a716-446655440000/status",
  "created_at": "2026-02-28T12:00:00Z"
}
```

#### 2. Check Job Status

```bash
curl http://localhost:8000/jobs/550e8400-e29b-41d4-a716-446655440000/status
```

**Response (Processing):**
```json
{
  "request_id": "550e8400-e29b-41d4-a716-446655440000",
  "status": "processing",
  "created_at": "2026-02-28T12:00:00Z",
  "updated_at": "2026-02-28T12:02:00Z",
  "progress": "Creating batch request..."
}
```

**Response (Completed):**
```json
{
  "request_id": "550e8400-e29b-41d4-a716-446655440000",
  "status": "completed",
  "created_at": "2026-02-28T12:00:00Z",
  "updated_at": "2026-02-28T12:15:00Z",
  "download_url": "https://drive.google.com/file/d/abc123xyz/view?usp=sharing",
  "file_size_mb": 15.4,
  "document_count": 2
}
```

#### 3. List User Jobs

```bash
curl "http://localhost:8000/jobs/user/1?limit=10&offset=0"
```

**Response:**
```json
{
  "user_id": 1,
  "jobs": [
    {
      "request_id": "550e8400-e29b-41d4-a716-446655440000",
      "status": "completed",
      "created_at": "2026-02-28T12:00:00Z",
      "download_url": "https://drive.google.com/file/d/abc123xyz/view"
    }
  ],
  "count": 1,
  "limit": 10,
  "offset": 0
}
```

### Sync PDF API (`/generate/pdf`)

#### Mode 1: Quick Demo (No Tracking)

```bash
curl -X POST http://localhost:8000/generate/pdf \
  -H "Content-Type: application/json" \
  -d '{
    "seed_images": ["https://ocr.space/Content/Images/receipt-ocr-original.webp"],
    "prompt_params": {
      "language": "English",
      "doc_type": "receipts",
      "num_solutions": 1,
      "enable_handwriting": false,
      "enable_visual_elements": false,
      "output_detail": "minimal"
    }
  }' \
  --output documents.zip
```

**Response:**
- Returns ZIP file directly (binary)
- No tracking headers
- File saved as `documents.zip`

#### Mode 2: Demo with Tracking

```bash
curl -X POST http://localhost:8000/generate/pdf \
  -H "Content-Type: application/json" \
  -d '{
    "user_id": 123,
    "seed_images": ["https://ocr.space/Content/Images/receipt-ocr-original.webp"],
    "prompt_params": {
      "language": "English",
      "doc_type": "business documents",
      "num_solutions": 2,
      "enable_handwriting": false,
      "output_detail": "minimal"
    }
  }' \
  --output documents.zip \
  -D headers.txt
```

**Response:**
- Returns ZIP file directly (binary)
- Headers saved to `headers.txt` contain:
  - `X-Request-ID: 550e8400-e29b-41d4-a716-446655440000`
  - `X-Status-URL: /jobs/550e8400-e29b-41d4-a716-446655440000/status`

**Check Status:**
```bash
# Extract request_id from headers.txt, then:
curl http://localhost:8000/jobs/550e8400-e29b-41d4-a716-446655440000/status
```

#### Mode 3: Full Production (Tracking + GDrive)

```bash
curl -X POST http://localhost:8000/generate/pdf \
  -H "Content-Type: application/json" \
  -d '{
    "user_id": 123,
    "google_drive_token": "ya29.a0AfH6SMBx...",
    "google_drive_refresh_token": "1//0gw...",
    "seed_images": ["https://ocr.space/Content/Images/receipt-ocr-original.webp"],
    "prompt_params": {
      "language": "English",
      "doc_type": "invoices",
      "num_solutions": 1,
      "enable_handwriting": false,
      "output_detail": "dataset"
    }
  }' \
  --output documents.zip \
  -D headers.txt
```

**Response:**
- Returns ZIP file immediately (binary)
- Google Drive upload happens in background
- Wait 10-30 seconds, then check status for GDrive URL:

```bash
curl http://localhost:8000/jobs/550e8400-e29b-41d4-a716-446655440000/status
```

**Response (after background upload):**
```json
{
  "request_id": "550e8400-e29b-41d4-a716-446655440000",
  "status": "completed",
  "created_at": "2026-03-01T10:00:00Z",
  "updated_at": "2026-03-01T10:00:45Z",
  "results": {
    "download_url": "https://drive.google.com/file/d/abc123xyz/view?usp=sharing",
    "file_size_mb": 0.15,
    "document_count": 1,
    "zip_filename": "docgenie_550e8400-e29b-41d4-a716-446655440000.zip"
  }
}
```

---

## Frontend Integration Example

### React + TypeScript

```typescript
import { useState, useEffect } from 'react';
interface JobStatus {
  request_id: string;
  status: 'queued' | 'processing' | 'generating' | 'completed' | 'failed';
  download_url?: string;
  error_message?: string;
}
function DocumentGenerator() {
  const [jobId, setJobId] = useState<string | null>(null);
  const [status, setStatus] = useState<JobStatus | null>(null);
  const [googleToken, setGoogleToken] = useState<string>('');
  // Step 1: Google OAuth (implement separately)
  const handleGoogleAuth = async () => {
    // Redirect to Google OAuth
    const clientId = 'YOUR_CLIENT_ID';
    const redirectUri = 'https://yourapp.com/auth/callback';
    const scope = 'https://www.googleapis.com/auth/drive.file';
    
    const authUrl = `https://accounts.google.com/o/oauth2/v2/auth?` +
      `client_id=${clientId}&` +
      `redirect_uri=${redirectUri}&` +
      `response_type=code&` +
      `scope=${scope}&` +
      `access_type=offline&` +
      `prompt=consent`;
    
    window.location.href = authUrl;
  };
  // Step 2: Submit job
  const handleGenerateDocuments = async () => {
    const response = await fetch('http://localhost:8000/generate/async', {
      method: 'POST',
      headers: { 'Content-Type': 'application/json' },
      body: JSON.stringify({
        user_id: 1,
        google_drive_token: googleToken,
        seed_images: ['https://example.com/seed.jpg'],
        prompt_params: {
          language: 'English',
          doc_type: 'receipts',
          num_solutions: 3
        }
      })
    });
    const job = await response.json();
    setJobId(job.request_id);
  };
  // Step 3: Poll status
  useEffect(() => {
    if (!jobId) return;
    const interval = setInterval(async () => {
      const response = await fetch(
        `http://localhost:8000/jobs/${jobId}/status`
      );
      const data = await response.json();
      setStatus(data);
      if (data.status === 'completed' || data.status === 'failed') {
        clearInterval(interval);
      }
    }, 30000); // Poll every 30 seconds
    return () => clearInterval(interval);
  }, [jobId]);
  return (
    <div>
      {!googleToken ? (
        <button onClick={handleGoogleAuth}>
          Connect Google Drive
        </button>
      ) : (
        <button onClick={handleGenerateDocuments}>
          Generate Documents
        </button>
      )}
      {status && (
        <div>
          <p>Status: {status.status}</p>
          {status.status === 'completed' && (
            <a href={status.download_url} target="_blank">
              Download Documents
            </a>
          )}
          {status.status === 'failed' && (
            <p>Error: {status.error_message}</p>
          )}
        </div>
      )}
    </div>
  );
}
```

---

## Troubleshooting

### Issue: "google_drive_token is required"

**Cause**: No token provided in request

**Solution**:
```bash
# Make sure you're passing the token
python test_async_api.py --google-token "ya29.a0AfH6SMBx..."
```

### Issue: "Failed to refresh Google Drive token"

**Cause**: Token expired and no refresh token provided

**Solutions**:
1. Get a new token (tokens expire in ~1 hour)
2. Include refresh token in request
3. Frontend should refresh tokens automatically

### Issue: "Google Drive upload failed: insufficient permissions"

**Cause**: Token doesn't have drive.file scope

**Solution**: Re-authorize with correct scope:
```
https://www.googleapis.com/auth/drive.file
```

### Issue: Worker not processing jobs

**Check 1**: Is Redis running?
```bash
redis-cli ping  # Should return "PONG"
```

**Check 2**: Is worker running?
```bash
# Check worker logs
journalctl -u docgenie-worker@1 -f
# Or check RQ info
rq info --url redis://localhost:6379/0
```

**Check 3**: Check failed queue
```bash
rq info --queue failed --url redis://localhost:6379/0
```

### Issue: Job stuck in "generating" status

**Cause**: Batch API taking longer than expected

**Solution**: Wait up to 30 minutes for batched requests. Check Anthropic dashboard:
https://console.anthropic.com/settings/batches

### Issue: Cannot access Google Drive link

**Cause**: File not shared properly

**Solution**: Check worker logs for sharing errors. File should have "anyone with link" permission.

---

## Performance Testing

### Test Batch API Cost Savings

```bash
# Generate 10 documents
time python test_async_api.py --google-token TOKEN --num-solutions 10
# Compare with direct API (for reference)
curl -X POST http://localhost:8000/generate \
  -H "Content-Type: application/json" \
  -d '{"seed_images": ["..."], "prompt_params": {"num_solutions": 10}}'
```

**Expected Results:**
- **Batched API**: 10-20 minutes, ~$2.50 per 1M tokens
- **Direct API**: 3-5 minutes, ~$5.00 per 1M tokens
- **Cost Savings**: 50%

---

## Next Steps

1. ✅ Test locally with script
2. ✅ Verify Google Drive upload
3. ✅ Test with your frontend
4. ✅ Deploy to production (see [DEPLOYMENT.md](DEPLOYMENT.md))
5. ✅ Set up monitoring (see [SCALING.md](SCALING.md))

---

## Additional Resources

- **API Documentation**: http://localhost:8000/docs
- **Deployment Guide**: [DEPLOYMENT.md](DEPLOYMENT.md)
- **Scaling Guide**: [SCALING.md](SCALING.md)
- **Google OAuth Docs**: https://developers.google.com/identity/protocols/oauth2
- **Anthropic Batch API**: https://docs.anthropic.com/en/docs/batch-api