# Jupyter Notebook Usage

This document shows how to use the document processing function in Jupyter notebooks for integration into larger processing pipelines.

## Simple Usage

```python
from processing.document_processor import process_document_with_redaction

# Process a single document
result = process_document_with_redaction(
    file_path="path/to/your/document.pdf",
    endpoint="your-azure-openai-endpoint",
    api_key="your-azure-openai-key",
    api_version="2024-02-15-preview",
    deployment="o3-mini"  # or "o4-mini", "o3", "o4"
)

# Access the results
original_md = result.original_document_md
redacted_md = result.redacted_document_md
input_tokens = result.input_tokens
output_tokens = result.output_tokens
cost = result.cost

print(f"Processing complete!")
print(f"Input tokens: {input_tokens:,}")
print(f"Output tokens: {output_tokens:,}")
print(f"Total cost: ${cost:.4f}")
```

## Batch Processing

```python
import os
from processing.document_processor import process_document_with_redaction

# Configuration
AZURE_OPENAI_ENDPOINT = "your-azure-openai-endpoint"
AZURE_OPENAI_KEY = "your-azure-openai-key"
AZURE_OPENAI_VERSION = "2024-02-15-preview"
AZURE_OPENAI_DEPLOYMENT = "o3-mini"

# Process multiple documents
pdf_directory = "path/to/pdf/files"
results = []

for filename in os.listdir(pdf_directory):
    if filename.endswith('.pdf'):
        file_path = os.path.join(pdf_directory, filename)
        
        print(f"Processing {filename}...")
        
        try:
            result = process_document_with_redaction(
                file_path=file_path,
                endpoint=AZURE_OPENAI_ENDPOINT,
                api_key=AZURE_OPENAI_KEY,
                api_version=AZURE_OPENAI_VERSION,
                deployment=AZURE_OPENAI_DEPLOYMENT
            )
            
            results.append({
                'filename': filename,
                'original_md': result.original_document_md,
                'redacted_md': result.redacted_document_md,
                'input_tokens': result.input_tokens,
                'output_tokens': result.output_tokens,
                'cost': result.cost
            })
            
            print(f"  ✓ Completed - Cost: ${result.cost:.4f}")
            
        except Exception as e:
            print(f"  ✗ Error processing {filename}: {e}")

# Summary
total_cost = sum(r['cost'] for r in results)
total_input_tokens = sum(r['input_tokens'] for r in results)
total_output_tokens = sum(r['output_tokens'] for r in results)

print(f"\nBatch processing complete!")
print(f"Documents processed: {len(results)}")
print(f"Total input tokens: {total_input_tokens:,}")
print(f"Total output tokens: {total_output_tokens:,}")
print(f"Total cost: ${total_cost:.4f}")
```

## Environment Variables

You can also use environment variables for configuration:

```python
import os
from dotenv import load_dotenv
from processing.document_processor import process_document_with_redaction

# Load environment variables
load_dotenv()

# Get configuration from environment
AZURE_OPENAI_ENDPOINT = os.getenv("AZURE_OPENAI_ENDPOINT")
AZURE_OPENAI_KEY = os.getenv("AZURE_OPENAI_KEY")
AZURE_OPENAI_VERSION = os.getenv("AZURE_OPENAI_VERSION")
AZURE_OPENAI_DEPLOYMENT = os.getenv("AZURE_OPENAI_DEPLOYMENT")

# Process document
result = process_document_with_redaction(
    file_path="document.pdf",
    endpoint=AZURE_OPENAI_ENDPOINT,
    api_key=AZURE_OPENAI_KEY,
    api_version=AZURE_OPENAI_VERSION,
    deployment=AZURE_OPENAI_DEPLOYMENT
)
```

## Return Value

The function returns a `ProcessingResult` object with the following attributes:

- `original_document_md`: Markdown version of the original document
- `redacted_document_md`: Markdown version with medication sections removed
- `input_tokens`: Number of input tokens used
- `output_tokens`: Number of output tokens generated
- `cost`: Total cost in USD

## Supported Models

The function supports the following Azure OpenAI deployment names:
- `o3-mini` (GPT-4o Mini) - Cheapest option
- `o4-mini` (GPT-4o Mini) - Same as o3-mini
- `o3` (GPT-3.5 Turbo) - Medium cost
- `o4` (GPT-4o) - Most expensive but most capable

## Error Handling

The function will raise exceptions for:
- File not found
- Invalid Azure OpenAI credentials
- API rate limits
- Network errors

Make sure to handle these appropriately in your pipeline.