# API Reference
## MCP Tools
### parse_document
Parse and convert any document to Markdown format.
**Input Schema:**
```json
{
"source": "string (required)",
"pipeline": "string (optional: standard|vlm|asr|auto)",
"options": {
"ocr_enabled": "boolean",
"ocr_language": "string",
"table_accuracy_mode": "string (fast|accurate)",
"pdf_backend": "string (dlparse_v4|pypdfium2)",
"enable_enrichments": "boolean"
}
}
```
**Returns:** Markdown text
**Example:**
```json
{
"name": "parse_document",
"arguments": {
"source": "/path/to/document.pdf",
"pipeline": "standard"
}
}
```
### parse_document_advanced
Advanced document parsing with detailed configuration options.
**Input Schema:**
```json
{
"source": "string (required)",
"pipeline": "string (optional)",
"ocr_enabled": "boolean",
"ocr_language": "string",
"table_accuracy_mode": "string",
"pdf_backend": "string",
"enable_enrichments": "boolean"
}
```
**Returns:** Markdown text
### get_job_status
Get the status of a processing job.
**Input Schema:**
```json
{
"job_id": "string (required)"
}
```
**Returns:** Job status JSON
**Response Format:**
```json
{
"job_id": "string",
"source_path": "string",
"pipeline": "string",
"status": "string",
"created_at": "ISO8601 timestamp",
"started_at": "ISO8601 timestamp",
"completed_at": "ISO8601 timestamp",
"duration_seconds": "number",
"retry_count": "number",
"error_message": "string"
}
```
### list_supported_formats
List all supported input formats and processing pipelines.
**Input Schema:**
```json
{}
```
**Returns:** Supported formats JSON
**Response Format:**
```json
{
"input_formats": ["pdf", "docx", ...],
"output_formats": ["markdown", "html", "json", "text", "doctags"],
"pipelines": ["standard", "vlm", "asr"]
}
```
### get_queue_statistics
Get current queue and processing statistics.
**Input Schema:**
```json
{}
```
**Returns:** Statistics JSON
**Response Format:**
```json
{
"queue": {
"current_size": "number",
"max_size": "number",
"is_full": "boolean",
"is_empty": "boolean"
},
"processing": {
"total_jobs": "number",
"active_jobs": "number",
"status_counts": {
"pending": "number",
"running": "number",
"completed": "number",
"failed": "number"
},
"average_duration_seconds": "number"
}
}
```
## Python API
### DocumentProcessor
Main document processing engine.
```python
from document_parser.engine.processor import DocumentProcessor
from document_parser.config.settings import get_settings
# Initialize
settings = get_settings()
processor = DocumentProcessor(settings)
# Process document
result = await processor.process_document(
source="document.pdf",
pipeline="standard",
options={"ocr_enabled": True}
)
```
#### Methods
**`async process_document(source, pipeline=None, options=None)`**
Process a document and return Markdown.
- **source** (str): File path or URL
- **pipeline** (str, optional): Processing pipeline
- **options** (dict, optional): Processing options
- **Returns**: str - Markdown content
**`get_supported_formats()`**
Get list of supported formats.
- **Returns**: dict - Supported formats
### DocumentParserServer
MCP server implementation.
```python
from document_parser.mcp.server import DocumentParserServer
from document_parser.config.settings import load_settings
# Initialize
settings = load_settings("config.yaml")
server = DocumentParserServer(settings)
# Run server
await server.run()
```
## Configuration API
### ApplicationSettings
Main configuration model.
```python
from document_parser.config.models import ApplicationSettings
settings = ApplicationSettings(
server=ServerSettings(max_concurrent_jobs=5),
processing=ProcessingSettings(default_pipeline="standard")
)
```
### Loading Configuration
```python
from document_parser.config.settings import load_settings, get_settings
# Load from file
settings = load_settings("config.yaml")
# Get singleton
settings = get_settings()
```
## Error Handling
### Exception Hierarchy
```
DocumentParserError (base)
├── ProcessingError
├── ConfigurationError
├── NetworkError
└── ValidationError
```
### Example
```python
from document_parser.core.exceptions import ProcessingError
try:
result = await processor.process_document("document.pdf")
except ProcessingError as e:
print(f"Processing failed: {e.message}")
if e.details:
print(f"Details: {e.details}")
```
