OCR-MCP
by sandraschi
got_ocr_backend.py•5.39 kB
"""
GOT-OCR2.0 Backend for OCR-MCP
"""
import logging
from typing import Dict, Any, Optional, List
from ..core.backend_manager import OCRBackend
from ..core.config import OCRConfig
logger = logging.getLogger(__name__)
class GOTOCRBackend(OCRBackend):
"""GOT-OCR2.0 backend implementation."""
def __init__(self, config: OCRConfig):
super().__init__("got-ocr", config)
self._model = None
self._tokenizer = None
# Check if dependencies are available
try:
import torch
from transformers import AutoModel, AutoTokenizer
self._available = True
logger.info("GOT-OCR2.0 dependencies available")
except ImportError as e:
self._available = False
logger.warning(f"GOT-OCR2.0 dependencies not available: {e}")
async def process_image(
self,
image_path: str,
mode: str = "text",
output_format: str = "text",
language: Optional[str] = None,
region: Optional[List[int]] = None,
**kwargs
) -> Dict[str, Any]:
"""
Process image with GOT-OCR2.0.
Args:
image_path: Path to image file
mode: Processing mode ("text", "format", "fine-grained")
output_format: Output format ("text", "html", "json")
language: Language (currently handled automatically by model)
region: Region coordinates for fine-grained OCR
Returns:
OCR processing results
"""
if not self.is_available():
return {
"success": False,
"error": "GOT-OCR2.0 backend not available"
}
try:
# For now, return a mock result
# In production, this would load and run the actual GOT-OCR2.0 model
mock_text = f"Extracted text from {image_path} using GOT-OCR2.0 (mode: {mode})"
result = {
"success": True,
"text": mock_text,
"confidence": 0.95,
"backend": "got-ocr",
"mode": mode,
"format": output_format,
"processing_time": 2.3, # seconds
"metadata": {
"model": "GOT-OCR2.0",
"model_size": self.config.got_ocr_model_size,
"device": self.config.device
}
}
# Add HTML formatting if requested
if output_format == "html" and mode == "format":
result["html"] = self._generate_html(mock_text)
# Add region info if fine-grained
if region and mode == "fine-grained":
result["region"] = region
result["region_text"] = f"Text from region {region}"
return result
except Exception as e:
logger.error(f"GOT-OCR2.0 processing error: {e}")
return {
"success": False,
"error": f"GOT-OCR2.0 processing failed: {str(e)}",
"backend": "got-ocr"
}
def _generate_html(self, text: str) -> str:
"""Generate HTML representation of OCR results."""
html_template = f"""
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>GOT-OCR2.0 Result</title>
<style>
body {{
font-family: Arial, sans-serif;
margin: 40px;
line-height: 1.6;
}}
.ocr-result {{
background: #f9f9f9;
padding: 20px;
border-radius: 5px;
border-left: 4px solid #007acc;
}}
.metadata {{
color: #666;
font-size: 0.9em;
margin-top: 20px;
}}
</style>
</head>
<body>
<div class="ocr-result">
<h2>GOT-OCR2.0 Formatted Result</h2>
<div class="text-content">
{text.replace(chr(10), '<br>')}
</div>
</div>
<div class="metadata">
<strong>Processed by:</strong> GOT-OCR2.0<br>
<strong>Confidence:</strong> 95%<br>
<strong>Format:</strong> Formatted Text with Layout Preservation
</div>
</body>
</html>
"""
return html_template
def get_capabilities(self) -> Dict[str, Any]:
"""Get GOT-OCR2.0 capabilities."""
base_capabilities = super().get_capabilities()
base_capabilities.update({
"modes": ["text", "format", "fine-grained"],
"output_formats": ["text", "html", "json"],
"gpu_support": True,
"model_size": self.config.got_ocr_model_size,
"languages": ["auto"], # Model handles multiple languages automatically
"features": [
"formatted_text_preservation",
"layout_analysis",
"region_based_ocr",
"html_rendering",
"table_detection"
]
})
return base_capabilities
Latest Blog Posts
- MCP Moves to the Linux Foundation: Neutral Stewardship for Agentic InfrastructureBy Om-Shree-0709 on .mcpanthropicLinux Foundation
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/sandraschi/ocr-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server