File size: 41,219 Bytes
a963d65 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 |
"""
Local Processor for FhirFlame Development
Core logic with optional Mistral API OCR and multimodal fallbacks
"""
import asyncio
import json
import uuid
import os
import io
import base64
from datetime import datetime
from typing import Dict, Any, Optional, List
from .monitoring import monitor
# PDF and Image Processing
try:
from pdf2image import convert_from_bytes
from PIL import Image
import PyPDF2
PDF_PROCESSING_AVAILABLE = True
except ImportError:
PDF_PROCESSING_AVAILABLE = False
class LocalProcessor:
"""Local processor with optional external fallbacks"""
def __init__(self):
self.use_mistral_fallback = os.getenv("USE_MISTRAL_FALLBACK", "false").lower() == "true"
self.use_multimodal_fallback = os.getenv("USE_MULTIMODAL_FALLBACK", "false").lower() == "true"
self.mistral_api_key = os.getenv("MISTRAL_API_KEY")
@monitor.track_operation("real_document_processing")
async def process_document(self, document_bytes: bytes, user_id: str, filename: str) -> Dict[str, Any]:
"""Process document with fallback capabilities and quality assertions"""
# Try external OCR if enabled and available
extracted_text = await self._extract_text_with_fallback(document_bytes, filename)
# Log OCR quality metrics
monitor.log_event("ocr_text_extracted", {
"text_extracted": len(extracted_text) > 0,
"text_length": len(extracted_text),
"filename": filename
})
monitor.log_event("ocr_minimum_length", {
"substantial_text": len(extracted_text) > 50,
"text_length": len(extracted_text)
})
# Extract medical entities from text
entities = self._extract_medical_entities(extracted_text)
# Log medical entity extraction
monitor.log_event("medical_entities_found", {
"entities_found": len(entities) > 0,
"entity_count": len(entities)
})
# Create FHIR bundle
fhir_bundle = self._create_simple_fhir_bundle(entities, user_id)
# Log FHIR validation
monitor.log_event("fhir_bundle_valid", {
"bundle_valid": fhir_bundle.get("resourceType") == "Bundle",
"resource_type": fhir_bundle.get("resourceType")
})
monitor.log_event("fhir_has_entries", {
"has_entries": len(fhir_bundle.get("entry", [])) > 0,
"entry_count": len(fhir_bundle.get("entry", []))
})
# Log processing with enhanced metrics
monitor.log_medical_processing(
entities_found=len(entities),
confidence=0.85,
processing_time=100.0,
processing_mode="file_processing",
model_used="enhanced_processor"
)
return {
"status": "success",
"processing_mode": self._get_processing_mode(),
"filename": filename,
"processed_by": user_id,
"entities_found": len(entities),
"fhir_bundle": fhir_bundle,
"extracted_text": extracted_text[:500] + "..." if len(extracted_text) > 500 else extracted_text,
"text_length": len(extracted_text)
}
async def _extract_text_with_fallback(self, document_bytes: bytes, filename: str) -> str:
"""Extract text with optional fallbacks"""
# Try Mistral API OCR first if enabled
if self.use_mistral_fallback and self.mistral_api_key:
try:
monitor.log_event("mistral_attempt_start", {
"document_size": len(document_bytes),
"api_key_present": bool(self.mistral_api_key),
"use_mistral_fallback": self.use_mistral_fallback
})
result = await self._extract_with_mistral(document_bytes)
monitor.log_event("mistral_success_in_fallback", {
"text_length": len(result),
"text_preview": result[:100] + "..." if len(result) > 100 else result
})
return result
except Exception as e:
import traceback
monitor.log_event("mistral_fallback_failed", {
"error": str(e),
"error_type": type(e).__name__,
"traceback": traceback.format_exc(),
"document_size": len(document_bytes),
"api_key_format": f"{self.mistral_api_key[:8]}...{self.mistral_api_key[-4:]}" if self.mistral_api_key else "none"
})
print(f"🚨 MISTRAL API FAILED: {type(e).__name__}: {str(e)}")
print(f"🚨 Full traceback: {traceback.format_exc()}")
# Try multimodal processor if enabled
if self.use_multimodal_fallback:
try:
return await self._extract_with_multimodal(document_bytes)
except Exception as e:
monitor.log_event("multimodal_fallback_failed", {"error": str(e)})
# CRITICAL: No dummy data in production - fail properly when OCR fails
raise Exception(f"Document text extraction failed for {filename}. All OCR methods exhausted. Cannot return dummy data for real medical processing.")
def _convert_pdf_to_images(self, pdf_bytes: bytes) -> List[bytes]:
"""Convert PDF to list of image bytes for Mistral vision processing"""
if not PDF_PROCESSING_AVAILABLE:
raise Exception("PDF processing libraries not available. Install pdf2image, Pillow, and PyPDF2.")
try:
# Convert PDF pages to PIL Images
monitor.log_event("pdf_conversion_debug", {
"step": "starting_pdf_conversion",
"pdf_size": len(pdf_bytes)
})
# Convert PDF to images (300 DPI for good OCR quality)
images = convert_from_bytes(pdf_bytes, dpi=300, fmt='PNG')
monitor.log_event("pdf_conversion_debug", {
"step": "pdf_converted_to_images",
"page_count": len(images),
"image_sizes": [(img.width, img.height) for img in images]
})
# Convert PIL Images to bytes
image_bytes_list = []
for i, img in enumerate(images):
# Convert to RGB if necessary (for JPEG compatibility)
if img.mode != 'RGB':
img = img.convert('RGB')
# Save as high-quality JPEG bytes
img_byte_arr = io.BytesIO()
img.save(img_byte_arr, format='JPEG', quality=95)
img_bytes = img_byte_arr.getvalue()
image_bytes_list.append(img_bytes)
monitor.log_event("pdf_conversion_debug", {
"step": f"page_{i+1}_converted",
"page_size": len(img_bytes),
"dimensions": f"{img.width}x{img.height}"
})
monitor.log_event("pdf_conversion_success", {
"total_pages": len(image_bytes_list),
"total_size": sum(len(img_bytes) for img_bytes in image_bytes_list)
})
return image_bytes_list
except Exception as e:
monitor.log_event("pdf_conversion_error", {
"error": str(e),
"error_type": type(e).__name__
})
raise Exception(f"PDF to image conversion failed: {str(e)}")
async def _extract_with_mistral(self, document_bytes: bytes) -> str:
"""Extract text using Mistral OCR API - using proper document understanding endpoint"""
import httpx
import base64
import tempfile
import os
# 🔍 DEBUGGING: Log entry to Mistral OCR function
monitor.log_event("mistral_ocr_start", {
"document_size": len(document_bytes),
"api_key_present": bool(self.mistral_api_key),
"api_key_format": f"sk-...{self.mistral_api_key[-4:]}" if self.mistral_api_key else "none"
})
# Detect file type and extension
def detect_file_info(data: bytes) -> tuple[str, str]:
if data.startswith(b'%PDF'):
return "application/pdf", ".pdf"
elif data.startswith(b'\xff\xd8\xff'): # JPEG
return "image/jpeg", ".jpg"
elif data.startswith(b'\x89PNG\r\n\x1a\n'): # PNG
return "image/png", ".png"
elif data.startswith(b'GIF87a') or data.startswith(b'GIF89a'): # GIF
return "image/gif", ".gif"
elif data.startswith(b'BM'): # BMP
return "image/bmp", ".bmp"
elif data.startswith(b'RIFF') and b'WEBP' in data[:12]: # WEBP
return "image/webp", ".webp"
elif data.startswith(b'II*\x00') or data.startswith(b'MM\x00*'): # TIFF
return "image/tiff", ".tiff"
elif data.startswith(b'\xd0\xcf\x11\xe0\xa1\xb1\x1a\xe1'): # DOC (OLE2)
return "application/msword", ".doc"
elif data.startswith(b'PK\x03\x04') and b'word/' in data[:1000]: # DOCX
return "application/vnd.openxmlformats-officedocument.wordprocessingml.document", ".docx"
else:
return "application/pdf", ".pdf"
mime_type, file_ext = detect_file_info(document_bytes)
# 🔍 DEBUGGING: Log document analysis
monitor.log_event("mistral_ocr_debug", {
"step": "document_analysis",
"mime_type": mime_type,
"file_extension": file_ext,
"document_size": len(document_bytes),
"document_start": document_bytes[:100].hex()[:50] + "..." if len(document_bytes) > 50 else document_bytes.hex()
})
try:
# 🔍 DEBUGGING: Log exact HTTP request details
monitor.log_event("mistral_http_debug", {
"step": "preparing_http_client",
"api_endpoint": "https://api.mistral.ai/v1/chat/completions",
"api_key_prefix": f"{self.mistral_api_key[:8]}..." if self.mistral_api_key else "none",
"timeout": 180.0,
"client_config": "httpx.AsyncClient() with default settings"
})
async with httpx.AsyncClient() as client:
# Handle PDF conversion to images
if mime_type == "application/pdf":
monitor.log_event("mistral_ocr_debug", {
"step": "pdf_detected_converting_to_images",
"pdf_size": len(document_bytes)
})
# Convert PDF to images
try:
image_bytes_list = self._convert_pdf_to_images(document_bytes)
monitor.log_event("mistral_ocr_debug", {
"step": "pdf_conversion_success",
"page_count": len(image_bytes_list)
})
except Exception as pdf_error:
monitor.log_event("mistral_ocr_debug", {
"step": "pdf_conversion_failed",
"error": str(pdf_error)
})
raise Exception(f"PDF conversion failed: {str(pdf_error)}")
# Process each page and combine results
all_extracted_text = []
for page_num, image_bytes in enumerate(image_bytes_list, 1):
monitor.log_event("mistral_ocr_debug", {
"step": f"processing_page_{page_num}",
"image_size": len(image_bytes)
})
# Convert image to base64
b64_data = base64.b64encode(image_bytes).decode()
# 🔍 DEBUGGING: Log exact HTTP request details
request_payload = {
"model": "pixtral-12b-2409",
"messages": [
{
"role": "user",
"content": [
{
"type": "text",
"text": f"""You are a strict OCR text extraction tool. Your job is to extract ONLY the actual text that appears in this image - nothing more, nothing less.
CRITICAL RULES:
- Extract ONLY text that is actually visible in the image
- Do NOT generate, invent, or create any content
- Do NOT add examples or sample data
- Do NOT fill in missing information
- If the image contains minimal text, return minimal text
- If the image is blank or contains no medical content, return what you actually see
For page {page_num}, extract exactly what text appears in this image:"""
},
{
"type": "image_url",
"image_url": {
"url": f"data:image/jpeg;base64,{b64_data[:50]}..." # Truncated for logging
}
}
]
}
],
"max_tokens": 8000,
"temperature": 0.0
}
monitor.log_event("mistral_http_request_start", {
"step": f"sending_request_page_{page_num}",
"url": "https://api.mistral.ai/v1/chat/completions",
"method": "POST",
"headers_count": 2,
"payload_size": len(str(request_payload)),
"b64_data_size": len(b64_data),
"timeout": min(300.0, 60.0 + (len(b64_data) / 100000)), # Dynamic timeout: 60s base + 1s per 100KB
"estimated_timeout": min(300.0, 60.0 + (len(b64_data) / 100000))
})
# Calculate dynamic timeout based on image size
dynamic_timeout = min(300.0, 60.0 + (len(b64_data) / 100000)) # Max 5 minutes
# API call for this page with dynamic timeout
response = await client.post(
"https://api.mistral.ai/v1/chat/completions",
headers={
"Authorization": f"Bearer {self.mistral_api_key}",
"Content-Type": "application/json"
},
json={
"model": "pixtral-12b-2409",
"messages": [
{
"role": "user",
"content": [
{
"type": "text",
"text": f"""You are a strict OCR text extraction tool. Your job is to extract ONLY the actual text that appears in this image - nothing more, nothing less.
CRITICAL RULES:
- Extract ONLY text that is actually visible in the image
- Do NOT generate, invent, or create any content
- Do NOT add examples or sample data
- Do NOT fill in missing information
- If the image contains minimal text, return minimal text
- If the image is blank or contains no medical content, return what you actually see
For page {page_num}, extract exactly what text appears in this image:"""
},
{
"type": "image_url",
"image_url": {
"url": f"data:image/jpeg;base64,{b64_data}"
}
}
]
}
],
"max_tokens": 8000,
"temperature": 0.0
},
timeout=dynamic_timeout
)
monitor.log_event("mistral_http_response_received", {
"step": f"response_page_{page_num}",
"status_code": response.status_code,
"response_size": len(response.content),
"headers": dict(response.headers),
"elapsed_seconds": response.elapsed.total_seconds() if hasattr(response, 'elapsed') else "unknown"
})
# Process response for this page
monitor.log_event("mistral_ocr_debug", {
"step": f"page_{page_num}_api_response",
"status_code": response.status_code
})
if response.status_code == 200:
result = response.json()
if 'choices' in result and len(result['choices']) > 0:
message = result['choices'][0].get('message', {})
page_text = message.get('content', '').strip()
if page_text:
cleaned_text = self._clean_ocr_text(page_text)
all_extracted_text.append(f"[PAGE {page_num}]\n{cleaned_text}")
monitor.log_event("mistral_ocr_debug", {
"step": f"page_{page_num}_extracted",
"text_length": len(cleaned_text)
})
else:
monitor.log_event("mistral_ocr_debug", {
"step": f"page_{page_num}_api_error",
"status_code": response.status_code,
"error": response.text
})
# Continue with other pages even if one fails
# Combine all pages
if all_extracted_text:
combined_text = "\n\n".join(all_extracted_text)
monitor.log_event("mistral_ocr_success", {
"mime_type": mime_type,
"total_pages": len(image_bytes_list),
"pages_processed": len(all_extracted_text),
"total_text_length": len(combined_text)
})
return f"[MISTRAL PDF PROCESSED - {len(image_bytes_list)} pages]\n\n{combined_text}"
else:
raise Exception("No text extracted from any PDF pages")
else:
# Handle non-PDF documents (images) - original logic
b64_data = base64.b64encode(document_bytes).decode()
b64_preview = b64_data[:100] + "..." if len(b64_data) > 100 else b64_data
monitor.log_event("mistral_ocr_debug", {
"step": "api_call_preparation",
"b64_data_length": len(b64_data),
"b64_preview": b64_preview,
"api_endpoint": "https://api.mistral.ai/v1/chat/completions",
"model": "pixtral-12b-2409"
})
# Calculate dynamic timeout based on image size
dynamic_timeout = min(300.0, 60.0 + (len(b64_data) / 100000)) # Max 5 minutes
monitor.log_event("mistral_http_request_start", {
"step": "sending_request_image",
"url": "https://api.mistral.ai/v1/chat/completions",
"method": "POST",
"mime_type": mime_type,
"b64_data_size": len(b64_data),
"timeout": dynamic_timeout,
"estimated_timeout": dynamic_timeout
})
response = await client.post(
"https://api.mistral.ai/v1/chat/completions",
headers={
"Authorization": f"Bearer {self.mistral_api_key}",
"Content-Type": "application/json"
},
json={
"model": "pixtral-12b-2409",
"messages": [
{
"role": "user",
"content": [
{
"type": "text",
"text": """You are a strict OCR text extraction tool. Your job is to extract ONLY the actual text that appears in this image - nothing more, nothing less.
CRITICAL RULES:
- Extract ONLY text that is actually visible in the image
- Do NOT generate, invent, or create any content
- Do NOT add examples or sample data
- Do NOT fill in missing information
- If the image contains minimal text, return minimal text
- If the image is blank or contains no medical content, return what you actually see
Extract exactly what text appears in this image:"""
},
{
"type": "image_url",
"image_url": {
"url": f"data:{mime_type};base64,{b64_data}"
}
}
]
}
],
"max_tokens": 8000,
"temperature": 0.0
},
timeout=dynamic_timeout
)
monitor.log_event("mistral_http_response_received", {
"step": "response_image",
"status_code": response.status_code,
"response_size": len(response.content),
"headers": dict(response.headers),
"elapsed_seconds": response.elapsed.total_seconds() if hasattr(response, 'elapsed') else "unknown"
})
# 🔍 DEBUGGING: Log API response
monitor.log_event("mistral_ocr_debug", {
"step": "api_response_received",
"status_code": response.status_code,
"response_headers": dict(response.headers),
"response_size": len(response.content),
"response_preview": response.text[:500] + "..." if len(response.text) > 500 else response.text
})
if response.status_code == 200:
result = response.json()
# 🔍 DEBUGGING: Log successful response parsing
monitor.log_event("mistral_ocr_debug", {
"step": "response_parsing_success",
"result_keys": list(result.keys()) if isinstance(result, dict) else "not_dict",
"choices_count": len(result.get("choices", [])) if isinstance(result, dict) else 0
})
# Log successful API response
monitor.log_event("mistral_api_success", {
"status_code": response.status_code,
"response_format": "valid"
})
# Extract text from Mistral chat completion response
if 'choices' in result and len(result['choices']) > 0:
message = result['choices'][0].get('message', {})
extracted_text = message.get('content', '').strip()
# Log OCR quality
monitor.log_event("mistral_response_has_content", {
"has_content": len(extracted_text) > 0,
"text_length": len(extracted_text)
})
if extracted_text:
# Clean up the response - remove any OCR processing artifacts
cleaned_text = self._clean_ocr_text(extracted_text)
# Log cleaned text quality
monitor.log_event("mistral_cleaned_text_substantial", {
"substantial": len(cleaned_text) > 20,
"text_length": len(cleaned_text)
})
# Log successful OCR metrics
monitor.log_event("mistral_ocr_success", {
"mime_type": mime_type,
"raw_length": len(extracted_text),
"cleaned_length": len(cleaned_text),
"cleaning_ratio": len(cleaned_text) / len(extracted_text) if extracted_text else 0
})
return f"[MISTRAL DOCUMENT AI PROCESSED - {mime_type}]\n\n{cleaned_text}"
else:
monitor.log_event("mistral_ocr_not_empty", {
"empty_response": True,
"mime_type": mime_type
})
monitor.log_event("mistral_ocr_empty_response", {"mime_type": mime_type})
raise Exception("Mistral OCR returned empty text content")
else:
monitor.log_event("mistral_response_format_valid", {
"format_valid": False,
"response_keys": list(result.keys()) if isinstance(result, dict) else "not_dict"
})
monitor.log_event("mistral_ocr_invalid_response", {"response": result})
raise Exception("Invalid response format from Mistral OCR API")
else:
# Handle API errors with detailed logging
error_msg = f"Mistral OCR API failed with status {response.status_code}"
try:
error_details = response.json()
error_msg += f": {error_details.get('message', 'Unknown error')}"
# Log specific error types for debugging
if response.status_code == 401:
monitor.log_event("mistral_auth_error", {"error": "Invalid API key"})
error_msg = "Mistral OCR authentication failed - check API key"
elif response.status_code == 429:
monitor.log_event("mistral_rate_limit", {"error": "Rate limit exceeded"})
error_msg = "Mistral OCR rate limit exceeded - try again later"
elif response.status_code == 413:
monitor.log_event("mistral_file_too_large", {"mime_type": mime_type})
error_msg = "Document too large for Mistral OCR processing"
else:
monitor.log_event("mistral_api_error", {
"status_code": response.status_code,
"error": error_details
})
except Exception:
error_text = response.text
error_msg += f": {error_text}"
monitor.log_event("mistral_unknown_error", {
"status_code": response.status_code,
"response": error_text
})
raise Exception(error_msg)
except Exception as e:
# 🔍 DEBUGGING: Log exception details
monitor.log_event("mistral_ocr_debug", {
"step": "exception_caught",
"exception_type": type(e).__name__,
"exception_message": str(e),
"exception_details": {
"args": e.args if hasattr(e, 'args') else "no_args",
"traceback_summary": f"{type(e).__name__}: {str(e)}"
}
})
# Re-raise with context for better debugging
raise Exception(f"Mistral OCR processing failed: {str(e)}")
def _clean_ocr_text(self, text: str) -> str:
"""Clean up OCR text output for medical documents"""
# Remove common OCR artifacts while preserving medical formatting
cleaned = text.strip()
# Remove any instruction responses or commentary
lines = cleaned.split('\n')
cleaned_lines = []
skip_patterns = [
"here is the extracted text",
"the extracted text is:",
"extracted text:",
"text content:",
"document content:",
]
for line in lines:
line_lower = line.lower().strip()
should_skip = any(pattern in line_lower for pattern in skip_patterns)
if not should_skip and line.strip():
cleaned_lines.append(line)
return '\n'.join(cleaned_lines)
async def _extract_with_multimodal(self, document_bytes: bytes) -> str:
"""Extract text using multimodal processor (simplified)"""
import base64
import sys
import os
# Add gaia system to path
gaia_path = os.path.join(os.path.dirname(__file__), "..", "..", "..", "gaia_agentic_system")
if gaia_path not in sys.path:
sys.path.append(gaia_path)
try:
from mcp_servers.multi_modal_processor_server import MultiModalProcessorServer
# Create processor instance
processor = MultiModalProcessorServer()
processor.initialize()
# Convert to base64
b64_data = base64.b64encode(document_bytes).decode()
# Analyze image for text extraction
result = await processor._analyze_image({
"image_data": b64_data,
"analysis_type": "text_extraction"
})
return result.get("extracted_text", "")
except Exception as e:
raise Exception(f"Multimodal processor failed: {str(e)}")
# Mock text method removed - never return dummy data for real medical processing
def _extract_medical_entities(self, text: str) -> dict:
"""Extract medical entities from actual OCR text using regex patterns"""
import re
entities = {
"patient_name": "Undefined",
"date_of_birth": "Undefined",
"conditions": [],
"medications": [],
"vitals": [],
"provider_name": "Undefined"
}
# Pattern for names (capitalized words, typically 2-3 parts)
name_patterns = [
r'Patient:?\s*([A-Z][a-z]+ [A-Z][a-z]+(?:\s+[A-Z][a-z]+)?)',
r'Name:?\s*([A-Z][a-z]+ [A-Z][a-z]+(?:\s+[A-Z][a-z]+)?)',
r'([A-Z][a-z]+,\s*[A-Z][a-z]+(?:\s+[A-Z][a-z]+)?)',
]
for pattern in name_patterns:
match = re.search(pattern, text)
if match:
entities["patient_name"] = match.group(1).strip()
break
# Pattern for dates of birth
dob_patterns = [
r'(?:DOB|Date of Birth|Born):?\s*(\d{1,2}[/-]\d{1,2}[/-]\d{2,4})',
r'(?:DOB|Date of Birth|Born):?\s*(\d{1,2}/\d{1,2}/\d{2,4})',
r'(?:DOB|Date of Birth|Born):?\s*([A-Z][a-z]+ \d{1,2},? \d{4})'
]
for pattern in dob_patterns:
match = re.search(pattern, text, re.IGNORECASE)
if match:
entities["date_of_birth"] = match.group(1).strip()
break
# Pattern for medical conditions
condition_keywords = [
r'(?:Diagnosis|Condition|History):?\s*([A-Z][a-z]+(?: [a-z]+)*)',
r'([A-Z][a-z]+(?:itis|osis|emia|pathy|trophy|plasia))',
r'(Hypertension|Diabetes|Asthma|COPD|Depression|Anxiety)'
]
for pattern in condition_keywords:
matches = re.findall(pattern, text, re.IGNORECASE)
for match in matches:
condition = match if isinstance(match, str) else match[0]
if condition and len(condition) > 2:
entities["conditions"].append(condition.strip())
# Pattern for medications
med_patterns = [
r'(?:Medication|Med|Rx):?\s*([A-Z][a-z]+(?:ol|ine|ide|ate|pril|statin))',
r'([A-Z][a-z]+(?:ol|ine|ide|ate|pril|statin))\s*\d+\s*mg',
r'(Lisinopril|Metformin|Aspirin|Ibuprofen|Acetaminophen)'
]
for pattern in med_patterns:
matches = re.findall(pattern, text, re.IGNORECASE)
for match in matches:
medication = match if isinstance(match, str) else match[0]
if medication and len(medication) > 2:
entities["medications"].append(medication.strip())
# Pattern for vital signs
vital_patterns = [
r'(?:BP|Blood Pressure):?\s*(\d{2,3}/\d{2,3})',
r'(?:Heart Rate|HR):?\s*(\d{2,3})\s*bpm',
r'(?:Temperature|Temp):?\s*(\d{2,3}(?:\.\d)?)\s*°?F?',
r'(?:Weight):?\s*(\d{2,3})\s*lbs?',
r'(?:Height):?\s*(\d+)\'?\s*(\d+)"?'
]
for pattern in vital_patterns:
matches = re.findall(pattern, text, re.IGNORECASE)
for match in matches:
vital = match if isinstance(match, str) else ' '.join(filter(None, match))
if vital:
entities["vitals"].append(vital.strip())
# Pattern for provider/doctor names
provider_patterns = [
r'(?:Dr\.|Doctor|Physician):?\s*([A-Z][a-z]+ [A-Z][a-z]+)',
r'Provider:?\s*([A-Z][a-z]+ [A-Z][a-z]+)',
r'Attending:?\s*([A-Z][a-z]+ [A-Z][a-z]+)'
]
for pattern in provider_patterns:
match = re.search(pattern, text)
if match:
entities["provider_name"] = match.group(1).strip()
break
return entities
def _create_simple_fhir_bundle(self, entities: dict, user_id: str) -> dict:
"""Create FHIR bundle from extracted entities"""
bundle_id = f"local-{uuid.uuid4()}"
# Parse patient name
patient_name = entities.get("patient_name", "Undefined")
if patient_name != "Undefined" and " " in patient_name:
name_parts = patient_name.split()
given_name = name_parts[0] if len(name_parts) > 0 else "Undefined"
family_name = " ".join(name_parts[1:]) if len(name_parts) > 1 else "Undefined"
else:
given_name = "Undefined"
family_name = "Undefined"
# Create bundle entries
entries = []
# Patient resource
patient_resource = {
"resource": {
"resourceType": "Patient",
"id": "local-patient",
"name": [{"given": [given_name], "family": family_name}]
}
}
# Add birth date if available
if entities.get("date_of_birth") != "Undefined":
patient_resource["resource"]["birthDate"] = entities["date_of_birth"]
entries.append(patient_resource)
# Add conditions as Condition resources
for i, condition in enumerate(entities.get("conditions", [])):
if condition:
entries.append({
"resource": {
"resourceType": "Condition",
"id": f"local-condition-{i}",
"subject": {"reference": "Patient/local-patient"},
"code": {
"text": condition
},
"clinicalStatus": {
"coding": [{
"system": "http://terminology.hl7.org/CodeSystem/condition-clinical",
"code": "active"
}]
}
}
})
# Add medications as MedicationStatement resources
for i, medication in enumerate(entities.get("medications", [])):
if medication:
entries.append({
"resource": {
"resourceType": "MedicationStatement",
"id": f"local-medication-{i}",
"subject": {"reference": "Patient/local-patient"},
"medicationCodeableConcept": {
"text": medication
},
"status": "active"
}
})
# Add vitals as Observation resources
for i, vital in enumerate(entities.get("vitals", [])):
if vital:
entries.append({
"resource": {
"resourceType": "Observation",
"id": f"local-vital-{i}",
"subject": {"reference": "Patient/local-patient"},
"status": "final",
"code": {
"text": "Vital Sign"
},
"valueString": vital
}
})
return {
"resourceType": "Bundle",
"id": bundle_id,
"type": "document",
"timestamp": datetime.now().isoformat(),
"entry": entries,
"_metadata": {
"processing_mode": self._get_processing_mode(),
"entities_found": len(entities.get("conditions", [])) + len(entities.get("medications", [])) + len(entities.get("vitals", [])),
"processed_by": user_id,
"patient_name": entities.get("patient_name", "Undefined"),
"provider_name": entities.get("provider_name", "Undefined")
}
}
def _get_processing_mode(self) -> str:
"""Determine current processing mode"""
if self.use_mistral_fallback and self.mistral_api_key:
return "local_processing_with_mistral_ocr"
elif self.use_multimodal_fallback:
return "local_processing_with_multimodal_fallback"
else:
return "local_processing_only"
# Global instance
local_processor = LocalProcessor() |