Spaces:
Sleeping
Sleeping
File size: 24,000 Bytes
6f9d4a7 f8f2a5f 6f9d4a7 f8f2a5f 6f9d4a7 f8f2a5f 6f9d4a7 f8f2a5f 6f9d4a7 f8f2a5f 6f9d4a7 f8f2a5f 6f9d4a7 f8f2a5f 6f9d4a7 f8f2a5f 6f9d4a7 f8f2a5f 6f9d4a7 f8f2a5f 6f9d4a7 f8f2a5f 6f9d4a7 f8f2a5f 6f9d4a7 f8f2a5f 6f9d4a7 f8f2a5f 6f9d4a7 f8f2a5f 6f9d4a7 a081bdc |
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 |
#!/usr/bin/env python3
"""
OCR Document Verification with Batch Processing & Required Document Checklist
Usage:
# Single file (backward compatible)
python ocrupdated2.py --file image.jpg --inputkeywords "keyword1 keyword2" --fuzzy --debug
# Multiple files with required document checklist
python ocrupdated2.py --file doc1.pdf doc2.jpg doc3.png --inputkeywords "Shaikh Anisa Rahat" --required PAN HSC AgeNationalityDomicile --fuzzy --debug
NOTE: Use spaces to separate required document types, NOT commas:
✅ --required PAN Aadhaar HSC
❌ --required PAN, Aadhaar, HSC
"""
import argparse
import re
import os
import tempfile
from collections import defaultdict
from paddleocr import PaddleOCR
import difflib
from concurrent.futures import ThreadPoolExecutor
import multiprocessing
import sys
# Optional PDF support
try:
import fitz # PyMuPDF
PDF_SUPPORT = True
except ImportError:
PDF_SUPPORT = False
print("Warning: PyMuPDF not installed. PDF support disabled. Install with: pip install PyMuPDF")
# Document keywords (unchanged)
DOC_KEYWORDS = {
"Aadhaar": [
"uidai", "aadhaar", "aadhar", "government of india", "भारत सरकार",
"आधार", "यूआईडीएआई", "प्रधानमंत्री", "जन्म तिथि", "पता", "लिंग",
"unique identification authority", "aadhaar number", "enrollment number"
],
"PAN": [
"permanent account number", "income tax", "incometaxindia", "pan",
"income tax department", "आयकर विभाग", "स्थायी खाता संख्या",
"taxpayer", "father's name", "पिता का नाम", "signature", "inc"
],
"Driving_License": [
"driving licence", "motor vehicles act", "rto", "mcwg", "lmv",
"transport department", "licence no", "valid till", "date of issue",
"ड्राइविंग लाइसेंस", "परिवहन विभाग", "challan", "regional transport office"
],
"Passport": [
"passport", "republic of india", "ministry of external affairs",
"passport number", "date of issue", "date of expiry", "surname",
"given names", "nationality indian", "पासपोर्ट", "गणराज्य", "विदेश मंत्रालय",
"consular", "visa"
],
"SSC": [
"secondary school certificate", "statement of marks", "ssc", "10th", "class x",
"board of secondary education", "maharashtra state board", "matriculation",
"roll number", "seat number", "subject code", "marks obtained", "grade", "pass"
],
"HSC": [
"higher secondary certificate", "statement of marks", "hsc", "12th", "class xii",
"board of higher secondary education", "maharashtra state board", "intermediate",
"stream", "science", "commerce", "arts", "marks obtained", "grade", "percentage"
],
"AgeNationalityDomicile": [
"certificate of age nationality and domicile", "domicile certificate",
"age nationality domicile", "tehsildar", "executive magistrate", "collector",
"certificate of residence", "domiciled in the state of", "citizen of india",
"residence proof", "maharashtra domicile", "satara", "karad", "taluka", "district"
],
"Ration_Card": [
"ration card", "food and civil supplies", "apl", "bpl", "aay", "antyodaya",
"ration card number", "family members", "head of family",
"राशन कार्ड", "खाद्य पुरवठा", "नागरी पुरवठा विभाग", "fps", "fair price shop"
],
"Cast_Certificate": [
"CASTE CERTIFICATE",
"FORM - 8",
"Rule No. 5(6)",
"De-Notified Tribe (Vimukt Jati)",
"Nomadic Tribe/Other Backward Class",
"Special Backward Category",
"recognised as",
"Government Resolution",
"Sub Divisional Officer",
"belonging to the State of Maharashtra"
],
"Income_Certificate": [
"१ वर्षासाठी उत्पन्नाचे प्रमाणपत्र",
"ऑफिस ऑफ नायब तहसीलदार",
"वार्षिक उत्पन्न",
"मिळालेले १ वर्षाचे उत्पन्न",
"कुटुंबातील सर्व सदस्यांचे",
"प्रमाणित करण्यात येते की",
"वैध राहील",
"Signature valid",
"Digitally Signed by"
],
"PCM_Score_Card": [
"MAH-MHT CET (PCM Group)",
"State Common Entrance Test Cell",
"Score Card",
"Physics",
"Chemistry",
"Mathematics",
"Total Percentile",
"Normalization document",
"Centralized Admission Process (CAP)",
"IP address of the Computer"
]
}
# Validate keyword uniqueness (unchanged)
_keyword_sets = {k: set(v) for k, v in DOC_KEYWORDS.items()}
for doc1 in DOC_KEYWORDS:
for doc2 in DOC_KEYWORDS:
if doc1 < doc2:
overlap = _keyword_sets[doc1].intersection(_keyword_sets[doc2])
if overlap:
print(f"⚠️ Warning: Overlap between {doc1} and {doc2}: {overlap}")
# NEW: Pre-compile regex patterns for performance
NOISE_PATTERN = re.compile(r'^[b-df-hj-np-tv-xz]{4,}$')
TOKEN_PATTERN = re.compile(r'[\u0900-\u097F]{2,}|\w{3,}')
STOPWORDS = {'the', 'and', 'of', 'in', 'to', 'for', 'is', 'on', 'by', 'with', 'at', 'from', 'a', 'an', 'this'}
def normalize_text(text):
"""Robust multilingual tokenization with noise filtering"""
text = text.lower()
# Extract Hindi Devanagari (2+ chars) OR English alphanumeric (3+ chars)
tokens = TOKEN_PATTERN.findall(text)
# Remove common English stopwords
tokens = [t for t in tokens if t not in STOPWORDS]
# Remove OCR noise (4+ consecutive consonants = garbage)
tokens = [t for t in tokens if not NOISE_PATTERN.match(t)]
return tokens
def pdf_to_images(pdf_path, max_pages=3):
"""Convert PDF pages to high-resolution temporary images"""
if not PDF_SUPPORT:
raise ValueError("PDF support not available. Install PyMuPDF")
doc = fitz.open(pdf_path)
total_pages = len(doc)
pages_to_process = min(total_pages, max_pages)
image_paths = []
temp_dir = tempfile.mkdtemp(prefix="ocr_pdf_")
for page_num in range(pages_to_process):
page = doc.load_page(page_num)
zoom = 2 # 2x resolution for better OCR
mat = fitz.Matrix(zoom, zoom)
pix = page.get_pixmap(matrix=mat)
img_path = os.path.join(temp_dir, f"page_{page_num + 1}.png")
pix.save(img_path)
image_paths.append(img_path)
doc.close()
return image_paths, total_pages, temp_dir
def process_page_ocr(img_path, page_num, ocr, debug):
"""Process a single page with OCR (for parallel execution)"""
try:
if debug:
print(f"\n--- Processing PDF Page {page_num} ---")
result = ocr.predict(input=img_path)
texts = []
for res in result:
texts.extend(res['rec_texts'])
return texts
except Exception as e:
print(f"❌ ERROR: OCR failed on page {page_num}: {str(e)}")
return []
def get_ocr_text(file_path, ocr, max_pages=3, debug=False):
"""Process image or PDF with OCR, returning all extracted text lines"""
all_texts = []
temp_dir = None
try:
if file_path.lower().endswith('.pdf'):
if not PDF_SUPPORT:
print("Error: PDF file provided but PyMuPDF not installed")
return []
image_paths, total_pages, temp_dir = pdf_to_images(file_path, max_pages)
print(f"Processing PDF: {total_pages} pages total, processing first {len(image_paths)} pages...")
# Process pages in parallel
max_workers = min(len(image_paths), 4)
with ThreadPoolExecutor(max_workers=max_workers) as executor:
future_to_page = {
executor.submit(process_page_ocr, img_path, i+1, ocr, debug): i
for i, img_path in enumerate(image_paths)
}
page_results = [None] * len(image_paths)
for future in future_to_page:
page_idx = future_to_page[future]
try:
page_results[page_idx] = future.result()
except Exception as e:
print(f"❌ ERROR: Failed to process page {page_idx+1}: {str(e)}")
page_results[page_idx] = []
for texts in page_results:
all_texts.extend(texts)
else:
result = ocr.predict(input=file_path)
for res in result:
all_texts.extend(res['rec_texts'])
except Exception as e:
print(f"❌ ERROR: Failed to process file {file_path}: {str(e)}")
return []
finally:
if temp_dir and os.path.exists(temp_dir):
import shutil
shutil.rmtree(temp_dir)
return all_texts
def fuzzy_match(token, target_set, threshold=0.75):
"""
Multi-level matching for OCR errors:
1. Exact match
2. Levenshtein distance
3. Substring containment
4. Hindi character-level similarity
"""
if token in target_set:
return token
# Levenshtein distance match
matches = difflib.get_close_matches(token, target_set, n=1, cutoff=threshold)
if matches:
return matches[0]
# Substring match (handles concatenated words)
for ocr_token in target_set:
if token in ocr_token or ocr_token in token:
return ocr_token
# Hindi-specific fuzzy matching
if any('\u0900' <= c <= '\u097F' for c in token):
for ocr_token in target_set:
if len(ocr_token) > 3:
similarity = difflib.SequenceMatcher(None, token, ocr_token).ratio()
if similarity > threshold:
return ocr_token
return None
def calculate_doc_type(ocr_tokens, debug=False):
"""
Enhanced document classification with CORRECTED tie-breaking logic.
Only compares documents that are ACTUALLY TIED (within 5% score).
"""
ocr_set = set(ocr_tokens)
ocr_combined = " ".join(ocr_tokens)
scores = {}
# Pre-calculate keyword sets once
doc_keyword_sets = {}
for doc_type, keywords in DOC_KEYWORDS.items():
doc_keyword_sets[doc_type] = set(k.lower() for k in keywords)
for doc_type, kw_set in doc_keyword_sets.items():
# Primary: exact/fuzzy token matches (weighted 2 for exact, 1.5 for fuzzy)
primary_matches = 0
for kw in kw_set:
if kw in ocr_set:
primary_matches += 2
elif fuzzy_match(kw, ocr_set):
primary_matches += 1.5
# Secondary: multi-word phrase matches in combined text
phrase_matches = sum(1 for kw in kw_set if " " in kw and kw in ocr_combined)
# Tertiary: title keyword bonus
title_keywords = [kw for kw in kw_set if any(word in kw for word in ["certificate", "card", "licence", "passport"])]
title_match = sum(1 for kw in title_keywords if kw in ocr_combined)
# Calculate weighted score
max_possible = len(kw_set) * 2
weighted_score = ((primary_matches + phrase_matches + title_match) / max_possible) * 100
scores[doc_type] = weighted_score
if debug:
print(f" {doc_type:<25}: {weighted_score:>6.1f}% ({primary_matches:.1f} + {phrase_matches} + {title_match})")
# Sort by score descending
sorted_scores = sorted(scores.items(), key=lambda x: x[1], reverse=True)
best_type, best_score = sorted_scores[0]
# Tie-breaking logic
if len(sorted_scores) > 1 and (sorted_scores[0][1] - sorted_scores[1][1]) < 5:
if debug:
print(f"\n⚠️ Tie detected between '{sorted_scores[0][0]}' and '{sorted_scores[1][0]}'!")
tied_docs = [(doc_type, score) for doc_type, score in sorted_scores
if (best_score - score) < 5]
if debug:
print(f"Tied documents: {[f'{doc}({score:.1f}%)' for doc, score in tied_docs]}")
unique_counts = {}
for doc_type, _ in tied_docs:
kw_set = doc_keyword_sets[doc_type]
other_tied_keywords = set()
for other_doc, _ in tied_docs:
if other_doc != doc_type:
other_tied_keywords.update(doc_keyword_sets[other_doc])
unique_keywords = kw_set - other_tied_keywords
unique_matches = sum(1 for kw in unique_keywords if fuzzy_match(kw, ocr_set))
unique_counts[doc_type] = unique_matches
if debug:
print(f" {doc_type:<25}: {unique_matches} unique matches ({len(unique_keywords)} available)")
if unique_counts and max(unique_counts.values()) > 0:
sorted_unique = sorted(unique_counts.items(), key=lambda x: x[1], reverse=True)
if len(sorted_unique) > 1 and sorted_unique[0][1] > sorted_unique[1][1]:
best_type = sorted_unique[0][0]
best_score = scores[best_type]
if debug:
print(f"✓ Tie broken: {best_type} wins with {unique_counts[best_type]} unique matches")
return best_type, best_score
def verify_keywords(ocr_tokens, user_keywords, use_fuzzy=False):
"""
Sequence-aware matching for multi-keyword inputs.
Checks if keywords appear consecutively in OCR text first.
"""
ocr_set = set(ocr_tokens)
ocr_combined = " ".join(ocr_tokens)
results = []
if len(user_keywords) > 1:
user_phrase = " ".join([kw.lower() if all(ord(c) < 128 for c in kw) else kw for kw in user_keywords])
if user_phrase in ocr_combined:
for kw in user_keywords:
results.append({
'keyword': kw,
'matched': True,
'matched_text': kw
})
return results
if use_fuzzy:
n = len(user_keywords)
ocr_phrases = [" ".join(ocr_tokens[i:i+n]) for i in range(len(ocr_tokens) - n + 1)]
phrase_match = fuzzy_match(user_phrase, set(ocr_phrases))
if phrase_match:
for kw in user_keywords:
results.append({
'keyword': kw,
'matched': True,
'matched_text': kw
})
return results
# Fallback to individual keyword matching
for kw in user_keywords:
kw_processed = kw.lower() if all(ord(c) < 128 for c in kw) else kw
matched = False
matched_text = None
if kw_processed in ocr_set:
matched = True
matched_text = kw_processed
elif " " in kw_processed and kw_processed in ocr_combined:
matched = True
matched_text = kw_processed
elif use_fuzzy:
matched_text = fuzzy_match(kw_processed, ocr_set)
if matched_text:
matched = True
results.append({
'keyword': kw,
'matched': matched,
'matched_text': matched_text or kw_processed if matched else None
})
return results
def main():
parser = argparse.ArgumentParser(description='OCR Document Verification with PDF Support')
parser.add_argument('--file', nargs='+', required=True, help='Paths to image or PDF files')
parser.add_argument('--inputkeywords', required=True, help='Space-separated keywords to verify')
parser.add_argument('--required', nargs='+', help='List of required document types')
parser.add_argument('--fuzzy', action='store_true', help='Enable fuzzy matching')
parser.add_argument('--debug', action='store_true', help='Show detailed OCR and scoring output')
parser.add_argument('--pages', type=int, default=3, help='Max pages to process for PDFs')
global args
args = parser.parse_args()
# Clean required list
required_list = []
if args.required:
for item in args.required:
parts = [part.strip() for part in item.split(',') if part.strip()]
required_list.extend(parts)
required_set = set(required_list)
# Initialize OCR once, reuse for all files
print("Initializing OCR engine (first run may take a few seconds)...")
try:
ocr_engine = PaddleOCR(
lang="mr",
use_doc_orientation_classify=False,
use_doc_unwarping=False,
use_textline_orientation=False,
max_batch_size=16,
num_workers=min(4, multiprocessing.cpu_count()),
)
# Test if OCR is working
test_result = ocr_engine.predict(input="")
if not test_result:
print("⚠️ WARNING: OCR engine test returned empty result. Models may not be loaded correctly.")
except Exception as e:
print(f"❌ CRITICAL ERROR: Failed to initialize OCR engine: {str(e)}")
print("Please ensure PaddleOCR is installed correctly and models are downloaded.")
sys.exit(1)
# Process each file and collect results
file_results = []
found_documents = set()
all_matched_keywords_per_file = []
print(f"\n{'='*60}")
print(f"PROCESSING {len(args.file)} FILES")
print(f"{'='*60}\n")
for idx, file_path in enumerate(args.file, 1):
print(f"--- FILE {idx}/{len(args.file)}: {os.path.basename(file_path)} ---")
# Check if file exists
if not os.path.exists(file_path):
print(f"❌ ERROR: File not found: {file_path}\n")
file_results.append({
'file': file_path,
'doc_type': 'Unknown',
'doc_score': 0,
'keywords_matched': [],
'status': 'ERROR'
})
continue
# Extract text from file
ocr_texts = get_ocr_text(file_path, ocr_engine, args.pages, args.debug)
if not ocr_texts:
print(f"⚠️ No text extracted from {file_path}")
print(" Possible causes:")
print(" - File is corrupted or empty")
print(" - OCR engine failed to process the file")
print(" - Text is not in supported language/format")
print(" Try running with --debug flag to see detailed OCR output\n")
file_results.append({
'file': file_path,
'doc_type': 'Unknown',
'doc_score': 0,
'keywords_matched': [],
'status': 'ERROR'
})
continue
# Show OCR summary even without debug if text is very short
if len(ocr_texts) < 5 and not args.debug:
print(f" ℹ️ Only {len(ocr_texts)} lines of text extracted. Run with --debug to see details.")
# Normalize tokens
ocr_tokens = normalize_text(" ".join(ocr_texts))
# Show token count
print(f" Extracted {len(ocr_tokens)} valid tokens from OCR text")
# Debug: Show normalized tokens
if args.debug:
print("="*60)
print("NORMALIZED TOKENS:")
print("="*60)
print(f"Total tokens: {len(ocr_tokens)}")
print(f"First 50 tokens: {', '.join(ocr_tokens[:50])}{'...' if len(ocr_tokens) > 50 else ''}")
print("="*60 + "\n")
# Document classification
if args.debug:
print("="*60)
print("DOCUMENT TYPE SCORING:")
print("="*60)
doc_type, doc_score = calculate_doc_type(ocr_tokens, debug=args.debug)
found_documents.add(doc_type)
if args.debug:
print("="*60 + "\n")
# Keyword verification
user_keywords = [kw.strip() for kw in args.inputkeywords.split()]
verification_results = verify_keywords(ocr_tokens, user_keywords, args.fuzzy)
# Status: ALL keywords must match in this file
all_matched = all(r['matched'] for r in verification_results)
status = "VERIFIED" if all_matched else "NOT VERIFIED"
# Store results for this file
file_results.append({
'file': file_path,
'doc_type': doc_type,
'doc_score': doc_score,
'keywords_matched': verification_results,
'status': status,
'all_keywords_matched': all_matched
})
# Track which keywords were matched in this file
matched_keywords_in_file = {r['keyword'] for r in verification_results if r['matched']}
all_matched_keywords_per_file.append(matched_keywords_in_file)
# Per-file output
print(f"\n{'='*60}")
print(f"Document Type: {doc_type} ({doc_score:.1f}% confidence)")
print(f"{'='*60}")
print(f"{'Keyword':<25} | {'Status':<10} | {'Matched Text'}")
print(f"{'-'*60}")
for r in verification_results:
status_icon = "✓" if r['matched'] else "✗"
matched_text = r['matched_text'] if r['matched_text'] else "Not found"
print(f"{r['keyword']:<25} | {status_icon:<10} | {matched_text}")
print(f"{'='*60}")
print(f"File Status: {status}")
print(f"{'='*60}\n")
# FINAL SUMMARY (unchanged)
print(f"\n{'='*60}")
print(f"FINAL SUMMARY")
print(f"{'='*60}")
# Required documents check
if required_set:
missing_docs = required_set - found_documents
print(f"\nRequired Documents: {', '.join(sorted(required_set))}")
print(f"Found Documents: {', '.join(sorted(found_documents)) if found_documents else 'None'}")
if missing_docs:
print(f"❌ Missing Documents: {', '.join(sorted(missing_docs))}")
docs_status = "NOT VERIFIED"
else:
print(f"✅ All required documents found!")
docs_status = "VERIFIED"
else:
docs_status = "N/A (no required list specified)"
missing_docs = set()
# Overall keyword verification across ALL files
all_user_keywords = set(args.inputkeywords.split())
keywords_found_across_files = set()
for file_keyword_set in all_matched_keywords_per_file:
keywords_found_across_files.update(file_keyword_set)
missing_keywords = all_user_keywords - keywords_found_across_files
print(f"\nKeywords to Find: {', '.join(sorted(all_user_keywords))}")
print(f"Keywords Found (across all files): {', '.join(sorted(keywords_found_across_files)) if keywords_found_across_files else 'None'}")
if missing_keywords:
print(f"❌ Missing Keywords: {', '.join(sorted(missing_keywords))}")
keywords_status = "NOT VERIFIED"
else:
print(f"✅ All keywords found across uploaded documents!")
keywords_status = "VERIFIED"
# Overall status
overall_status = "VERIFIED" if (docs_status == "VERIFIED" and keywords_status == "VERIFIED") else "NOT VERIFIED"
print(f"\n{'='*60}")
print(f"Documents Status: {docs_status}")
print(f"Keywords Status: {keywords_status}")
print(f"OVERALL STATUS: {overall_status}")
print(f"{'='*60}")
if __name__ == "__main__":
main() |