Spaces:
Sleeping
Sleeping
File size: 55,353 Bytes
c362ce2 459923e 6df0748 459923e 6df0748 459923e 7ac4443 459923e 7ac4443 459923e 7ac4443 459923e 7ac4443 459923e 7ac4443 459923e 7ac4443 459923e 7ac4443 459923e 7ac4443 459923e 7ac4443 459923e 7ac4443 459923e 7ac4443 459923e 1f043fb 1e6c9c5 1f043fb 1e6c9c5 1f043fb 459923e 9df92e6 459923e 9df92e6 459923e 7ac4443 459923e bc0ab93 1e6c9c5 bc0ab93 1e6c9c5 bc0ab93 459923e 9df92e6 459923e 9df92e6 459923e 7ac4443 459923e 64684cb 459923e 7ac4443 1f3ea22 7ac4443 1f3ea22 7ac4443 0102648 1f3ea22 7ac4443 1e6c9c5 1f3ea22 7ac4443 1f3ea22 1e6c9c5 1f3ea22 1e6c9c5 1f3ea22 1e6c9c5 1f3ea22 1e6c9c5 1f3ea22 1e6c9c5 1f3ea22 1e6c9c5 1f3ea22 1e6c9c5 1f3ea22 7ac4443 1f3ea22 7ac4443 1f3ea22 7ac4443 1f3ea22 7ac4443 1f3ea22 7ac4443 1f3ea22 7ac4443 1f3ea22 7ac4443 1f3ea22 1e6c9c5 1f3ea22 1e6c9c5 1f3ea22 1e6c9c5 1f3ea22 7ac4443 6df0748 459923e 6df0748 459923e 6df0748 459923e 64684cb 459923e 6df0748 459923e 6df0748 459923e 9df92e6 6df0748 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 |
import logging
logger = logging.getLogger(__name__)
logger.info("Importing Feedback.py...")
import openai
from docx import Document
import json
import re
import os
import tiktoken
from typing import List, Dict, Tuple, Optional, Any
import unicodedata
class Grader:
def __init__(self, api_key, config: Optional[Dict[str, Any]] = None):
logger.info("Initializing Grader...")
self.api_key = api_key
openai.api_key = self.api_key
try:
self.client = openai.OpenAI(api_key=self.api_key)
except AttributeError:
self.client = openai
try:
self.encoding = tiktoken.encoding_for_model("gpt-4o")
logger.info("Successfully initialized tiktoken encoding")
except Exception as e:
logger.warning(f"Failed to initialize tiktoken: {e}")
self.encoding = None
# Fixed config, no runtime update
self.config = {
'enable_validation': True,
'enable_enhanced_logging': True,
'fallback_to_legacy': True,
'aggregate_scores': True,
'log_missing_categories': True
}
logger.info(f"Grader initialized with config: {self.config}")
def count_tokens(self, text):
if not self.encoding:
return len(text) // 4
try:
return len(self.encoding.encode(text))
except Exception as e:
logger.warning(f"Error counting tokens: {e}")
return len(text) // 4
def process_full_text(self, text):
if not text:
return text, 0, False
# Store original text for comparison
original_text = text
# More conservative character filtering - only remove truly problematic control characters
# Keep more Unicode characters that might be meaningful
text = ''.join(char for char in text if (
unicodedata.category(char)[0] != 'C' or # Keep control chars
char in '\n\r\t' or # Keep newlines, returns, tabs
unicodedata.category(char) in ['Cc', 'Cf', 'Cs'] # Only remove specific control categories
))
# Normalize Unicode but be more careful
text = unicodedata.normalize('NFKC', text)
# More selective character replacements - only replace if they cause issues
replacements = {
'\u201c': '"', # Left double quotation mark
'\u201d': '"', # Right double quotation mark
'\u2018': "'", # Left single quotation mark
'\u2019': "'", # Right single quotation mark
'\u2013': '-', # En dash
'\u2014': '--', # Em dash (replace with two dashes)
'\u2022': '•', # Bullet
'\u00a0': ' ', # Non-breaking space
'\u2026': '...', # Horizontal ellipsis
}
for old_char, new_char in replacements.items():
text = text.replace(old_char, new_char)
# Log if significant changes were made
if len(text) != len(original_text):
logger.info(f"Text processing: {len(original_text)} -> {len(text)} characters")
if len(text) < len(original_text) * 0.95: # If more than 5% was removed
logger.warning(f"Significant text reduction detected: {len(original_text)} -> {len(text)} characters")
token_count = self.count_tokens(text)
logger.info(f"Full text token count: {token_count} - NO TRUNCATION")
return text, token_count, False
def read_file(self, file_path):
logger.info(f"Reading file: {file_path}")
if file_path.endswith('.txt'):
with open(file_path, 'r', encoding='utf-8') as file:
return file.read().strip()
elif file_path.endswith('.docx'):
doc = Document(file_path)
return '\n'.join([para.text for para in doc.paragraphs]).strip()
else:
raise ValueError("Unsupported file format. Please use .txt or .docx files.")
def extract_json_from_text(self, text):
try:
return json.loads(text)
except json.JSONDecodeError as e:
logger.warning(f"Initial JSON parsing failed: {str(e)}")
logger.info(f"Raw response text: {text[:500]}...") # Log first 500 chars for debugging
start_idx = text.find('{')
end_idx = text.rfind('}')
if start_idx == -1 or end_idx == -1:
logger.error("No JSON object markers found in response")
raise ValueError("No valid JSON object found in the response")
json_str = text[start_idx:end_idx + 1]
logger.info(f"Extracted JSON string: {json_str[:200]}...") # Log first 200 chars
# Remove markdown formatting
json_str = json_str.replace('```json', '').replace('```', '')
# Remove control characters except newlines, returns, tabs
json_str = ''.join(char for char in json_str if (
unicodedata.category(char)[0] != 'C' or
char in '\n\r\t' or
unicodedata.category(char) in ['Cc', 'Cf', 'Cs']
))
# Normalize Unicode
json_str = unicodedata.normalize('NFKC', json_str)
# Replace common problematic characters
replacements = {
'\u201c': '"', # Left double quotation mark
'\u201d': '"', # Right double quotation mark
'\u2018': "'", # Left single quotation mark
'\u2019': "'", # Right single quotation mark
'\u2013': '-', # En dash
'\u2014': '--', # Em dash
'\u2022': '•', # Bullet
'\u00a0': ' ', # Non-breaking space
'\u2026': '...', # Horizontal ellipsis
}
for old_char, new_char in replacements.items():
json_str = json_str.replace(old_char, new_char)
# Clean up whitespace and formatting
json_str = re.sub(r'[\r\n\t]+', ' ', json_str)
json_str = re.sub(r'\s+', ' ', json_str)
# Remove trailing commas before closing brackets/braces
json_str = re.sub(r',\s*([}\]])', r'\1', json_str)
# Ensure property names are quoted
json_str = re.sub(r'([{,])\s*([a-zA-Z_][a-zA-Z0-9_]*)\s*:', r'\1"\2":', json_str)
# Handle escaped quotes properly
json_str = json_str.replace('\\"', '___ESCAPED_QUOTE___')
json_str = re.sub(r'(?<!\\)\'', '"', json_str)
json_str = json_str.replace('___ESCAPED_QUOTE___', '\\"')
# Additional fixes for common JSON issues
# Fix unquoted string values
json_str = re.sub(r':\s*([a-zA-Z][a-zA-Z0-9\s]*?)(?=\s*[,}])', r': "\1"', json_str)
# Fix missing quotes around property names that might have been missed
json_str = re.sub(r'([{,])\s*([a-zA-Z_][a-zA-Z0-9_]*)\s*:', r'\1"\2":', json_str)
logger.info(f"Cleaned JSON string: {json_str[:200]}...") # Log first 200 chars after cleaning
try:
parsed_json = json.loads(json_str)
logger.info("JSON parsing successful after cleaning")
return parsed_json
except json.JSONDecodeError as e2:
logger.error(f"JSON parsing still failed after cleaning: {str(e2)}")
logger.error(f"Problematic JSON: {json_str}")
# Try to create a fallback response
fallback_response = {
"categories": {
"grammar_punctuation": {
"analysis": "Unable to parse detailed response. Basic grammar analysis completed.",
"issues": [{"type": "parsing_error", "before": "JSON parsing failed", "after": "Please try again", "explanation": "Technical error in response parsing"}],
"positive_points": ["Essay submitted successfully"],
"suggestions": ["Please try submitting again"]
},
"vocabulary_usage": {
"analysis": "Unable to parse detailed response. Basic vocabulary analysis completed.",
"issues": [{"type": "parsing_error", "before": "JSON parsing failed", "after": "Please try again", "explanation": "Technical error in response parsing"}],
"positive_points": ["Essay submitted successfully"],
"suggestions": ["Please try submitting again"]
},
"sentence_structure": {
"analysis": "Unable to parse detailed response. Basic structure analysis completed.",
"issues": [{"type": "parsing_error", "before": "JSON parsing failed", "after": "Please try again", "explanation": "Technical error in response parsing"}],
"positive_points": ["Essay submitted successfully"],
"suggestions": ["Please try submitting again"]
},
"content_relevance": {
"analysis": "Unable to parse detailed response. Basic content analysis completed.",
"issues": [{"type": "parsing_error", "before": "JSON parsing failed", "after": "Please try again", "explanation": "Technical error in response parsing"}],
"positive_points": ["Essay submitted successfully"],
"suggestions": ["Please try submitting again"]
},
"argument_development": {
"analysis": "Unable to parse detailed response. Basic argument analysis completed.",
"issues": [{"type": "parsing_error", "before": "JSON parsing failed", "after": "Please try again", "explanation": "Technical error in response parsing"}],
"positive_points": ["Essay submitted successfully"],
"suggestions": ["Please try submitting again"]
},
"evidence_citations": {
"analysis": "Unable to parse detailed response. Basic evidence analysis completed.",
"issues": [{"type": "parsing_error", "before": "JSON parsing failed", "after": "Please try again", "explanation": "Technical error in response parsing"}],
"positive_points": ["Essay submitted successfully"],
"suggestions": ["Please try submitting again"]
},
"structure_organization": {
"analysis": "Unable to parse detailed response. Basic organization analysis completed.",
"issues": [{"type": "parsing_error", "before": "JSON parsing failed", "after": "Please try again", "explanation": "Technical error in response parsing"}],
"positive_points": ["Essay submitted successfully"],
"suggestions": ["Please try submitting again"]
},
"conclusion_quality": {
"analysis": "Unable to parse detailed response. Basic conclusion analysis completed.",
"issues": [{"type": "parsing_error", "before": "JSON parsing failed", "after": "Please try again", "explanation": "Technical error in response parsing"}],
"positive_points": ["Essay submitted successfully"],
"suggestions": ["Please try submitting again"]
}
},
"essay_structure": {
"has_clear_introduction": {"value": True, "explanation": "Basic analysis completed"},
"has_structured_body": {"value": True, "explanation": "Basic analysis completed"},
"has_logical_conclusion": {"value": True, "explanation": "Basic analysis completed"},
"uses_transitions": {"value": True, "explanation": "Basic analysis completed"},
"maintains_tone": {"value": True, "explanation": "Basic analysis completed"}
},
"overall_feedback": "Technical error occurred during analysis. Please try submitting your essay again.",
"improvement_priorities": ["Try submitting again", "Check essay format", "Ensure proper text encoding"],
"error": "JSON parsing failed",
"original_error": str(e),
"cleaned_error": str(e2)
}
logger.warning("Returning fallback response due to JSON parsing failure")
return fallback_response
def grade_answer_with_gpt(self, student_answer, training_context):
logger.info(f"Processing full essay text: {self.count_tokens(student_answer)} tokens - NO TRUNCATION")
return self._grade_answer_legacy(student_answer, training_context)
def _grade_answer_legacy(self, student_answer, training_context):
original_text = student_answer
original_tokens = self.count_tokens(student_answer)
logger.info(f"Full essay token count: {original_tokens} - NO TRUNCATION")
student_answer, final_tokens, was_truncated = self.process_full_text(student_answer)
system_instructions = """
You are an expert English examiner specializing in CSS (Central Superior Services) essay evaluation. You MUST provide comprehensive feedback for EVERY aspect of the essay. Each feedback question/topic is MANDATORY.
EVALUATION CATEGORIES (ALL MANDATORY):
Grammar & Punctuation
Vocabulary Usage
Sentence Structure
Content Relevance & Depth
Argument Development
Evidence & Citations
Structure & Organization
Conclusion Quality
For each category, you MUST:
- SCAN THE ENTIRE ESSAY and identify EVERY instance where the writing does NOT meet CSS Essay standards for that category.
- For each issue, provide:
- before: the exact problematic sentence or phrase from the essay
- after: the improved or corrected version
- explanation: why it is an issue and how the correction improves it
- Return a COMPREHENSIVE LIST of ALL such issues found in the essay, not just a summary or a few examples.
- If no issues found, provide positive reinforcement with specific examples
- Focus on ISSUE DETECTION rather than scoring
VOCABULARY ANALYSIS REQUIREMENTS:
- Identify ALL problematic words (too simple, incorrect usage, misspellings)
- For each vocabulary issue, provide:
- before: the exact problematic word
- after: the corrected/improved word
- explanation: why the change is needed
- Suggest academic/sophisticated alternatives
- Check for repetitive word usage
ESSAY STRUCTURE ANALYSIS (ALL MANDATORY):
Evaluate each of these aspects with true/false and detailed explanations according to CSS Examiner standards:
1) Introduction & Thesis:
- Clear Thesis Statement: Is there a clear, well-defined thesis statement?
- Engaging Introduction: Does the introduction capture reader's attention?
- Background Context: Is there sufficient background context provided?
2) Body Development:
- Topic Sentences: Are there clear topic sentences for each paragraph?
- Supporting Evidence: Is each argument supported with evidence?
- Logical Flow: Do ideas flow logically from one to the next?
- Paragraph Coherence: Are paragraphs well-connected and coherent?
3) Content Quality:
- Relevance to Topic: Is all content relevant to the essay topic?
- Depth of Analysis: Does the essay provide deep, thorough analysis?
- Use of Examples: Are specific examples used to illustrate points?
- Critical Thinking: Does the essay demonstrate critical thinking?
4) Evidence & Citations:
- Factual Accuracy: Are all facts accurate and verifiable?
- Source Credibility: Are sources credible and authoritative?
- Proper Citations: Are sources properly cited?
- Statistical Data: Is statistical data used appropriately?
5) Conclusion:
- Summary of Arguments: Does the conclusion summarize main arguments?
- Policy Recommendations: Are policy recommendations provided?
- Future Implications: Are future implications discussed?
- Strong Closing: Does the essay have a strong, memorable closing?
TOPIC-SPECIFIC ANALYSIS:
Based on the essay topic, provide specialized feedback on:
- How well the essay addresses the specific question/topic
- Whether all aspects of the topic are covered
- If the essay demonstrates understanding of the topic's complexity
- Suggestions for better topic coverage
IMPORTANT REQUIREMENTS:
1. EVERY category MUST have feedback (no empty responses)
2. EVERY essay structure aspect MUST be evaluated
3. Provide specific examples from the essay
4. Give actionable improvement suggestions
5. Consider the CSS exam context and standards
6. Return ONLY valid JSON - no additional text
7. Ensure all feedback is constructive and educational
8. Use ONLY the exact JSON structure provided below
9. FOCUS ON FINDING AND DOCUMENTING ISSUES RATHER THAN SCORING
EXACT JSON FORMAT TO RETURN:
{
"categories": {
"grammar_punctuation": {
"analysis": "Detailed analysis of grammar and punctuation issues found in the essay",
"issues": [
{
"type": "grammar",
"before": "original text with error",
"after": "corrected text",
"explanation": "Explanation of the grammar rule violated"
},
{
"type": "grammar",
"before": "another error example",
"after": "corrected version",
"explanation": "Explanation for this correction"
}
],
"positive_points": ["Good point 1", "Good point 2"],
"suggestions": ["Suggestion 1", "Suggestion 2"]
},
"vocabulary_usage": {
"analysis": "Comprehensive analysis of vocabulary usage, word choice, and language sophistication",
"issues": [
{
"type": "vocabulary",
"before": "problematic_word",
"after": "improved_word",
"explanation": "Why this word needs improvement (too simple, incorrect usage, etc.)"
},
{
"type": "repetition",
"before": "repeated_word",
"after": "alternative_word",
"explanation": "Word is overused, suggest alternative"
}
],
"positive_points": ["Good vocabulary choices"],
"suggestions": ["Use more academic vocabulary", "Avoid repetition"]
},
"sentence_structure": {
"analysis": "Analysis of sentence variety, complexity, and structure. List ALL problematic sentences with before/after/explanation.",
"issues": [
{
"before": "This is a very long sentence it has no punctuation and is hard to read.",
"after": "This is a very long sentence. It has no punctuation and is hard to read.",
"explanation": "Split run-on sentence for clarity and punctuation."
},
{
"before": "He go to school every day.",
"after": "He goes to school every day.",
"explanation": "Subject-verb agreement error."
}
],
"positive_points": ["Good sentence variety"],
"suggestions": ["Vary sentence length", "Use complex sentences"]
},
"content_relevance": {
"analysis": "Analysis of how well content addresses the topic. List ALL irrelevant or off-topic content with before/after/explanation.",
"issues": [
{
"before": "The essay discusses unrelated historical events.",
"after": "Removed unrelated content.",
"explanation": "Content is not relevant to the essay topic."
},
{
"before": "Personal anecdotes not related to the topic.",
"after": "Removed personal anecdote.",
"explanation": "Personal stories are not relevant in a CSS essay unless directly related to the topic."
}
],
"positive_points": ["Content is relevant"],
"suggestions": ["Add more depth"]
},
"argument_development": {
"analysis": "Analysis of argument strength and logical flow. List ALL weak or missing arguments with before/after/explanation.",
"issues": [
{
"before": "The essay lacks a clear argument.",
"after": "Added a clear thesis statement and supporting arguments.",
"explanation": "CSS essays require a clear argument and logical development."
},
{
"before": "Arguments are not supported by evidence.",
"after": "Added supporting evidence for each argument.",
"explanation": "Arguments must be supported by evidence in a CSS essay."
}
],
"positive_points": ["Good arguments"],
"suggestions": ["Strengthen arguments"]
},
"evidence_citations": {
"analysis": "Analysis of evidence quality and citation usage. List ALL missing or weak evidence/citations with before/after/explanation.",
"issues": [
{
"before": "No sources are cited in the essay.",
"after": "Added citations for all factual claims.",
"explanation": "CSS essays require proper citation of evidence."
},
{
"before": "Uses vague evidence like 'many people say'.",
"after": "Replaced with specific, credible sources.",
"explanation": "Evidence must be specific and credible in a CSS essay."
}
],
"positive_points": ["Some evidence provided"],
"suggestions": ["Add more evidence"]
},
"structure_organization": {
"analysis": "Analysis of essay organization and structure. List ALL organizational issues with before/after/explanation.",
"issues": [
{
"before": "Paragraphs are not clearly separated.",
"after": "Added clear paragraph breaks.",
"explanation": "CSS essays require clear paragraph structure."
},
{
"before": "Ideas are presented in a random order.",
"after": "Reorganized ideas for logical flow.",
"explanation": "Ideas should be organized logically in a CSS essay."
}
],
"positive_points": ["Good organization"],
"suggestions": ["Improve transitions"]
},
"conclusion_quality": {
"analysis": "Analysis of conclusion effectiveness. List ALL issues with the conclusion with before/after/explanation.",
"issues": [
{
"before": "The essay ends abruptly without a conclusion.",
"after": "Added a clear, summarizing conclusion.",
"explanation": "CSS essays require a strong conclusion."
},
{
"before": "Conclusion repeats the introduction without adding value.",
"after": "Rewrote conclusion to synthesize main points and provide closure.",
"explanation": "Conclusion should synthesize, not repeat."
}
],
"positive_points": ["Good conclusion"],
"suggestions": ["Strengthen conclusion"]
}
},
"essay_structure": {
"Introduction & Thesis": {
"Clear Thesis Statement": {"value": true, "explanation": "Clear thesis statement present"},
"Engaging Introduction": {"value": true, "explanation": "Introduction captures reader's attention"},
"Background Context": {"value": true, "explanation": "Sufficient background context provided"}
},
"Body Development": {
"Topic Sentences": {"value": true, "explanation": "Clear topic sentences for each paragraph"},
"Supporting Evidence": {"value": true, "explanation": "Arguments supported with evidence"},
"Logical Flow": {"value": true, "explanation": "Ideas flow logically from one to the next"},
"Paragraph Coherence": {"value": true, "explanation": "Paragraphs well-connected and coherent"}
},
"Content Quality": {
"Relevance to Topic": {"value": true, "explanation": "All content relevant to essay topic"},
"Depth of Analysis": {"value": true, "explanation": "Essay provides deep, thorough analysis"},
"Use of Examples": {"value": true, "explanation": "Specific examples used to illustrate points"},
"Critical Thinking": {"value": true, "explanation": "Essay demonstrates critical thinking"}
},
"Evidence & Citations": {
"Factual Accuracy": {"value": true, "explanation": "All facts accurate and verifiable"},
"Source Credibility": {"value": true, "explanation": "Sources credible and authoritative"},
"Proper Citations": {"value": true, "explanation": "Sources properly cited"},
"Statistical Data": {"value": true, "explanation": "Statistical data used appropriately"}
},
"Conclusion": {
"Summary of Arguments": {"value": true, "explanation": "Conclusion summarizes main arguments"},
"Policy Recommendations": {"value": true, "explanation": "Policy recommendations provided"},
"Future Implications": {"value": true, "explanation": "Future implications discussed"},
"Strong Closing": {"value": true, "explanation": "Essay has strong, memorable closing"}
}
},
"overall_feedback": "Comprehensive overall feedback summarizing all aspects",
"improvement_priorities": ["Priority 1", "Priority 2", "Priority 3"]
}
"""
messages = [
{"role": "system", "content": system_instructions},
{"role": "user", "content": f"Student's Essay:\n\n{student_answer}"}
]
try:
response = self.client.chat.completions.create(
model="gpt-4.1",
messages=messages,
max_tokens=8000,
temperature=0,
)
feedback_raw = response.choices[0].message.content
feedback_dict = self.extract_json_from_text(feedback_raw)
# Transform to app format for compatibility
transformed_feedback = self.transform_feedback_to_app_format(feedback_dict)
return transformed_feedback
except Exception as e:
logger.error(f"Error in grade_answer_with_gpt: {str(e)}")
raise RuntimeError(f"Failed to grade answer using GPT: {str(e)}")
def grade_answer_with_question(self, student_answer, question):
logger.info(f"Processing full essay text for question: {self.count_tokens(student_answer)} tokens - NO TRUNCATION")
return self._grade_answer_with_question_legacy(student_answer, question)
def _grade_answer_with_question_legacy(self, student_answer, question):
original_text = student_answer
original_tokens = self.count_tokens(student_answer)
logger.info(f"Full essay token count: {original_tokens} - NO TRUNCATION")
student_answer, final_tokens, was_truncated = self.process_full_text(student_answer)
system_instructions = f"""
You are an expert English examiner specializing in CSS (Central Superior Services) essay evaluation.
You are evaluating an essay based on the specific question: '{question}'
EVALUATION CATEGORIES (ALL MANDATORY):
Grammar & Punctuation
Vocabulary Usage
Sentence Structure
Content Relevance & Depth
Argument Development
Evidence & Citations
Structure & Organization
Conclusion Quality
QUESTION-SPECIFIC ANALYSIS (MOST IMPORTANT):
Evaluate how well the essay addresses the question: '{question}'
- Does the essay directly answer the question?
- Are all aspects of the question covered?
- Is the response relevant and focused?
- Does the essay demonstrate understanding of the question's complexity?
For each category, you MUST:
- Provide a detailed analysis (minimum 2-3 sentences)
- List ALL issues found with specific examples
- For each issue, provide:
- before: the original text
- after: the corrected or improved text
- explanation: why the change is needed
- If no issues found, provide positive reinforcement with specific examples
- Focus on ISSUE DETECTION rather than scoring
VOCABULARY ANALYSIS REQUIREMENTS:
- Identify ALL problematic words (too simple, incorrect usage, misspellings)
- For each vocabulary issue, provide:
- before: the exact problematic word
- after: the corrected/improved word
- explanation: why the change is needed
- Suggest academic/sophisticated alternatives
- Check for repetitive word usage
ESSAY STRUCTURE ANALYSIS (ALL MANDATORY):
Evaluate each of these aspects with true/false and detailed explanations according to CSS Examiner standards:
1) Introduction & Thesis:
- Clear Thesis Statement: Is there a clear, well-defined thesis statement?
- Engaging Introduction: Does the introduction capture reader's attention?
- Background Context: Is there sufficient background context provided?
2) Body Development:
- Topic Sentences: Are there clear topic sentences for each paragraph?
- Supporting Evidence: Is each argument supported with evidence?
- Logical Flow: Do ideas flow logically from one to the next?
- Paragraph Coherence: Are paragraphs well-connected and coherent?
3) Content Quality:
- Relevance to Topic: Is all content relevant to the essay topic?
- Depth of Analysis: Does the essay provide deep, thorough analysis?
- Use of Examples: Are specific examples used to illustrate points?
- Critical Thinking: Does the essay demonstrate critical thinking?
4) Evidence & Citations:
- Factual Accuracy: Are all facts accurate and verifiable?
- Source Credibility: Are sources credible and authoritative?
- Proper Citations: Are sources properly cited?
- Statistical Data: Is statistical data used appropriately?
5) Conclusion:
- Summary of Arguments: Does the conclusion summarize main arguments?
- Policy Recommendations: Are policy recommendations provided?
- Future Implications: Are future implications discussed?
- Strong Closing: Does the essay have a strong, memorable closing?
QUESTION-SPECIFIC FEEDBACK:
Provide specialized feedback on how well the essay addresses the question: '{question}'
- Specific aspects of the question covered
- Missing aspects that should be addressed
- Suggestions for better question coverage
IMPORTANT REQUIREMENTS:
1. EVERY category MUST have feedback (no empty responses)
2. EVERY essay structure aspect MUST be evaluated
3. Provide specific examples from the essay
4. Give actionable improvement suggestions
5. Consider the CSS exam context and standards
6. Return ONLY valid JSON - no additional text
7. Ensure all feedback is constructive and educational
8. Focus heavily on how well the essay answers the specific question: '{question}'
9. Use ONLY the exact JSON structure provided below
10. FOCUS ON FINDING AND DOCUMENTING ISSUES RATHER THAN SCORING
EXACT JSON FORMAT TO RETURN:
{{
"categories": {{
"grammar_punctuation": {{
"analysis": "Detailed analysis of grammar and punctuation issues found in the essay",
"issues": [
{{
"type": "grammar",
"before": "original text with error",
"after": "corrected text",
"explanation": "Explanation of the grammar rule violated"
}}
],
"positive_points": ["Good point 1", "Good point 2"],
"suggestions": ["Suggestion 1", "Suggestion 2"]
}},
"vocabulary_usage": {{
"analysis": "Comprehensive analysis of vocabulary usage, word choice, and language sophistication",
"issues": [
{{
"type": "vocabulary",
"before": "problematic_word",
"after": "improved_word",
"explanation": "Why this word needs improvement (too simple, incorrect usage, etc.)"
}},
{{
"type": "repetition",
"before": "repeated_word",
"after": "alternative_word",
"explanation": "Word is overused, suggest alternative"
}}
],
"positive_points": ["Good vocabulary choices"],
"suggestions": ["Use more academic vocabulary", "Avoid repetition"]
}},
"sentence_structure": {{
"analysis": "Analysis of sentence variety, complexity, and structure",
"issues": [
{{
"type": "structure",
"before": "problematic sentence",
"after": "improved sentence",
"explanation": "Why this sentence structure needs improvement"
}}
],
"positive_points": ["Good sentence variety"],
"suggestions": ["Vary sentence length", "Use complex sentences"]
}},
"content_relevance": {{
"analysis": "Analysis of how well content addresses the topic",
"issues": [],
"positive_points": ["Content is relevant"],
"suggestions": ["Add more depth"]
}},
"argument_development": {{
"analysis": "Analysis of argument strength and logical flow",
"issues": [],
"positive_points": ["Good arguments"],
"suggestions": ["Strengthen arguments"]
}},
"evidence_citations": {{
"analysis": "Analysis of evidence quality and citation usage",
"issues": [],
"positive_points": ["Some evidence provided"],
"suggestions": ["Add more evidence"]
}},
"structure_organization": {{
"analysis": "Analysis of essay organization and structure",
"issues": [],
"positive_points": ["Good organization"],
"suggestions": ["Improve transitions"]
}},
"conclusion_quality": {{
"analysis": "Analysis of conclusion effectiveness",
"issues": [],
"positive_points": ["Good conclusion"],
"suggestions": ["Strengthen conclusion"]
}}
}},
"essay_structure": {{
"Introduction & Thesis": {{
"Clear Thesis Statement": {{"value": true, "explanation": "Clear thesis statement present"}},
"Engaging Introduction": {{"value": true, "explanation": "Introduction captures reader's attention"}},
"Background Context": {{"value": true, "explanation": "Sufficient background context provided"}}
}},
"Body Development": {{
"Topic Sentences": {{"value": true, "explanation": "Clear topic sentences for each paragraph"}},
"Supporting Evidence": {{"value": true, "explanation": "Arguments supported with evidence"}},
"Logical Flow": {{"value": true, "explanation": "Ideas flow logically from one to the next"}},
"Paragraph Coherence": {{"value": true, "explanation": "Paragraphs well-connected and coherent"}}
}},
"Content Quality": {{
"Relevance to Topic": {{"value": true, "explanation": "All content relevant to essay topic"}},
"Depth of Analysis": {{"value": true, "explanation": "Essay provides deep, thorough analysis"}},
"Use of Examples": {{"value": true, "explanation": "Specific examples used to illustrate points"}},
"Critical Thinking": {{"value": true, "explanation": "Essay demonstrates critical thinking"}}
}},
"Evidence & Citations": {{
"Factual Accuracy": {{"value": true, "explanation": "All facts accurate and verifiable"}},
"Source Credibility": {{"value": true, "explanation": "Sources credible and authoritative"}},
"Proper Citations": {{"value": true, "explanation": "Sources properly cited"}},
"Statistical Data": {{"value": true, "explanation": "Statistical data used appropriately"}}
}},
"Conclusion": {{
"Summary of Arguments": {{"value": true, "explanation": "Conclusion summarizes main arguments"}},
"Policy Recommendations": {{"value": true, "explanation": "Policy recommendations provided"}},
"Future Implications": {{"value": true, "explanation": "Future implications discussed"}},
"Strong Closing": {{"value": true, "explanation": "Essay has strong, memorable closing"}}
}}
}},
"overall_feedback": "Comprehensive overall feedback summarizing all aspects",
"improvement_priorities": ["Priority 1", "Priority 2", "Priority 3"],
"question_specific_feedback": {{
"question": "{question}",
"question_coverage": "Analysis of how well the essay addresses the specific question",
"covered_aspects": ["Aspect 1", "Aspect 2"],
"missing_aspects": ["Missing aspect 1", "Missing aspect 2"],
"strengths": ["Strength 1", "Strength 2"],
"improvement_suggestions": ["Suggestion 1", "Suggestion 2"]
}}
}}
"""
messages = [
{"role": "system", "content": system_instructions},
{"role": "user", "content": f"Question: {question}\n\nStudent's Essay:\n\n{student_answer}"}
]
try:
response = self.client.chat.completions.create(
model="gpt-4.1",
messages=messages,
max_tokens=8000,
temperature=0,
)
feedback_raw = response.choices[0].message.content
feedback_dict = self.extract_json_from_text(feedback_raw)
# Transform to app format for compatibility
transformed_feedback = self.transform_feedback_to_app_format(feedback_dict)
return transformed_feedback
except Exception as e:
logger.error(f"Error in grade_answer_with_question: {str(e)}")
raise RuntimeError(f"Failed to grade answer using GPT: {str(e)}")
def analyze_grammar_only(self, text: str) -> Dict[str, Any]:
logger.info("Starting grammar-only analysis")
if not text.strip():
return {'line_by_line_grammar': [], 'overall_grammar_summary': {'error': 'No text provided'}}
text = self.process_full_text(text)[0]
lines = text.split('\n')
all_line_grammar = []
for line_index, line in enumerate(lines):
if not line.strip():
all_line_grammar.append({
'line_number': line_index + 1,
'line_content': line,
'line_type': 'empty',
'grammar_score': 100,
'grammar_issues': [],
'positive_points': ['Proper line spacing'],
'suggestions': []
})
continue
try:
line_grammar = self._analyze_line_grammar_only(line, line_index + 1)
all_line_grammar.append(line_grammar)
except Exception as e:
logger.error(f"Error analyzing line {line_index + 1} for grammar: {str(e)}")
all_line_grammar.append({
'line_number': line_index + 1,
'line_content': line,
'line_type': 'error',
'grammar_score': 0,
'grammar_issues': [{'type': 'processing_error', 'description': str(e)}],
'positive_points': [],
'suggestions': ['Please review this line manually']
})
return {'line_by_line_grammar': all_line_grammar}
def _analyze_line_grammar_only(self, line: str, line_number: int) -> Dict[str, Any]:
system_prompt = f"""You are an expert English grammar examiner. Analyze this single line of text for GRAMMAR AND PUNCTUATION issues ONLY.\n\n...\nReturn JSON format:\n{{ ... }}"""
messages = [
{"role": "system", "content": system_prompt},
{"role": "user", "content": f"Line {line_number}: {line}"}
]
try:
response = self.client.chat.completions.create(
model="gpt-4.1",
messages=messages,
max_tokens=8000,
temperature=0.3,
)
feedback_raw = response.choices[0].message.content
feedback_dict = self.extract_json_from_text(feedback_raw)
feedback_dict['line_number'] = line_number
feedback_dict['line_content'] = line
return feedback_dict
except Exception as e:
logger.error(f"Error in grammar-only line analysis: {str(e)}")
return {
'line_number': line_number,
'line_content': line,
'line_type': 'error',
'grammar_score': 0,
'grammar_issues': [{'type': 'processing_error', 'description': str(e)}],
'positive_points': [],
'suggestions': ['Please review this line manually']
}
def transform_feedback_to_app_format(self, feedback_dict):
"""
Transform the detailed feedback format to the format expected by the app.
This ensures compatibility with existing API responses and focuses on showing issues.
"""
try:
# Check if we have the new detailed format
if "categories" in feedback_dict:
# Transform from new detailed format to app format
evaluation_and_scoring = []
# Map category names to app format
category_mapping = {
"grammar_punctuation": "Grammar & Punctuation",
"vocabulary_usage": "Vocabulary Usage",
"sentence_structure": "Sentence Structure",
"content_relevance": "Content Relevance & Depth",
"argument_development": "Argument Development",
"evidence_citations": "Evidence & Citations",
"structure_organization": "Structure & Organization",
"conclusion_quality": "Conclusion Quality"
}
for category_key, category_name in category_mapping.items():
if category_key in feedback_dict["categories"]:
category_data = feedback_dict["categories"][category_key]
# Transform issues to app format with detailed information
issues_list = []
for issue in category_data.get("issues", []):
issue_info = {
"before": issue.get("before", ""),
"after": issue.get("after", ""),
"explanation": issue.get("explanation", "")
}
issues_list.append(issue_info)
# Create the evaluation and scoring entry (focus on issues, not scores)
evaluation_and_scoring.append({
"label": category_name,
"analysis": category_data.get("analysis", f"{category_name} analysis completed"),
"issuesCount": len(issues_list),
"issuesList": issues_list,
"positivePoints": category_data.get("positive_points", [])
})
# Transform essay structure to match the desired format with new CSS Examiner criteria
essay_structure = []
if "essay_structure" in feedback_dict:
structure_data = feedback_dict["essay_structure"]
# Introduction & Thesis section
intro_features = []
if 'Introduction & Thesis' in structure_data:
intro_data = structure_data['Introduction & Thesis']
for key, value in intro_data.items():
is_correct = value.get('value', True)
explanation = value.get('explanation', '')
error_message = f"Missing: {key.lower()}. {explanation}" if not is_correct else None
intro_features.append({
"label": key,
"isCorrect": is_correct,
"errorMessage": error_message
})
essay_structure.append({
"label": "Introduction & Thesis",
"features": intro_features
})
# Body Development section
body_features = []
if 'Body Development' in structure_data:
body_data = structure_data['Body Development']
for key, value in body_data.items():
is_correct = value.get('value', True)
explanation = value.get('explanation', '')
error_message = f"Missing: {key.lower()}. {explanation}" if not is_correct else None
body_features.append({
"label": key,
"isCorrect": is_correct,
"errorMessage": error_message
})
essay_structure.append({
"label": "Body Development",
"features": body_features
})
# Content Quality section
content_features = []
if 'Content Quality' in structure_data:
content_data = structure_data['Content Quality']
for key, value in content_data.items():
is_correct = value.get('value', True)
explanation = value.get('explanation', '')
error_message = f"Missing: {key.lower()}. {explanation}" if not is_correct else None
content_features.append({
"label": key,
"isCorrect": is_correct,
"errorMessage": error_message
})
essay_structure.append({
"label": "Content Quality",
"features": content_features
})
# Evidence & Citations section
evidence_features = []
if 'Evidence & Citations' in structure_data:
evidence_data = structure_data['Evidence & Citations']
for key, value in evidence_data.items():
is_correct = value.get('value', True)
explanation = value.get('explanation', '')
error_message = f"Missing: {key.lower()}. {explanation}" if not is_correct else None
evidence_features.append({
"label": key,
"isCorrect": is_correct,
"errorMessage": error_message
})
essay_structure.append({
"label": "Evidence & Citations",
"features": evidence_features
})
# Conclusion section
conclusion_features = []
if 'Conclusion' in structure_data:
conclusion_data = structure_data['Conclusion']
for key, value in conclusion_data.items():
is_correct = value.get('value', True)
explanation = value.get('explanation', '')
error_message = f"Missing: {key.lower()}. {explanation}" if not is_correct else None
conclusion_features.append({
"label": key,
"isCorrect": is_correct,
"errorMessage": error_message
})
essay_structure.append({
"label": "Conclusion",
"features": conclusion_features
})
# Create the transformed response with focus on issues
transformed_response = {
"evaluationAndScoring": evaluation_and_scoring,
"essayStructure": essay_structure,
"overall_feedback": feedback_dict.get("overall_feedback", "Comprehensive analysis completed"),
"improvement_priorities": feedback_dict.get("improvement_priorities", []),
"total_issues_found": sum(len(section.get("issuesList", [])) for section in evaluation_and_scoring),
"vocabulary_issues": [
issue for section in evaluation_and_scoring
if section["label"] == "Vocabulary Usage"
for issue in section.get("issuesList", [])
],
"grammar_issues": [
issue for section in evaluation_and_scoring
if section["label"] == "Grammar & Punctuation"
for issue in section.get("issuesList", [])
]
}
# Add question-specific feedback if present
if "question_specific_feedback" in feedback_dict:
transformed_response["question_specific_feedback"] = feedback_dict["question_specific_feedback"]
return transformed_response
else:
# Already in app format, return as is
return feedback_dict
except Exception as e:
logger.error(f"Error transforming feedback format: {str(e)}")
# Return fallback format
return {
"evaluationAndScoring": [
{
"label": "Grammar & Punctuation",
"analysis": "Basic analysis completed",
"issuesCount": 0,
"issuesList": [],
"positivePoints": ["Essay submitted successfully"]
}
],
"essayStructure": [
{
"label": "Introduction & Thesis",
"features": [
{
"label": "Clear Thesis Statement",
"isCorrect": False,
"errorMessage": "Missing: clear thesis statement. The essay lacks a clear, well-defined thesis statement that guides the reader."
}
]
}
],
"overall_feedback": "Analysis completed with basic feedback",
"improvement_priorities": ["Try submitting again"]
}
def rephrase_text_with_gpt(self, essay_text: str, system_prompt: str = None) -> dict:
"""
Rephrase and correct the essay to meet CSS (Central Superior Services) standards.
Provides comprehensive corrections for grammar, structure, style, and content.
"""
if system_prompt is None:
system_prompt = """You are an expert CSS (Central Superior Services) essay examiner and editor. Your task is to provide the BEST VERSION of the given essay by making comprehensive improvements while maintaining the original meaning and intent.
IMPORTANT REQUIREMENTS:
1. CORRECT ALL GRAMMAR AND PUNCTUATION ERRORS
2. IMPROVE SENTENCE STRUCTURE AND FLOW
3. ENHANCE VOCABULARY USAGE WITH APPROPRIATE ACADEMIC LANGUAGE
4. STRENGTHEN ARGUMENT DEVELOPMENT AND LOGICAL FLOW
5. IMPROVE ESSAY STRUCTURE (Introduction, Body, Conclusion)
6. ADD TRANSITIONAL PHRASES FOR BETTER COHERENCE
7. ENSURE PROPER PARAGRAPH ORGANIZATION
8. MAINTAIN CSS EXAM STANDARDS AND EXPECTATIONS
9. KEEP THE ORIGINAL MEANING AND ARGUMENTS INTACT
10. USE FORMAL ACADEMIC TONE APPROPRIATE FOR CSS EXAMS
CORRECTION GUIDELINES:
- Fix all grammatical errors (subject-verb agreement, tense consistency, etc.)
- Correct punctuation (commas, semicolons, apostrophes, etc.)
- Improve sentence variety and complexity
- Enhance vocabulary with sophisticated academic terms
- Strengthen topic sentences and supporting evidence
- Add logical transitions between paragraphs
- Ensure clear thesis statement and conclusion
- Maintain professional tone throughout
- Follow CSS essay format and style requirements
Return ONLY the corrected essay text - no explanations, no markdown formatting, just the improved essay ready for CSS examination."""
try:
messages = [
{"role": "system", "content": system_prompt},
{"role": "user", "content": f"Please provide the BEST VERSION of this CSS essay with all corrections applied:\n\n{essay_text}"}
]
# Calculate appropriate max_tokens based on input length
input_tokens = self.count_tokens(essay_text)
# Allow for 2x the input length plus extra for corrections
max_tokens_needed = min(input_tokens * 2 + 2000, 16000) # Cap at 16k tokens
response = self.client.chat.completions.create(
model="gpt-4.1",
messages=messages,
max_tokens=max_tokens_needed, # Dynamic token limit
temperature=0.3, # Lower temperature for more consistent corrections
)
rephrased_text = response.choices[0].message.content.strip()
# Clean up any potential formatting artifacts
rephrased_text = rephrased_text.replace('```', '').replace('**', '').replace('*', '')
rephrased_text = rephrased_text.strip()
# Verify that we didn't lose significant content
original_words = len(essay_text.split())
rephrased_words = len(rephrased_text.split())
if rephrased_words < original_words * 0.7: # If we lost more than 30% of content
logger.warning(f"Significant content loss detected: {original_words} -> {rephrased_words} words")
# Return original text with a note
return {
"rephrased_text": essay_text,
"error": f"Content loss detected ({original_words} -> {rephrased_words} words). Returning original text.",
"warning": "Rephrasing may have truncated content"
}
logger.info(f"Rephrasing successful: {original_words} -> {rephrased_words} words")
return {"rephrased_text": rephrased_text, "error": None}
except Exception as e:
logger.error(f"Error in rephrasing essay: {str(e)}")
return {"rephrased_text": essay_text, "error": str(e)} |