utils.py

# utils.py
from fuzzywuzzy import fuzz
import re

def extract_keyword(text: str, symptoms: list = None) -> str:
    """
    Extracts a primary keyword from a given text, prioritizing symptom matches if available,
    otherwise, the first relevant word.
    Handles both "Question: " prefixed strings and direct symptom strings.
    """
    if text.startswith("Question: "):
        # Remove "Question: " prefix and process question text
        question = text[10:].strip()
        words = question.split()
        if not words:
            return "Unknown"

        # Common words to skip - expanded list
        common_words = {
            'what', 'is', 'are', 'how', 'why', 'can', 'do', 'does', 'i', 'have', 'my', 'a', 'an', 'the',
            'in', 'of', 'and', 'or', 'for', 'with', 'from', 'about', 'some', 'any', 'this', 'that',
            'there', 'be', 'to', 'me', 'am', 'feel', 'feeling', 'experiencing', 'symptoms', 'issue',
            'problem', 'cause', 'causes', 'tell', 'me', 'more', 'information', 'on', 'about', 'a',
            'an', 'the', 'my', 'your', 'its', 'their', 'our', 'his', 'her', 'its', 'them', 'us', 'you',
            'i', 'we', 'he', 'she', 'it', 'they', 'this', 'that', 'these', 'those', 'which', 'who',
            'whom', 'whose', 'where', 'when', 'why', 'how', 'what', 'if', 'then', 'else', 'or', 'and',
            'but', 'because', 'as', 'until', 'while', 'of', 'at', 'by', 'for', 'with', 'about', 'against',
            'between', 'into', 'through', 'during', 'before', 'after', 'above', 'below', 'to', 'from',
            'up', 'down', 'in', 'out', 'on', 'off', 'over', 'under', 'again', 'further', 'then', 'once',
            'here', 'there', 'when', 'where', 'why', 'all', 'any', 'both', 'each', 'few', 'more', 'most',
            'other', 'some', 'such', 'no', 'nor', 'not', 'only', 'own', 'same', 'so', 'than', 'too',
            'very', 's', 't', 'can', 'will', 'just', 'don', 'should', 'now'
        }

        # Fuzzy matching against symptoms (if provided)
        if symptoms:
            best_match = None
            highest_score = 0
            # Prioritize multi-word symptoms if they match well
            for symptom_phrase in sorted(symptoms, key=len, reverse=True):
                for i in range(len(words)):
                    for j in range(i + 1, len(words) + 1):
                        phrase_from_question = " ".join(words[i:j]).lower()
                        score = fuzz.token_sort_ratio(phrase_from_question, symptom_phrase.lower())
                        if score > 85 and score > highest_score:
                            best_match = symptom_phrase
                            highest_score = score
            if best_match:
                return best_match.capitalize()

        # Fallback: pick the first non-common word longer than 2 characters
        for word in words:
            word_lower = word.lower()
            if word_lower not in common_words and len(word_lower) > 2:
                if not re.match(r'^\d+$', word_lower) and not re.match(r'^\w$', word_lower):
                    return word.capitalize()
        
        # Last resort: if no good keyword, take the first non-common word
        for word in words:
            word_lower = word.lower()
            if word_lower not in common_words:
                return word.capitalize()
        
        return words[0].capitalize() if words else "Unknown"
    else:
        # For symptom checker inputs (comma-separated symptoms)
        # The history stores the raw selected symptoms here, so we just return the first one or the full list if short
        symptom_list_str = text.strip()
        if symptom_list_str:
            # If it's a short list of symptoms, return the whole thing
            if len(symptom_list_str.split(',')) <= 3:
                return symptom_list_str.capitalize()
            else: # Otherwise, just the first symptom
                first_symptom = symptom_list_str.split(',')[0].strip()
                return first_symptom.capitalize()
        return "Unknown"