convert_dataset.py

import json
import os

def process_standard_structure(data, filename):
    """
    מעבד קבצים עם מבנה סטנדרטי (תנ"ך, ש"ס, ירושלמי, משנה).
    מבנה: קטגוריה -> תת-קטגוריה -> ספרים -> {שם_ספר: {pages: מספר}}
    """
    records = []
    main_category = data.get('name')
    content_type = data.get('content_type')

    for subcategory in data.get('subcategories', []):
        subcategory_name = subcategory.get('name')
        for book_name, book_info in subcategory.get('books', {}).items():
            record = {
                "source_file": filename,
                "main_category": main_category,
                "subcategory": subcategory_name,
                "book": book_name,
                "content_type": content_type,
                "count": book_info.get('pages')
            }
            records.append(record)
    return records

def process_rambam(data, filename):
    """
    מעבד את קובץ הרמב"ם, השומר על החלוקה הפנימית ל'חלקים' (parts).
    """
    records = []
    main_category = data.get('name')
    content_type = data.get('content_type')

    for subcategory in data.get('subcategories', []):
        subcategory_name = subcategory.get('name')
        for book_name, book_info in subcategory.get('books', {}).items():
            parts = book_info.get('parts', [])
            # חישוב סך הפרקים מכל החלקים
            total_chapters = sum(part.get('end', 0) - part.get('start', 0) + 1 for part in parts)
            
            record = {
                "source_file": filename,
                "main_category": main_category,
                "subcategory": subcategory_name,
                "book": book_name,
                "content_type": content_type,
                "count": total_chapters,
                "parts": parts  # שימור המבנה המפורט המקורי
            }
            records.append(record)
    return records

def process_halakha(data, filename):
    """
    מעבד את קובץ ההלכה, המכיל מבנים שונים (טור/שו"ע, משנה ברורה).
    """
    records = []
    main_category = data.get('name')

    for subcategory in data.get('subcategories', []):
        subcategory_name = subcategory.get('name')
        content_type = subcategory.get('content_type')

        for book_name, book_info in subcategory.get('books', {}).items():
            record = {
                "source_file": filename,
                "main_category": main_category,
                "subcategory": subcategory_name,
                "book": book_name,
                "content_type": content_type
            }

            # טיפול במבנה של משנה ברורה / ביאור הלכה
            if 'parts' in book_info:
                parts = book_info.get('parts', [])
                # חישוב סך הסימנים מכל החלקים
                total_simanim = sum(part.get('end', 0) - part.get('start', 0) + 1 for part in parts)
                record['count'] = total_simanim
                record['parts'] = parts # שימור המבנה המפורט
            
            # טיפול במבנה של טור / שו"ע
            elif 'pages' in book_info:
                record['count'] = book_info.get('pages')

            # הוספת החרגות אם קיימות
            if 'exclude' in book_info:
                record['excluded_units'] = book_info.get('exclude')
            
            records.append(record)
            
    return records


def main():
    """
    פונקציה ראשית שמנהלת את תהליך ההמרה.
    """
    # מיפוי שם קובץ לפונקציית העיבוד המתאימה
    file_processors = {
        'shas.json': process_standard_structure,
        'yerushalmi.json': process_standard_structure,
        'mishna.json': process_standard_structure,
        'tanach.json': process_standard_structure,
        'rambam.json': process_rambam,
        'halakha.json': process_halakha
    }

    all_records = []
    input_files = list(file_processors.keys())

    for filename in input_files:
        if not os.path.exists(filename):
            print(f"Warning: File '{filename}' not found. Skipping.")
            continue
        
        try:
            with open(filename, 'r', encoding='utf-8') as f:
                data = json.load(f)
            
            processor_func = file_processors[filename]
            records = processor_func(data, filename)
            all_records.extend(records)
            print(f"Processed '{filename}', added {len(records)} records.")

        except json.JSONDecodeError:
            print(f"Error: Could not decode JSON from '{filename}'.")
        except Exception as e:
            print(f"An unexpected error occurred while processing '{filename}': {e}")

    output_filename = 'judaic_texts_dataset.jsonl'
    try:
        with open(output_filename, 'w', encoding='utf-8') as f_out:
            for record in all_records:
                # שימוש ב-ensure_ascii=False כדי לשמור על עברית תקינה
                f_out.write(json.dumps(record, ensure_ascii=False) + '\n')
        
        print(f"\nSuccessfully created dataset file: '{output_filename}' with {len(all_records)} total records.")

    except Exception as e:
        print(f"An error occurred while writing the output file: {e}")


if __name__ == '__main__':
    main()