script that can create json file from the raw files and vice versa

manage_dataset_structure.py

@@ -0,0 +1,128 @@
+import json
+from pathlib import Path
+import re
+
+JSON_TITLE = 'ruwikibench_articles.json'
+
+def create_json():
+    # CREATES JSON FROM DATASET
+    base_dir = Path('.')
+    articles_list = base_dir / 'small_articles_data.txt'
+    html_dir = base_dir / 'Html'
+    sources_dir = base_dir / 'Sources'
+    sources_list_dir = base_dir / 'Downloaded_Sources_List'
+    output_file = base_dir / JSON_TITLE
+
+    with open(articles_list, 'r', encoding='utf-8') as f:
+        article_titles = [line.strip() for line in f if line.strip()]
+
+    dataset = []
+
+    for title in sorted(article_titles):
+        true_title = title
+        title = re.sub(r'[<>:"/\\|?*]', '', title)
+        entry = {'article_name': true_title, 'article_cleared_name': title}
+
+        html_file = html_dir / f'{title}.html'
+        with open(html_file, 'r', encoding='utf-8') as f:
+            entry['html'] = f.read()
+
+        sources_list_file = sources_list_dir / f'{title}.json'.replace(' ', '_')
+        with open(sources_list_file, 'r', encoding='utf-8') as f:
+            source_ids = json.load(f)
+        print(f'Number of links for article {title}: {len(source_ids)}')
+        article_sources_dir = sources_dir / title
+        entry['sources'] = []
+
+        source_files = sorted(
+            article_sources_dir.glob('source_*.txt'),
+            key=lambda x: int(x.stem.split('_')[1])
+        )
+
+        for idx, (source_id, source_file) in enumerate(zip(source_ids, source_files)):
+            with open(source_file, 'r', encoding='utf-8') as f:
+                source_text = f.read()
+
+            entry['sources'].append({
+                'source_id': source_id,
+                'source_text': source_text,
+                'file': source_file.name
+            })
+
+        dataset.append(entry)
+
+        print(f'Processed: {title}')
+
+    with open(output_file, 'w', encoding='utf-8') as f:
+        json.dump(dataset, f, ensure_ascii=False, indent=2)
+
+    print(f'Articles processed: {len(dataset)}')
+
+def load_json():
+    # CHECKS JSON DATA
+    with open(JSON_TITLE, 'r', encoding='utf-8') as f:
+        data = json.load(f)
+
+    print(len(data))
+    i = 17
+    print(data[i]['article_name'])
+    print(data[i]['article_cleared_name'])
+    print(data[i]['html'][:100])
+    print(data[i]['sources'][10]['source_id'])
+    print(data[i]['sources'][10]['source_text'][:100])
+    print(data[i]['sources'][10]['file'])
+
+def decompose_json():
+    # DECOMPOSES JSON TO NORMAL FILE SYSTEM
+    base_dir = Path('.')
+    json_file = base_dir / JSON_TITLE
+
+    html_dir = base_dir / 'Html'
+    sources_dir = base_dir / 'Sources'
+    sources_list_dir = base_dir / 'Downloaded_Sources_List'
+
+    html_dir.mkdir(exist_ok=True)
+    sources_dir.mkdir(exist_ok=True)
+    sources_list_dir.mkdir(exist_ok=True)
+
+    with open(json_file, 'r', encoding='utf-8') as f:
+        dataset = json.load(f)
+
+    article_titles = []
+
+    for article in dataset:
+        true_title = article['article_name']
+        title = article['article_cleared_name']
+        article_titles.append(true_title)
+
+        html_file = html_dir / f'{title}.html'
+        with open(html_file, 'w', encoding='utf-8') as f:
+            f.write(article['html'])
+
+        source_ids = [source['source_id'] for source in article['sources']]
+        sources_list_file = sources_list_dir / f'{title}.json'.replace(' ', '_')
+        with open(sources_list_file, 'w', encoding='utf-8') as f:
+            json.dump(source_ids, f, ensure_ascii=False, indent=2)
+
+        article_sources_dir = sources_dir / title
+        article_sources_dir.mkdir(exist_ok=True)
+
+        for idx, source in enumerate(article['sources'], 1):
+            source_file = article_sources_dir / source['file']
+            with open(source_file, 'w', encoding='utf-8') as f:
+                f.write(source['source_text'])
+
+        print(f'Reinstalled: {title}')
+
+    articles_list = base_dir / 'small_articles_data.txt'
+    with open(articles_list, 'w', encoding='utf-8') as f:
+        f.write('\n'.join(article_titles))
+
+    print(f'\nDone!')
+
+
+#create_json()
+
+#load_json()
+
+#decompose_json()