experiemental: youtube comment fs

comment_fs/create_comment_tree.py

@@ -0,0 +1,37 @@
+#!/usr/bin/env python3
+import json
+from collections import defaultdict
+
+def build_comment_tree(comments):
+    tree = defaultdict(list)
+    root_comments = []
+
+    print(f"Total comments: {len(comments)}")  # Debug info
+
+    for comment in comments:
+        if comment['parent'] == "root":
+            root_comments.append(comment)
+        else:
+            tree[comment['parent']].append(comment)
+
+    print(f"Root comments: {len(root_comments)}")  # Debug info
+
+    def build_subtree(comment):
+        return {
+            #"id": comment['id'],
+            "text": comment['text'],
+            "replies": [build_subtree(reply) for reply in tree[comment['id']]]
+        }
+
+    return [build_subtree(comment) for comment in root_comments]
+
+with open('comments.jsonl', 'r', encoding='utf-8') as f:
+    comments = [json.loads(line) for line in f]
+
+comment_tree = build_comment_tree(comments)
+
+print(f"Final tree length: {len(comment_tree)}")  # Debug info
+
+with open('comment_tree.json', 'w') as f:
+    json.dump(comment_tree, f, ensure_ascii=False, indent=2)
+