b4add2762c
- import_batch.py: 批量处理多个Notion导出zip·追加模式·UUID去重 - import_notion.py: 单次导入·目录标题匹配·批量插入 - brain.db本地已建(100MB): 15,451页·4,053入口 · 冰朔光湖世界(9,324) + 人格记忆总索引(135) + 光湖世界入口导航(193) + 曜冥纪元(5,799)
104 lines
4.0 KiB
Python
104 lines
4.0 KiB
Python
#!/usr/bin/env python3
|
|
"""Notion Full Export → SQLite brain.db importer"""
|
|
import sqlite3, os, re
|
|
|
|
DB = "/Users/bingshuolingdianyuanhe/WorkBuddy/2026-06-04-13-44-41/persona-brain-db/brain.db"
|
|
CONTENT = "/tmp/full-export/export"
|
|
|
|
def parse_md(filepath):
|
|
fn = os.path.basename(filepath).replace('.md', '')
|
|
uuid = None
|
|
title = fn
|
|
m = re.search(r'([a-f0-9]{32})$', fn)
|
|
if m:
|
|
uuid = m.group(1)
|
|
title = fn[:fn.rindex(uuid)].strip().rstrip(' ·').strip()
|
|
with open(filepath, 'r', encoding='utf-8', errors='replace') as f:
|
|
raw = f.read()[:12000]
|
|
return {"title": title, "uuid": uuid, "content": raw}
|
|
|
|
def main():
|
|
conn = sqlite3.connect(DB)
|
|
c = conn.cursor()
|
|
c.execute("DELETE FROM documents")
|
|
conn.commit()
|
|
|
|
# 第一遍: 建立目录→UUID映射
|
|
# Notion export: page .md filename = "Title UUID.md", child directory = "Title/"
|
|
# Directory name may NOT include UUID, only the title
|
|
dir_to_uuid = {}
|
|
# Build a mapping from title-only to uuid
|
|
title_to_uuid = {}
|
|
for walk_root, dirs, files in os.walk(CONTENT):
|
|
for f in files:
|
|
if not f.endswith('.md'): continue
|
|
fpath = os.path.join(walk_root, f)
|
|
page = parse_md(fpath)
|
|
if not page['uuid'] or not page['title']: continue
|
|
# Map this page's uuid to any subdirectory whose basename matches the title
|
|
# Also check the full filename (without .md) as fallback
|
|
title_to_uuid[page['title'].lower().strip()] = page['uuid']
|
|
|
|
# Second pass on dirs: match dir names to page titles
|
|
for walk_root, dirs, files in os.walk(CONTENT):
|
|
for d in dirs:
|
|
dpath = os.path.join(walk_root, d)
|
|
dname = d.lower().strip()
|
|
if dname in title_to_uuid:
|
|
dir_to_uuid[dpath] = title_to_uuid[dname]
|
|
else:
|
|
# Try prefix match: directory name might be truncated
|
|
for title, uuid in title_to_uuid.items():
|
|
if title.startswith(dname) or dname.startswith(title):
|
|
if len(dname) > 3 and len(title) > 3:
|
|
dir_to_uuid[dpath] = uuid
|
|
break
|
|
|
|
# 第二遍: 导入所有页面
|
|
pages = []
|
|
for walk_root, dirs, files in sorted(os.walk(CONTENT)):
|
|
for f in sorted(files):
|
|
if not f.endswith('.md'): continue
|
|
fpath = os.path.join(walk_root, f)
|
|
page = parse_md(fpath)
|
|
if not page['uuid']: continue
|
|
rel = os.path.relpath(fpath, CONTENT)
|
|
depth = max(0, rel.count('/'))
|
|
parent_id = dir_to_uuid.get(walk_root)
|
|
pages.append({**page, "depth": depth, "parent_id": parent_id, "path": rel})
|
|
|
|
# 批量插入
|
|
c.execute("BEGIN TRANSACTION")
|
|
for i, p in enumerate(pages):
|
|
c.execute("""
|
|
INSERT OR REPLACE INTO documents (document_id, title, content, parent_id, path, content_type)
|
|
VALUES (?, ?, ?, ?, ?, 'notion')
|
|
""", (p['uuid'], p['title'], p['content'], p['parent_id'], p['path']))
|
|
if (i+1) % 1000 == 0:
|
|
conn.commit()
|
|
c.execute("BEGIN TRANSACTION")
|
|
print(f" {i+1}/{len(pages)}...")
|
|
conn.commit()
|
|
|
|
# 统计
|
|
c.execute("SELECT count(*) FROM documents")
|
|
total = c.fetchone()[0]
|
|
c.execute("SELECT count(*) FROM documents WHERE parent_id IS NULL")
|
|
roots = c.fetchone()[0]
|
|
c.execute("SELECT count(DISTINCT parent_id) FROM documents WHERE parent_id IS NOT NULL")
|
|
parents = c.fetchone()[0]
|
|
|
|
print(f"\n导入完成: {total}页 ({roots}根页面, {parents}父节点)")
|
|
|
|
c.execute("SELECT document_id, title FROM documents WHERE parent_id IS NULL ORDER BY title")
|
|
for row in c.fetchall()[:30]:
|
|
print(f" [{row[0][:8]}..] {row[1][:70]}")
|
|
|
|
c.execute("SELECT count(*) FROM sqlite_master WHERE type='table'")
|
|
tables = c.fetchone()[0]
|
|
conn.close()
|
|
print(f"Done. {tables} tables total.")
|
|
|
|
if __name__ == '__main__':
|
|
main()
|