From 4a1628a4f94b0a3c720295b8d534c3562f1161c6 Mon Sep 17 00:00:00 2001
From: songtianlun <tianlun.song@foxmail.com>
Date: Wed, 4 Sep 2024 16:20:19 +0800
Subject: [PATCH] add migrate blog

---
 migrate-blog/.python-version                  |   1 +
 migrate-blog/fix_hexo_md_pre_frontmatter.py   |  20 +++
 .../migrate_from_hexo_to_wordpress.py         | 131 +++++++++++++++++
 .../migrate_from_hexo_to_wordpress_2.py       | 133 ++++++++++++++++++
 4 files changed, 285 insertions(+)
 create mode 100644 migrate-blog/.python-version
 create mode 100644 migrate-blog/fix_hexo_md_pre_frontmatter.py
 create mode 100644 migrate-blog/migrate_from_hexo_to_wordpress.py
 create mode 100644 migrate-blog/migrate_from_hexo_to_wordpress_2.py

diff --git a/migrate-blog/.python-version b/migrate-blog/.python-version
new file mode 100644
index 0000000..78c9a28
--- /dev/null
+++ b/migrate-blog/.python-version
@@ -0,0 +1 @@
+3.9.12
diff --git a/migrate-blog/fix_hexo_md_pre_frontmatter.py b/migrate-blog/fix_hexo_md_pre_frontmatter.py
new file mode 100644
index 0000000..7cbc202
--- /dev/null
+++ b/migrate-blog/fix_hexo_md_pre_frontmatter.py
@@ -0,0 +1,20 @@
+import os
+
+def fix_frontmatter(directory):
+    for root, _, files in os.walk(directory):
+        for file in files:
+            if file.endswith('.md'):
+                file_path = os.path.join(root, file)
+                with open(file_path, 'r+', encoding='utf-8') as f:
+                    content = f.read()
+                    # 检查文件开头是否缺少 '---'
+                    if not content.startswith('---'):
+                        # 在开头添加 '---'
+                        fixed_content = '---\n' + content
+                        f.seek(0)
+                        f.write(fixed_content)
+                        print(f"Fixed frontmatter in: {file_path}")
+
+if __name__ == "__main__":
+    hexo_posts_dir = '/home/songtianlun/Sync/Develop/frytea/source/_posts'  # Hexo 文章目录
+    fix_frontmatter(hexo_posts_dir)
diff --git a/migrate-blog/migrate_from_hexo_to_wordpress.py b/migrate-blog/migrate_from_hexo_to_wordpress.py
new file mode 100644
index 0000000..1ccfdd0
--- /dev/null
+++ b/migrate-blog/migrate_from_hexo_to_wordpress.py
@@ -0,0 +1,131 @@
+import os
+import re
+import yaml
+import markdown
+from wordpress_xmlrpc import Client, WordPressPost
+from wordpress_xmlrpc.methods import posts
+from datetime import datetime
+
+# WordPress 设置
+wp_url = "https://blog.lnf1.skybyte.me/xmlrpc.php"
+wp_username = "songtianlun"
+wp_password = "sotilu,WP2024"
+
+# Hexo 文章目录
+hexo_root_dir = "/home/songtianlun/Sync/Develop/frytea/source"
+
+
+def parse_hexo_post(file_path):
+    with open(file_path, 'r', encoding='utf-8') as f:
+        content = f.read()
+
+    # 提取 front matter，允许更灵活的分隔符
+    front_matter_match = re.match(r'^[-*]{3,}\s+(.*?)\s+[-*]{3,}\s*(.*)', content, re.DOTALL)
+    if front_matter_match:
+        front_matter_str = front_matter_match.group(1)
+        post_content = front_matter_match.group(2).strip()
+
+        # 使用 PyYAML 解析 front matter
+        try:
+            front_matter = yaml.safe_load(front_matter_str)
+        except yaml.YAMLError:
+            # 如果 YAML 解析失败，使用手动解析
+            front_matter = {}
+            current_key = None
+            for line in front_matter_str.split('\n'):
+                if ':' in line and not line.strip().startswith('-'):
+                    key, value = line.split(':', 1)
+                    current_key = key.strip()
+                    front_matter[current_key] = value.strip() or None
+                elif current_key and line.strip():
+                    if isinstance(front_matter[current_key], list):
+                        front_matter[current_key].append(line.strip())
+                    elif front_matter[current_key] is None:
+                        front_matter[current_key] = line.strip()
+                    else:
+                        front_matter[current_key] = [front_matter[current_key], line.strip()]
+    else:
+        front_matter = {}
+        post_content = content
+
+    return front_matter, post_content
+
+
+def get_categories_from_path(file_path, root_dir):
+    rel_path = os.path.relpath(file_path, root_dir)
+    dir_path = os.path.dirname(rel_path)
+    categories = dir_path.split(os.sep)
+    return [cat for cat in categories if cat and cat != '_posts']
+
+
+def migrate_to_wordpress(title, content, categories, tags, date, author):
+    client = Client(wp_url, wp_username, wp_password)
+
+    post = WordPressPost()
+    post.title = title
+    post.content = markdown.markdown(content)
+    post.post_status = 'publish'
+    post.terms_names = {
+        'category': categories,
+        'post_tag': tags
+    }
+
+    if date:
+        post.date = date
+
+    post_id = client.call(posts.NewPost(post))
+    return post_id
+
+
+def parse_date(date_value):
+    if isinstance(date_value, datetime):
+        return date_value
+
+    if not isinstance(date_value, str):
+        return None
+
+    date_formats = [
+        "%Y-%m-%d %H:%M:%S",
+        "%Y-%m-%d %H:%M:%S%z",
+        "%Y-%m-%d %H:%M:%S %z",
+        "%Y-%m-%d"
+    ]
+    for fmt in date_formats:
+        try:
+            return datetime.strptime(date_value, fmt)
+        except ValueError:
+            continue
+    return None
+
+
+def main():
+    for root, dirs, files in os.walk(hexo_root_dir):
+        for filename in files:
+            if filename.endswith('.md'):
+                file_path = os.path.join(root, filename)
+                front_matter, content = parse_hexo_post(file_path)
+
+                title = front_matter.get('title', 'Untitled').strip()
+                tags = front_matter.get('tags', [])
+                if isinstance(tags, str):
+                    tags = [tag.strip() for tag in tags.split(',') if tag.strip()]
+                date = parse_date(front_matter.get('date'))
+                author = front_matter.get('author')
+
+                dir_categories = get_categories_from_path(file_path, hexo_root_dir)
+
+                categories = front_matter.get('categories', [])
+                if isinstance(categories, str):
+                    categories = [categories]
+                categories = list(set(categories + dir_categories))
+
+                post_id = migrate_to_wordpress(title, content, categories, tags, date, author)
+                print(f"Migrated post: {title} (ID: {post_id})")
+                print(f" filename: {filename}")
+                print(f"  Categories: {categories}")
+                print(f"  Tags: {tags}")
+                print(f"  Date: {date}")
+
+
+if __name__ == "__main__":
+    main()
\ No newline at end of file
diff --git a/migrate-blog/migrate_from_hexo_to_wordpress_2.py b/migrate-blog/migrate_from_hexo_to_wordpress_2.py
new file mode 100644
index 0000000..6448f5d
--- /dev/null
+++ b/migrate-blog/migrate_from_hexo_to_wordpress_2.py
@@ -0,0 +1,133 @@
+import os
+import frontmatter
+from wordpress_xmlrpc import Client, WordPressPost
+from wordpress_xmlrpc.methods.posts import GetPosts, NewPost
+from datetime import datetime
+
+# 配置部分
+hexo_posts_dir = '/home/songtianlun/Sync/Develop/frytea/source/_posts'  # Hexo 文章目录
+wordpress_url = "https://blog.lnf1.skybyte.me/xmlrpc.php"
+wordpress_username = 'songtianlun'
+wordpress_password = 'sotilu,WP2024'
+
+# 初始化 WordPress 客户端
+wp = Client(wordpress_url, wordpress_username, wordpress_password)
+
+# 上传统计信息
+total_posts = 0
+successful_posts = 0
+failed_posts = 0
+duplicate_posts = []
+
+
+def get_all_wp_posts():
+    print("Fetching existing posts from WordPress...")
+    all_posts = []
+    offset = 0
+    increment = 20  # 每次获取的文章数量，可以调整
+    while True:
+        batch = wp.call(GetPosts({'number': increment, 'offset': offset}))
+        if not batch:
+            break
+        all_posts.extend(batch)
+        offset += increment
+    print(f"Total existing posts fetched: {len(all_posts)}")
+    return {post.title: post.id for post in all_posts}
+
+
+def scan_directory_for_posts(directory, category_prefix=""):
+    posts = []
+    post_count = 1
+    for root, dirs, files in os.walk(directory):
+        category = category_prefix + os.path.basename(root)  # 使用文件夹作为类别
+        for file in files:
+            if file.endswith('.md'):
+                file_path = os.path.join(root, file)
+                post_title, post_data = process_md_file(file_path)
+                if post_title and post_data:
+                    posts.append((post_title, post_data, category, post_count))
+                    post_count += 1
+    return posts
+
+
+def process_md_file(file_path):
+    with open(file_path, 'r', encoding='utf-8') as f:
+        post_data = frontmatter.load(f)
+
+    # 构造 WordPress 文章
+    post_title = post_data['title'] if 'title' in post_data else os.path.basename(file_path)
+    return post_title, post_data
+
+
+def create_wordpress_post(post_title, post_data, category):
+    # 构造 WordPress 文章
+    post = WordPressPost()
+    post.title = post_title
+    post.content = post_data.content
+    post.terms_names = {
+        'category': [category],
+        'post_tag': post_data['tags'] if 'tags' in post_data else [],
+    }
+
+    if 'date' in post_data:
+        post.date = datetime.strptime(str(post_data['date']), '%Y-%m-%d %H:%M:%S')
+
+    post.post_status = 'publish'  # 或 'draft' 保存为草稿
+
+    return post
+
+
+def upload_post_with_retries(post, post_number, retries=3):
+    global successful_posts, failed_posts
+    attempt = 0
+    while attempt < retries:
+        try:
+            post_id = wp.call(NewPost(post))
+            print(f"Post {post_number}: '{post.title}' - Upload successful. WordPress ID: {post_id}")
+            successful_posts += 1
+            return True
+        except Exception as e:
+            attempt += 1
+            print(f"Post {post_number}: '{post.title}' - Attempt {attempt} failed with error: {e}")
+            if attempt == retries:
+                print(f"Post {post_number}: '{post.title}' - Failed to upload after {retries} attempts.")
+                failed_posts += 1
+                return False
+
+
+def upload_posts_to_wordpress(posts_to_upload):
+    global total_posts
+    total_posts = len(posts_to_upload)
+    for post, post_number in posts_to_upload:
+        try:
+            upload_post_with_retries(post, post_number)
+        except Exception:
+            print(f"Post {post_number}: '{post.title}' - Failed permanently.")
+
+
+if __name__ == "__main__":
+    existing_wp_posts = get_all_wp_posts()
+    posts = scan_directory_for_posts(hexo_posts_dir)
+    posts_to_upload = []
+
+    for post_title, post_data, category, post_number in posts:
+        if post_title in existing_wp_posts:
+            print(f"Post '{post_title}' skipped: already exists.")
+            duplicate_posts.append((post_title, post_number))
+        else:
+            post = create_wordpress_post(post_title, post_data, category)
+            posts_to_upload.append((post, len(posts_to_upload) + 1))
+
+    upload_posts_to_wordpress(posts_to_upload)
+
+    print("Migration completed!")
+    print(f"Total posts processed: {total_posts}")
+    print(f"Successfully uploaded: {successful_posts}")
+    print(f"Failed uploads: {failed_posts}")
+    if duplicate_posts:
+        print("\nDuplicate posts found:")
+        for title, post_number in duplicate_posts:
+            print(f"Title: {title} - File Number: {post_number}")
+    else:
+        print("\nNo duplicate posts found.")
+