process_posts.py

from collections import defaultdict
import json
from math import ceil
from pathlib import Path
import datetime as dt

# pip3 install pyyaml
import yaml

POSTS_DIR = Path(__file__).parent / "content/posts/"

POSTS_COUNT_JSON_PATH = Path(__file__).parent / "data/posts_count.json"

TZ_UTC = dt.timezone(dt.timedelta(hours=0))


def get_all_posts():
    """
    遍历位于 posts 中所有的 Markdown 文档
    """
    return list(POSTS_DIR.glob("**/*.md"))


def parse_post_metadata(path: Path):
    """
    解析 Markdown 头部的 hugo metadata 信息
    """
    with path.open("r", encoding="utf-8") as f:
        metadata_lines = [f.readline()]
        for line in f:
            if line.strip(" ").strip("\n") != "---":
                metadata_lines.append(line)
            else:
                # it's the end of markdown metadata
                return yaml.safe_load("".join(metadata_lines))
            
    print(f"failed to load metadata: {path}")
    raise Exception("failed to load metadata")


def gen_folder_name(post: Path):
    """
    生成新的父目录名称，后续会将 posts 移动到此目录中

    生成规则举例：
    1. 按 年/月-日 或者 年/季度 对文章进行分类
    2. 按词云的权重或者相似度对文章进行智能分类
    3. 其他...
    """
    post_metadata = parse_post_metadata(post)
    post_time: dt.datetime = post_metadata["date"]
    if not isinstance(post_time, dt.datetime):
        raise Exception(f"invalid date: {post_time}")

    # 每三个月一个季度，一年 4 个季度
    quarter = ceil(post_time.month / 3)

    # 目前暂按 年/季度 进行分类
    return f"{post_time.year}/Q{quarter}"


def restructure_posts(posts):
    processed_files = []
    for p in posts:
        if p in processed_files:
            continue

        # 新的 Posts 父目录
        new_post_dir = POSTS_DIR / gen_folder_name(p)
        new_post_dir.mkdir(parents=True, exist_ok=True)

        # 将 posts 的所有文件移动到新的位置
        for file in p.parent.iterdir():
            processed_files.append(file)

            new_parent = new_post_dir / file.parent.name
            new_parent.mkdir(parents=True, exist_ok=True)
            new_path = new_parent / file.name

            print(new_path)
            file.rename(new_path)

        if not any(p.parent.iterdir()):
            # 删除 post 空的旧目录
            p.parent.rmdir()


def yearmonth_range(start_year, start_month, end_year, end_month):
    year, month = start_year, start_month
    while (year, month) <= (end_year, end_month):
        yield year, month

        month += 1
        if month > 12:
            month = 1
            year += 1


def gen_posts_count_by_month(posts):
    data = defaultdict(int)
    utcnow = dt.datetime.utcnow().replace(tzinfo=TZ_UTC)
    earliest = utcnow
    for p in posts:
        post_metadata = parse_post_metadata(p)

        if post_metadata['draft']:
            continue

        post_time: dt.datetime = post_metadata["date"]
        post_time = post_time
        date_str = f"{post_time.year}-{post_time.month:02}"
        data[date_str] += 1

        if post_time < earliest:
            earliest = post_time

    for year, month in yearmonth_range(earliest.year, earliest.month, utcnow.year, utcnow.month):
        date_str = f"{year}-{month:02}"
        if date_str not in data:
            data[date_str] = 0

    sorted_data = sorted(data.items())
    date_list = [it[0] for it in sorted_data]
    count_list = [it[1] for it in sorted_data]

    with POSTS_COUNT_JSON_PATH.open("w+", encoding="utf-8") as f:
        json.dump(
            {"xAxis": date_list, "yAxis": count_list},
            f,
            ensure_ascii=False,
            indent=2,
            sort_keys=True,
        )


def main():
    posts = get_all_posts()
    # 调整 posts 文件夹结构
    restructure_posts(posts)

    # 生成按月计数的文章统计数据，用于绘制图表
    gen_posts_count_by_month(posts)


if __name__ == "__main__":
    main()