"""
主要将plantuml的mindmap写法转为vega可用的json文件
"""
import sys
import re
import json
import os
import base64
import zlib


def converter(puml_path: str):
    """
    传入puml文件路径进行解析转化
    1. 标题都是以*开头, 且一个*的都是根节点
    2. 父级节点只会比子级节点少一个*，如果当前节点比下一个节点少于超过一个*，puml就无法通过
    3. 如果下一个节点比上一个节点少*，就去对应列表里面找最后一个
    :param puml_path:
    :return:
    """
    print(f"开始处理{puml_path}...")
    levels = {}
    json_results = []
    # 找到同级的md文档目录
    md_dir_path = f"{puml_path.replace(puml_path.split('/')[-1], '')}md"
    data_json_path = f"{puml_path.replace(puml_path.split('/')[-1], '')}data.json"
    with open(puml_path, 'r') as f:
        notes = extract_notes(md_dir_path, f.read())

    with open(puml_path, 'r') as f:
        lines = [line for line in f.readlines()]
        title_index = 1
        for index, line in enumerate(lines):
            # 标题的*后面只会出现三种情况：空格、:、[
            if line.startswith('*'):
                stars, name, color, links = extract_stars_name_links_color(line)
                levels[stars] = (line, title_index)
                parent = levels.get(stars[:-1])
                node = {
                    "id": title_index,
                    "name": name,
                    # "size": len(name)
                    # "link": 'https://www.google.com'
                }
                if parent:
                    node["parent"] = parent[1]
                if links:
                    # 如果是有链接，就变成子节点
                    link_count = 1
                    for link_name, link in links.items():
                        title_index += 1
                        wrap_link_name = get_wrap_name(f"链接{link_count}: {link_name}")
                        child_node = {
                            "id": title_index,
                            "name": wrap_link_name,
                            "link": link,
                            "parent": node['id'],
                            "note": f'[来自{node["name"]}的链接]({link})'
                        }
                        json_results.append(child_node)
                        link_count += 1
                if color:
                    node["color"] = '#' + color
                if index < len(lines) and lines[index + 1].startswith('<code>'):
                    note = notes.pop(0)
                    # print(f"弹出的注释：{note}")
                    node['note'] = note
                json_results.append(node)
                title_index += 1
    result_path = puml_path.replace('.puml', '.json')
    with open(data_json_path, 'w') as f:
        f.write(json.dumps(json_results))


def extract_stars_name_links_color(line=''):
    color = None
    links = re.findall('\[\[(.*?)\]\]', line)
    link_dict = {}
    for link in links:
        href, title = link.split(' ', 1)
        line = line.replace(f"[[{href} {title}]]", f" {title}")
        link_dict[title] = href
    try:
        stars = re.split('[ :\[]', line)[0]
        name = line[len(stars):]
        if name.startswith('[#'):  # 如果有颜色
            color = re.findall('\[#(.*?)\]', name)[0]
            name = name.split(']')[1]
        if name.startswith(':'):  # 如果有注释
            name = name[1:]
    except:
        print(line)
    wrap_name = get_wrap_name(name)
    return stars, wrap_name, color, link_dict


def get_wrap_name(name):
    # 统一添加换行符
    wrap_name = []
    space_count = 0
    for char in name:
        if char == ' ':
            space_count += 1
        if space_count == 3:
            char = '\n'
            space_count = 0
        wrap_name.append(char)
    return ''.join(wrap_name)


# [Kroki!](https://kroki.io/)
def get_kroki_preprocessors():
    preprocessors = {
        "puml": "plantuml",
        "seqdiag": "sequediag",
        "svgbob": "svgbob",
        "ditaa": "ditaa",

    }
    return preprocessors


# 将puml/mermaid等内容提交给kroki获取在线图片链接
def get_kroki_link(file_path, preprocessor):
    with open(file_path, 'r') as f:
        content = f.read()
        encoded_cotnent = base64.urlsafe_b64encode(zlib.compress(content.encode('utf8'))).decode('utf8')
        return f"https://kroki.io/{preprocessor}/svg/{encoded_cotnent}"


# 1. 提取内容
# 2. 如果是md文档地址，就取文档地址内容为note
# 3. 对note的内容进行处理
def extract_notes(md_dir_path, text=''):
    #     text = '''
    #         ****:tail -n 80 customSpec.json
    # <code>
    #
    # 此命令显示 Wasm 二进制字段后面的最后部分，
    # 包括运行时使用的几个托盘的详细信息，
    # 例如 sudo 和 balances 托盘。
    # </code>;
    # ****:Modify the name field to identify this chain specification as a custom chain specification.
    # <code>
    #
    # "name": "My Custom Testnet",
    # </code>
    # ****:Modify aura field to specify the nodes
    # <code>
    #     '''
    # 同时匹配换行符
    # (?:pattern) 来解决方括号不适用的场景
    # [正则匹配所有字符（包括换行）_尐东东的博客-CSDN博客_正则匹配所有字符](https://blog.csdn.net/u011158908/article/details/105666329)
    notes = re.findall('\<code\>((?:.|\n)*?)\</code\>', text)
    # 考虑到html默认只支持br换行，所以这里统一替换成br
    # notes = [note.replace('\n', '<br>') for note in notes]
    preprocessors = get_kroki_preprocessors()

    def inner_note_replace(note):
        # 如果是md文件地址，就替换
        if note.startswith("md_file:"):
            md_file = note.replace("md_file:", "")
            with open(f"{md_dir_path}/{md_file}", 'r') as f:
                note = f.read()
        # 考虑到plantuml的';'是元素结束符，所以这里将'";"'换成';'
        note = note.replace('";"', ';')
        # 提取其中的markdown图片链接，如果是puml后缀，就单独处理
        # ![](xxx.puml) -> ![xxx.puml](new_path)
        img_links = re.findall('!\[(.*?)\]\((.*?)\)', note)
        for img_link in img_links:
            img_name, img_path = img_link
            suffix = img_path.split('.')[-1]
            if suffix in preprocessors.keys():
                kroki_link = get_kroki_link(img_path, preprocessors[suffix])
                file_name = img_path.split('/')[-1]
                note = note.replace(f"![]({img_path})",
                                    f"- [{file_name}点开大图]({kroki_link})\n![{file_name}]({kroki_link})")
        return note

    notes = [inner_note_replace(note.strip()) for note in notes]
    return notes


def extract_links(text=''):
    links = re.findall('\[\[(.*?)\]\]', text)
    link_dict = {}
    for link in links:
        href, title = link.split(' ', 1)
        link_dict[title] = href
    return link_dict


if __name__ == "__main__":
    if len(sys.argv) != 2:
        print("请传入puml文件路径...")
    else:
        puml_path = sys.argv[1]
        puml_paths = []
        if os.path.isdir(puml_path):
            puml_paths = [f"{puml_path}/{item}" for item in os.listdir(puml_path) if item.endswith('.puml')]
        elif not puml_path.endswith('.puml'):
            print("请传入puml文件...")
        else:
            puml_paths.append(puml_path)
        for puml_path in puml_paths:
            converter(puml_path)

<head>
    <script src="https://cdn.jsdelivr.net/npm/vega@5"></script>
    <script src="https://cdn.jsdelivr.net/npm/vega-lite@5"></script>
    <script src="https://cdn.jsdelivr.net/npm/vega-embed@6"></script>
</head>
<body>
<div id="vis"/>
</code>
<script>
    vegaEmbed(
        '#vis',
        'vega/tree.vg.json'
    );
</script>
</body>

<head>
    <script src="https://cdn.jsdelivr.net/npm/vega@5"></script>
</head>
<body>
<div id="view"></div>
<script type="text/javascript">
    var view;

    fetch('vega/circle_packing.vg.json')
        .then(res => res.json())
        .then(spec => render(spec))
        .catch(err => console.error(err));

    function render(spec) {
        view = new vega.View(vega.parse(spec), {
            renderer:  'svg',  // renderer (canvas or svg)
            container: '#view',   // parent DOM container
            hover:     true       // enable hover processing
        });
        return view.runAsync();
    }
</script>
</body>

• Four Views of the Same Data: Scatterplot, Clustered Barchart, Clustered Barchart, Stacked Barchart
• Question: How Is a Value Distributed?: Categorical Histogram , Quantitative Histogram , Smoothed Histogram , Box Plot, Categorical Density Plot, Continuous Density Plot
• Question: How Do Groups Differ from Each Other?: Bar Chart, Paired (or Multiple) Series Bar Chart, Pie (or Doughnut ) Chart, Heatmap
• Question: Do Invidual Items Fall Into Groups? Is There a Relationship Between Attributes of Items?: Scatterplot, Line Chart, Stacked Area Chart
• Question: How Are Objects Related To Each Other in a Network or Hierarchy?: Node-Link View (Force-Directed Layout), Circular Network Layout, Adjacency Matrix, Tree View, Treemap, Sunburst plot
• Question: Where Are Objects Located?: Choropleth, Dotplot map
• Question: What Is In This Text?: Word Cloud
• Chapter 6: Multiviews: Small Multiples, Split by Dimension, Small Multiples, Showing Different Measures, SPLOM, Cross-Selection