test
test2
import os
import shutil
import markdown
import re
import json
from pathlib import Path
from PIL import Image
import time
from collections import deque, defaultdict
import subprocess
# 設定
vault_dir = Path("Obsidian_Vault_001")
output_dir = Path(r"Y:\nginx_site_data") / "asyncd_filament_database"
top_url = "https://mizugane.tcpexposer.com/"
home_url = "https://mizugane.tcpexposer.com/asyncd_filament_database/"
site_title = "asyncd_filament_database"
max_depth=4
# HTML テンプレート
HEAD = """<html><head><meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<link rel='icon' href='{top_url}img/favicon.ico'>
<link rel='stylesheet' href='{top_url}css/style2.css'>
{additional_css}
<script src="https://unpkg.com/budoux/bundle/budoux-ja.min.js"></script>
<meta property="og:title" content="{site_title}">
<meta property="og:description" content="{title}">
<meta property="og:image" content="{top_url}img/asyncd_filament_umekome.webp">
<meta property="og:url" content="{top_url}">
<meta property="og:type" content="website">
<meta name="twitter:card" content="summary_large_image">
<meta name="twitter:title" content="{site_title}">
"""
SITE_TITLE = """
<div class=site_title ><a href="{home_url}" >{site_title}</a></div>
<hr/>
"""
TITLE = """
<h1>{title}</h1>
"""
TOP_TITLE = """
<h1>{site_title}</h1>
"""
FOOT_INDEX = """
<br><hr/><br><div class="footer-nav">
<a href="../">目次へ</a>
<br><br><br><br>
</div>
</body></html>"""
FOOT_TOP = """
<br><hr/><br>
<a href="{top_url}">トップページへ</a>
<br><br><br><br>
</div>
</body></html>"""
assets_dir = output_dir / "assets"
# Markdown 初期化
md = markdown.Markdown(
extensions=[
'extra', # テーブル・フェンスコード対応
'fenced_code', # ``` を明示的に有効化
'codehilite', # シンタックスハイライト
],
extension_configs={
'codehilite': {'linenums': False}
},
output_format='html5'
)
internal_link_map = {}
graph = defaultdict(set)
# ── キャッシュディレクトリ＆メタ情報読み込み ──
vault_name = vault_dir.name
cache_dir = vault_dir.parent / f"{vault_name}_tmp"
cache_dir.mkdir(parents=True, exist_ok=True)
cache_meta_path = cache_dir / "cache_meta.json"
if cache_meta_path.exists():
with open(cache_meta_path, 'r', encoding='utf-8') as f:
cache_meta = json.load(f)
else:
cache_meta = {}
# ── オリジナル削除時はキャッシュもクリア ──
removed = []
for rel, meta in cache_meta.items():
orig_path = vault_dir / rel
if not orig_path.exists():
tmp_file = cache_dir / meta['out_name']
if tmp_file.exists():
tmp_file.unlink()
removed.append(rel)
for rel in removed:
cache_meta.pop(rel, None)
if removed:
with open(cache_meta_path, 'w', encoding='utf-8') as f:
json.dump(cache_meta, f, indent=2)
def prepare_output():
# ── output_dir の初期化はそのまま ──
if output_dir.exists():
if output_dir.is_file():
output_dir.unlink()
else:
shutil.rmtree(output_dir)
output_dir.mkdir(parents=True, exist_ok=True)
# ── assets_dir の初期化 ──
if assets_dir.exists():
if assets_dir.is_file():
assets_dir.unlink() # ファイルなら削除
else:
shutil.rmtree(assets_dir) # ディレクトリなら中身ごと削除
assets_dir.mkdir(parents=True, exist_ok=True)
def replace_links(md_text):
css_classes = []
def repl_embed(match):
filename_part = match.group(1)
parts = filename_part.split('|', 1)
base_filename = parts[0].strip()
size_spec = parts[1].strip() if len(parts) > 1 else None
base_name = os.path.basename(base_filename)
file_ext = os.path.splitext(base_name)[1].lower()
audio_ext = ('.wav', '.flac', '.ogg', '.m4a')
video_ext = ('.mp4', '.mov', '.avi', '.mkv', '.webm')
if file_ext in audio_ext:
mp3_name = os.path.splitext(base_name)[0] + '.mp3'
return f'<audio controls src="{home_url}assets/{mp3_name}"></audio>'
if file_ext in video_ext:
webm_name = os.path.splitext(base_name)[0] + '.webm'
size_attr = ''
if size_spec:
w, h = size_spec.split('x',1)
if w.isdigit() and h.isdigit():
size_attr = f' width="{w}" height="{h}"'
return f'<video controls src="{home_url}assets/{webm_name}"{size_attr}></video>'
if file_ext in ('.png', '.jpg', '.jpeg', '.webp'):
width = height = None
if size_spec:
w, h = size_spec.split('x', 1)
try:
width, height = int(w), int(h)
except:
pass
class_attr = ''
if width and height:
cls = f'img-w{width}-h{height}'
css_classes.append((cls, width, height))
class_attr = f' class="{cls}"'
webp_name = os.path.splitext(base_name)[0] + ".webp"
return f'<img src="{home_url}assets/{webp_name}" alt="{webp_name}"{class_attr}>'
else:
# その他ファイルはリンク
return f'<a href="{home_url}assets/{base_name}">{base_name}</a>'
def repl_internal(match):
filename = match.group(1)
link_path = internal_link_map.get(filename)
if link_path:
return f'<a href="{home_url}{link_path}/">{filename}</a>'
else:
return filename
# マークダウン中の <a href="https://mizugane.tcpexposer.com/asyncd_filament_database/assets/...">...</a> を置換
md_text = re.sub(r'!\[\[(.+?)\]\]', repl_embed, md_text)
# 内部リンクはそのまま
md_text = re.sub(r'(?<!\!)\[\[(.+?)\]\]', repl_internal, md_text)
# CSSクラスの重複除去（既存処理）
seen = set()
unique = []
for cls, w, h in css_classes:
if (cls, w, h) not in seen:
seen.add((cls, w, h))
unique.append((cls, w, h))
return md_text, unique
def create_parent_index(parent_dir, title, link):
if '.obsidian' in str(parent_dir):
return
parent_html_path = output_dir / parent_dir / "index.html"
parent_html_path.parent.mkdir(parents=True, exist_ok=True)
with open(parent_html_path, 'a', encoding='utf-8') as f:
f.write(f'<a href="{home_url}{link}/">{title}</a><br>\n')
def create_directory_index(directory_path, title, links):
directory_html_path = output_dir / directory_path / "index.html"
directory_html_path.parent.mkdir(parents=True, exist_ok=True)
full_dir_path = vault_dir / directory_path
subdirs = []
if full_dir_path.exists():
subdirs = [item.name for item in full_dir_path.iterdir() if item.is_dir()]
subdir_links_html = ""
if subdirs:
subdir_links_html = "<div class='directory-links-box'>\n"
subdir_links_html += "<b>ディレクトリ</b><br>\n"
for subdir in sorted(subdirs):
subdir_url = f"{home_url}{(directory_path / subdir).as_posix()}/"
subdir_links_html += f"<a href=\"{subdir_url}\">{subdir}</a><br>\n"
subdir_links_html += "</div>\n"
links_html = "\n".join([
f'<a href=\"{home_url}{(directory_path / link).as_posix()}/\">{link}</a><br>'
for link in links
])
with open(directory_html_path, 'w', encoding='utf-8') as f:
f.write(HEAD.format(title=title, top_url=top_url, site_title=site_title, additional_css="", home_url=home_url) + SITE_TITLE.format(home_url=home_url, site_title=site_title) + TITLE.format(title=title) + subdir_links_html + links_html + FOOT_INDEX)
def collect_nodes_and_links(start_node):
visited = {}
queue = deque([(start_node, 0)])
visited[start_node] = 0
while queue:
node, depth = queue.popleft()
if depth >= max_depth:
continue
for neighbor in graph.get(node, []):
if neighbor not in visited or visited[neighbor] > depth + 1:
visited[neighbor] = depth + 1
queue.append((neighbor, depth + 1))
nodes_in_scope = [n for n in visited if visited[n] <= max_depth]
links_in_scope = []
seen_edges = set()
for node in nodes_in_scope:
for neighbor in graph.get(node, []):
if neighbor in nodes_in_scope:
edge = tuple(sorted([node, neighbor]))
if edge not in seen_edges:
seen_edges.add(edge)
links_in_scope.append({"source": edge[0], "target": edge[1]})
return nodes_in_scope, links_in_scope
def process_file(md_path, rel_path):
if '.obsidian' in str(md_path):
return
with open(md_path, 'r', encoding='utf-8') as f:
md_text = f.read()
md_text, css_classes = replace_links(md_text)
md.reset()
html_body = md.convert(md_text)
html_body = re.sub(
r'<p>\s*<a href="([^"]+\.mp3)">[^<]+</a>\s*</p>',
r'<audio controls src="\1"></audio>',
html_body
)
title = rel_path.stem
parent_dir = rel_path.parent
if parent_dir != Path('.'):
create_parent_index(parent_dir, title, rel_path.with_suffix("").as_posix())
output_page_dir = output_dir / rel_path.with_suffix("")
output_path = output_page_dir / "index.html"
output_path.parent.mkdir(parents=True, exist_ok=True)
# CSSファイルの生成
css_content = ""
for cls, w, h in css_classes:
css_content += f".{cls} {{\n width: {w}px;\n height: {h}px;\n}}\n"
additional_css = ''
if css_content:
css_path = output_page_dir / "image_styles.css"
with open(css_path, 'w', encoding='utf-8') as f:
f.write(css_content)
additional_css = '<link rel="stylesheet" href="image_styles.css">'
# JSONデータ生成（既存処理）
current_page = title
nodes_in_scope, links_in_scope = collect_nodes_and_links(current_page)
nodes_data = []
for node in nodes_in_scope:
node_path = internal_link_map.get(node)
if node_path:
url = f"{home_url}{node_path}"
nodes_data.append({"id": node, "path": url})
json_data = {
"nodes": nodes_data,
"links": links_in_scope
}
json_path = output_dir / rel_path.with_suffix("") / "data.json"
with open(json_path, 'w', encoding='utf-8') as f:
json.dump(json_data, f, indent=2)
# Generate D3 visualization code
d3_code = f"""
<!-- SVG and D3.js Integration -->
<div id="graph-container"
class="graph-container"
data-current="{current_page}"
data-path="data.json">
<svg ></svg>
</div>
<script src="https://d3js.org/d3.v7.min.js"></script>
<script src="{top_url}js/graph.js" type="module"></script>
"""
head_content = HEAD.format(
title=title,
top_url=top_url,
site_title=site_title,
additional_css=additional_css,
home_url=home_url
)
site_title_content = SITE_TITLE.format(home_url=home_url, site_title=site_title)
title_content = TITLE.format(title=title)
with open(output_path, 'w', encoding='utf-8') as f:
f.write(
head_content
+ site_title_content
+ title_content
+ "<budoux-ja>\n"
+ html_body
+ "\n</budoux-ja>"
+ d3_code
+ FOOT_INDEX
)
def save_cache_meta():
with open(cache_meta_path, 'w', encoding='utf-8') as f:
json.dump(cache_meta, f, indent=2)
def copy_asset_with_cache(file_path: Path):
rel = str(file_path.relative_to(vault_dir))
mtime = file_path.stat().st_mtime
ext = file_path.suffix.lower()
# 拡張子に応じた出力名
if ext in ('.png','.jpg','.jpeg','.webp'):
out_name = file_path.stem + ".webp"
elif ext in ('.wav','.flac','.ogg','.m4a'):
out_name = file_path.stem + ".mp3"
elif ext in ('.mp4','.mov','.avi','.mkv','.webm'):
out_name = file_path.stem + ".webm"
else:
out_name = file_path.name
# キャッシュヒット判定
meta = cache_meta.get(rel)
if meta and meta['mtime'] == mtime and (cache_dir/meta['out_name']).exists():
# 既存キャッシュを assets_dir へコピー
shutil.copy(cache_dir / meta['out_name'], assets_dir / out_name)
return
# キャッシュに無ければ新規圧縮 or コピー
target_cache = cache_dir / out_name
if ext in ('.png','.jpg','.jpeg','.webp'):
img = Image.open(file_path)
img.thumbnail((1024,1024), Image.LANCZOS)
img.save(target_cache, 'WEBP', quality=85)
elif ext in ('.mp4','.mov','.avi','.mkv','.webm'):
subprocess.run([
'ffmpeg','-y','-i',str(file_path),
'-c:v','libvpx-vp9','-b:v','0','-crf','30',
'-c:a','libopus', str(target_cache)
], check=True)
elif ext in ('.wav','.flac','.ogg','.m4a'):
subprocess.run([
'ffmpeg','-y','-i',str(file_path),
'-codec:a','libmp3lame','-b:a','64k', str(target_cache)
], check=True)
else:
shutil.copy(file_path, target_cache)
# assets_dir へコピー
shutil.copy(target_cache, assets_dir / out_name)
# メタ情報を更新
cache_meta[rel] = {"out_name": out_name, "mtime": mtime}
save_cache_meta()
def create_top_index(all_md_paths):
tree = {}
for rel_path in all_md_paths:
parts = Path(rel_path).parts
current = tree
for part in parts[:-1]:
current = current.setdefault(part, {})
current.setdefault('_files', []).append(parts[-1])
def render_tree(node, depth=0):
html = ""
indent = " " * 4 * depth
for file in sorted(node.get('_files', [])):
file_path = '/'.join(node_path + [file])
html += f"{indent}<a href=\"{home_url}{file_path}/\">{file}</a><br>\n"
for dir_name in sorted(k for k in node.keys() if k != '_files'):
node_path.append(dir_name)
if '.obsidian' not in dir_name:
html += f"{indent}<b>{dir_name}/</b><br>\n"
html += render_tree(node[dir_name], depth + 1)
node_path.pop()
return html
node_path = []
body_html = render_tree(tree)
parent_dir_links_html = "<div class='directory-links-box'>\n"
parent_dir_links_html += "<b>ディレクトリ</b><br>\n"
subdirs = sorted([d.name for d in Path(vault_dir).iterdir() if d.is_dir() and '.obsidian' not in d.name])
for subdir in subdirs:
subdir_url = f"{home_url}{subdir}/"
parent_dir_links_html += f"<a href=\"{subdir_url}\">{subdir}</a><br>\n"
parent_dir_links_html += "</div>\n"
output_path = output_dir / "index.html"
with open(output_path, 'w', encoding='utf-8') as f:
f.write(
HEAD.format(title=site_title, top_url=top_url, site_title=site_title, additional_css="", home_url=home_url)
+ TOP_TITLE.format(site_title=site_title)
+ parent_dir_links_html
+ body_html
+ FOOT_TOP.format(top_url=top_url)
)
def build_site():
global internal_link_map, graph
# Build internal_link_map
for root, dirs, files in os.walk(vault_dir):
if '.obsidian' in dirs:
dirs.remove('.obsidian')
for file in files:
if file.lower().endswith(".md"):
path = Path(root) / file
rel_path = path.relative_to(vault_dir)
internal_link_map[file[:-3]] = rel_path.with_suffix("").as_posix()
# Build graph
graph = defaultdict(set)
for root, dirs, files in os.walk(vault_dir):
if '.obsidian' in dirs:
dirs.remove('.obsidian')
for file in files:
if file.lower().endswith(".md"):
path = Path(root) / file
rel_path = path.relative_to(vault_dir)
current_page = rel_path.stem
with open(path, 'r', encoding='utf-8') as f:
md_text = f.read()
links = re.findall(r'(?<!\!)\[\[(.+?)\]\]', md_text)
for link in links:
linked_page = link.strip()
if linked_page in internal_link_map:
graph[current_page].add(linked_page)
graph[linked_page].add(current_page)
# Process files
directory_links = {}
all_md_links = []
for root, dirs, files in os.walk(vault_dir):
if '.obsidian' in dirs:
dirs.remove('.obsidian')
for file in files:
path = Path(root) / file
rel_path = path.relative_to(vault_dir)
if file.lower().endswith(".md"):
process_file(path, rel_path)
parent_dir = rel_path.parent
directory_links.setdefault(parent_dir, []).append(rel_path.stem)
all_md_links.append(rel_path.with_suffix("").as_posix())
else:
copy_asset_with_cache(path)
# Create directory indexes
for directory, links in directory_links.items():
create_directory_index(directory, directory.name, links)
# Create top index
create_top_index(all_md_links)
if __name__ == "__main__":
prepare_output()
build_site()
print("Build complete.")