ジッピーにモゴミのwikiのapi叩いて記事のhtmlをbootstrap風に出力しろと命令した結果出てきたゴミ

@ntxIii7WlrtWGxtxxpSH / 更新: 2026/01/04 23:06

紛うことなきゴミなので扱いには要注意

TEXT 5.96KB
  1. #!/usr/bin/python3
  2. # -*- coding: utf-8 -*-
  4. import sys
  5. import requests
  6. from bs4 import BeautifulSoup
  7. from urllib.parse import urljoin, quote
  8. import re
  9. import html
  11. WIKI_BASE = "https://wiki.yjsnpi.nu/wiki"
  12. API_ENDPOINT = "https://wiki.yjsnpi.nu/w/api.php"
  14. if len(sys.argv) < 2:
  15. print("usage: {} ページ名".format(sys.argv[0]), file=sys.stderr)
  16. sys.exit(1)
  18. page_name = sys.argv[1]
  20. def fetch_parsed_html(page):
  21. params = {
  22. "action": "parse",
  23. "page": page,
  24. "prop": "text|displaytitle",
  25. "format": "json",
  26. }
  27. r = requests.get(API_ENDPOINT, params=params, timeout=15)
  28. r.raise_for_status()
  29. data = r.json()
  30. if "error" in data:
  31. raise RuntimeError("API error: {}".format(data["error"]))
  32. if "parse" not in data:
  33. raise RuntimeError("ページが見つかりません: " + page)
  34. return data["parse"].get("displaytitle") or page, data["parse"]["text"]["*"]
  36. # 簡易 style -> Bootstrap class マッピング
  37. STYLE_CLASS_MAP = [
  38. # 代表的な置換パターン (順に評価)
  39. (re.compile(r'^\s*float\s*:\s*right\s*;?\s*$', re.I), "float-end"),
  40. (re.compile(r'^\s*float\s*:\s*left\s*;?\s*$', re.I), "float-start"),
  41. (re.compile(r'^\s*display\s*:\s*block\s*;?\s*$', re.I), "d-block"),
  42. (re.compile(r'^\s*display\s*:\s*inline-block\s*;?\s*$', re.I), "d-inline-block"),
  43. (re.compile(r'^\s*text-align\s*:\s*center\s*;?\s*$', re.I), "text-center"),
  44. (re.compile(r'^\s*text-align\s*:\s*right\s*;?\s*$', re.I), "text-end"),
  45. (re.compile(r'^\s*text-align\s*:\s*left\s*;?\s*$', re.I), "text-start"),
  46. (re.compile(r'^\s*width\s*:\s*100%\s*;?\s*$', re.I), "w-100"),
  47. (re.compile(r'^\s*max-width\s*:\s*100%\s*;?\s*$', re.I), "img-fluid"),
  48. (re.compile(r'^\s*margin\s*:\s*0\s*;\s*$', re.I), "m-0"),
  49. (re.compile(r'^\s*margin-left\s*:\s*auto\s*;\s*margin-right\s*:\s*auto\s*;?\s*$', re.I), "mx-auto"),
  50. (re.compile(r'^\s*margin-top\s*:\s*(\d+)px\s*;?\s*$', re.I), lambda m: f"mt-{int(int(m.group(1))/4)}"), # rough
  51. (re.compile(r'^\s*padding\s*:\s*(\d+)px\s*;?\s*$', re.I), lambda m: f"p-{int(int(m.group(1))/4)}"),
  52. ]
  54. def style_to_bootstrap_classes(style_value):
  55. if not style_value:
  56. return None
  57. # normalize ; spacing
  58. style_value = style_value.strip()
  59. # attempt exact-match mappings first
  60. classes = []
  61. # split multiple declarations and try heuristic mapping for each
  62. parts = [p.strip() + (';' if not p.strip().endswith(';') else '') for p in style_value.split(';') if p.strip()]
  63. for part in parts:
  64. mapped = None
  65. for pattern, cls in STYLE_CLASS_MAP:
  66. m = pattern.match(part)
  67. if m:
  68. mapped = cls(m) if callable(cls) else cls
  69. break
  70. if mapped:
  71. classes.append(mapped)
  72. if not classes:
  73. # fallback: if style contains color/background/font-size etc., keep nothing or map generically
  74. return None
  75. # join unique, filter None
  76. uniq = []
  77. for c in classes:
  78. if c and c not in uniq:
  79. uniq.append(c)
  80. return " ".join(uniq) if uniq else None
  82. def rewrite_images_links_styles(html_str):
  83. soup = BeautifulSoup(html_str, "html.parser")
  85. # images
  86. for img in soup.find_all("img"):
  87. src = img.get("src")
  88. if src:
  89. abs_src = urljoin(WIKI_BASE, src)
  90. encoded = quote(abs_src, safe=":/?&=#%")
  91. new_src = f"https://external-content.duckduckgo.com/iu/?u={encoded}"
  92. img["src"] = new_src
  93. if img.has_attr("srcset"):
  94. del img["srcset"]
  95. # convert inline style on img to classes if possible
  96. if img.has_attr("style"):
  97. cls = style_to_bootstrap_classes(img["style"])
  98. del img["style"]
  99. if cls:
  100. existing = img.get("class", [])
  101. img["class"] = existing + cls.split()
  103. # links
  104. for a in soup.find_all("a", href=True):
  105. href = a["href"]
  106. if href.startswith("/wiki/") or "/wiki/" in href:
  107. new_href = href.replace("/wiki/", "/w/")
  108. a["href"] = new_href
  109. elif href.startswith("/w/index.php") and "title=" in href:
  110. import urllib.parse as _up
  111. parsed = _up.urlparse(href)
  112. qs = _up.parse_qs(parsed.query)
  113. title = qs.get("title", [None])[0]
  114. if title:
  115. a["href"] = "/w/" + quote(title, safe="/#?&=%")
  116. elif href.startswith(WIKI_BASE):
  117. new_href = href.replace(WIKI_BASE, "/w")
  118. a["href"] = new_href
  119. # convert style on links
  120. if a.has_attr("style"):
  121. cls = style_to_bootstrap_classes(a["style"])
  122. del a["style"]
  123. if cls:
  124. existing = a.get("class", [])
  125. a["class"] = existing + cls.split()
  127. # all other elements: convert style -> class where possible
  128. for tag in soup.find_all(True):
  129. if tag.has_attr("style"):
  130. cls = style_to_bootstrap_classes(tag["style"])
  131. del tag["style"]
  132. if cls:
  133. existing = tag.get("class", [])
  134. tag["class"] = existing + cls.split()
  136. return str(soup)
  138. def remove_mw_editsection(html_str):
  139. soup = BeautifulSoup(html_str, "html.parser")
  140. for span in soup.find_all("span", class_="mw-editsection"):
  141. span.decompose() # ノードごと削除
  142. return str(soup)
  144. def save_html(title, body_html, out_filename):
  145. with open(out_filename, "w", encoding="utf-8") as f:
  146. f.write(body_html)
  148. try:
  149. title, html_content = fetch_parsed_html(page_name)
  150. except Exception as e:
  151. print("parse API に失敗:", e, file=sys.stderr)
  152. sys.exit(2)
  154. modified = rewrite_images_links_styles(html_content)
  155. modified = remove_mw_editsection(modified)
  156. out_filename = "{}.html".format(page_name.replace("/", "_"))
  157. save_html(title, modified, out_filename)
  158. print("保存完了:", out_filename)
タグ: #ゴミ #淫夢

迫真広告部
野獣ノート
怪文書投稿プラットフォーム
風戸雄琉チャレンジ 挑戦者募集中
この記事のURLをXに貼って雄琉がｼｭﾊﾞｯて来なければ勝ち
最近追加されたノート