#!/usr/bin/env python3

"""Add i18n infrastructure to AI tool pages in /workspace/ai/"""

import os, re, html as html_mod

AI_DIR = '/workspace/ai'

EXISTING_I18N_FILES = set(['ai-title','ai-email','ai-resume','ai-summary',

'ai-interview','ai-japanese-essay','ai-code-explain'])

EXISTING_I18N = {f+'.html' for f in EXISTING_I18N_FILES}

def has_i18n(content):

return 'data-i18n' in content or 'pageTranslations' in content or 'ZT_PAGE' in content

def find_meta(content, name):

m = re.search(rf'<meta\s+name="{name}"\s+content="([^"]*)"', content)

return html_mod.unescape(m.group(1)) if m else ''

def find_title(content):

m = re.search(r'<title>([^<]*)</title>', content)

return html_mod.unescape(m.group(1)) if m else 'AI Tool'

def extract_ui_texts(content):

"""Extract Chinese UI text nodes likely needing translation."""

texts = {}

# h1, h2 text

for tag in ['h1','h2','h3']:

for m in re.finditer(rf'<{tag}[^>]*>([^<]+)</{tag}>', content):

t = m.group(1).strip()

if len(t) >= 2 and t not in texts:

texts[t] = '_' + re.sub(r'\s+', '_', t)[:30]

# button text

for m in re.finditer(r'<button[^>]*>([^<]+)</button>', content):

t = m.group(1).strip()

if len(t) >= 2 and t not in texts:

key = re.sub(r'[^a-zA-Z0-9]+','_',t).lower().strip('_')[:20]

texts[t] = key or 'btn'

# label text

for m in re.finditer(r'<label[^>]*>([^<]+)</label>', content):

t = m.group(1).strip()

if len(t) >= 2 and t not in texts:

texts[t] = re.sub(r'[^a-zA-Z0-9]+','_',t).lower().strip('_')[:20]

# placeholder text in inputs

for m in re.finditer(r'placeholder="([^"]+)"', content):

t = html_mod.unescape(m.group(1))

if len(t) >= 2 and t not in texts:

texts[t] = 'placeholder_' + re.sub(r'[^a-zA-Z0-9]+','_',t[:10]).lower().strip('_')

# Other notable text spans

# p with specific class patterns

for m in re.finditer(r'<(p|span)[^>]*>([^<]{2,40})</\1>', content):

t = m.group(2).strip()

if len(t) >= 2 and t not in texts and not t.startswith('<'):

texts[t] = re.sub(r'[^a-zA-Z0-9]+','_',t[:15]).lower().strip('_')

return texts

def make_translations(page_title, meta_desc, ui_texts):

"""Build translations dict for 4 languages."""

tool_name = re.sub(r'\s*[-–—|].*$', '', page_title).strip()

if not tool_name:

tool_name = page_title

zh = {'pageTitle': page_title, 'metaDesc': meta_desc}

en = {'pageTitle': tool_name + ' - Free Online Tool | ZenTools', 'metaDesc': 'Free online tool. Process locally in your browser.'}

ja = {'pageTitle': tool_name + ' - 無料オンラインツール | ZenTools', 'metaDesc': 'ブラウザで動作する無料オンラインツール。'}

vi = {'pageTitle': tool_name + ' - Công cụ Trực tuyến Miễn phí | ZenTools', 'metaDesc': 'Công cụ trực tuyến miễn phí.'}

for zh_text, key in ui_texts.items():

zh[key] = zh_text

en[key] = zh_text # placeholder - human review needed

ja[key] = zh_text

vi[key] = zh_text

return {'zh': zh, 'en': en, 'ja': ja, 'vi': vi}

def format_translations_js(translations):

"""Format translations as compact JS object."""

langs = {}

for lang, data in translations.items():

parts = []

for k, v in data.items():

escaped = v.replace('\\', '\\\\').replace("'", "\\'").replace('\n', '\\n').replace('\r', '')

parts.append(f"{k}: '{escaped}'")

langs[lang] = '{' + ', '.join(parts) + '}'

entries = ', '.join(f'{lang}: {obj}' for lang, obj in langs.items())

return 'var pageTranslations={' + entries + '};'

def inject_i18n(fpath):

with open(fpath, 'r', encoding='utf-8') as f:

content = f.read()

if has_i18n(content):

return False

page_title = find_title(content)

meta_desc = find_meta(content, 'description')

ui_texts = extract_ui_texts(content)

translations = make_translations(page_title, meta_desc, ui_texts)

trans_js = format_translations_js(translations)

slug = os.path.splitext(os.path.basename(fpath))[0]

# CSS for lang select

lang_css = '''

.lang-select{background:rgba(255,255,255,0.08);color:#fff;border:1px solid var(--border,rgba(255,255,255,0.12));border-radius:8px;padding:8px 10px;font-size:13px;cursor:pointer}.lang-select:focus{outline:none;border-color:var(--accent,#3b82f6)}'''

# Inject CSS into existing style tag or create one

style_match = re.search(r'</style>', content)

if style_match:

content = content[:style_match.start()] + lang_css + '\n' + content[style_match.start():]

else:

content = content.replace('</head>', f'<style>{lang_css}</style>\n</head>', 1)

# Inject translations script before </head>

trans_script = f'<script>{trans_js}</script>'

content = content.replace('</head>', trans_script + '\n</head>', 1)

# Add data-i18n to title tag

content = re.sub(r'<title>', '<title data-i18n="pageTitle">', content, 1)

# Add lang select to header

lang_html = '<select id="langSelect" class="lang-select"><option value="zh">中文</option><option value="ja">日本語</option><option value="en">English</option><option value="vi">Tiếng Việt</option></select>'

header_end_patterns = [

r'(</div>\s*(?:<div class="main|<main))', # header div close before main

r'(</header>\s*)', # header tag close

]

# Try inserting lang select at end of header

for pat in header_end_patterns:

m = re.search(pat, content)

if m:

content = content[:m.start(1)] + lang_html + '\n' + content[m.start(1):]

break

else:

# Check for .actions or .header-right in header

m = re.search(r'(<div class="(?:actions|header-right|header-actions)[^>]*>)', content)

if m:

content = content[:m.end(1)] + '\n' + lang_html + content[m.end(1):]

else:

# Fallback: insert after logo area in header

m = re.search(r'(.logo[^<]*</div>\s*)', content)

if m:

content = content[:m.end(1)] + '\n' + lang_html + '\n' + content[m.end(1):]

# Add data-i18n to elements with known text

for zh_text, key in ui_texts.items():

if key in ['pageTitle', 'metaDesc']:

continue

# Skip if already has data-i18n

escaped = re.escape(zh_text)

# h1/h2/h3

for tag in ['h1','h2','h3']:

content = re.sub(

rf'(<{tag}[^>]*?)>{escaped}</{tag}>',

rf'\1 data-i18n="{key}">{zh_text}</{tag}>',

content, 1

)

# buttons

content = re.sub(

rf'(<button[^>]*?)>{escaped}</button>',

rf'\1 data-i18n="{key}">{zh_text}</button>',

content, 1

)

# labels

content = re.sub(

rf'(<label[^>]*?)>{escaped}</label>',

rf'\1 data-i18n="{key}">{zh_text}</label>',

content, 1

)

# placeholder

content = re.sub(

rf'(placeholder="){escaped}(")',

rf'\1{zh_text}\2 data-i18n-placeholder="{key}"',

content, 1

)

# p, span, div

for tag in ['p','span','div']:

content = re.sub(

rf'(<{tag}[^>]*?)>{escaped}</{tag}>',

rf'\1 data-i18n="{key}">{zh_text}</{tag}>',

content, 1

)

# Add setLanguage function before </body>

i18n_js = '''

<script>

(function(){var t=window.pageTranslations||{};function apply(l){var d=t[l]||t.zh||{};document.documentElement.lang=l;var ti=document.querySelector('title[data-i18n]');if(ti){var pk=ti.getAttribute('data-i18n');if(d[pk])document.title=d[pk];}document.querySelectorAll('[data-i18n]').forEach(function(e){var k=e.getAttribute('data-i18n');if(d[k]&&k!=='pageTitle')e.textContent=d[k];});document.querySelectorAll('[data-i18n-placeholder]').forEach(function(e){var k=e.getAttribute('data-i18n-placeholder');if(d[k])e.placeholder=d[k];});var mc=document.querySelector('meta[name="description"]');if(mc&&d.metaDesc)mc.content=d.metaDesc;}var s=document.getElementById('langSelect');if(s){s.addEventListener('change',function(){apply(this.value);localStorage.setItem('zentools_lang',this.value);});var sv=localStorage.getItem('zentools_lang')||'zh';s.value=sv;apply(sv);}})();

</script>'''

content = content.replace('</body>', i18n_js + '\n</body>', 1)

with open(fpath, 'w', encoding='utf-8') as f:

f.write(content)

return True

# Process all ai/ files

processed = []

skipped = []

for fname in sorted(os.listdir(AI_DIR)):

if not fname.endswith('.html'):

continue

fpath = os.path.join(AI_DIR, fname)

if inject_i18n(fpath):

processed.append(fname)

else:

skipped.append(fname)

print(f"Processed: {len(processed)} files")

for f in processed:

print(f" {f}")

print(f"\nSkipped (already has i18n): {len(skipped)} files")

for f in skipped:

print(f" {f}")