1import { GENERIC_FONTS } from '../../shared/constants.mjs';
2import { checkSourceDesignSystem } from '../../design-system.mjs';
3import { isFullPage } from '../../shared/page.mjs';
4import { finding } from '../../findings.mjs';
5import { filterByProviders } from '../../registry/antipatterns.mjs';
6import { profileFindings, profileStep } from '../../profile/profiler.mjs';
7 
8// ---------------------------------------------------------------------------
9// Regex fallback (non-HTML files: CSS, JSX, TSX, etc.)
10// ---------------------------------------------------------------------------
11 
12const hasRounded = (line) => /\brounded(?:-\w+)?\b/.test(line);
13const hasBorderRadius = (line) => /border-radius/i.test(line);
14const isSafeElement = (line) => /<(?:blockquote|nav[\s>]|pre[\s>]|code[\s>]|a\s|input[\s>]|span[\s>])/i.test(line);
15 
16/** Strip HTML to plain text — drops script/style/comments/tags so
17 *  content-text analyzers don't false-positive on code or CSS. */
18function stripHtmlToText(html) {
19  return html
20    .replace(/<script\b[^>]*>[\s\S]*?<\/script>/gi, ' ')
21    .replace(/<style\b[^>]*>[\s\S]*?<\/style>/gi, ' ')
22    .replace(/<!--[\s\S]*?-->/g, ' ')
23    .replace(/<[^>]+>/g, ' ')
24    .replace(/\s+/g, ' ');
25}
26 
27const PAGE_ANALYZER_EXTS = new Set(['.html', '.htm', '.astro', '.vue', '.svelte']);
28 
29function extFromFilePath(filePath) {
30  return filePath ? (filePath.match(/\.\w+$/)?.[0] || '').toLowerCase() : '';
31}
32 
33function shouldRunPageAnalyzers(content, filePath) {
34  if (!isFullPage(content)) return false;
35  const ext = extFromFilePath(filePath);
36  return !ext || PAGE_ANALYZER_EXTS.has(ext);
37}
38 
39function isNeutralBorderColor(str) {
40  const m = str.match(/solid\s+(#[0-9a-f]{3,8}|rgba?\([^)]+\)|\w+)/i);
41  if (!m) return false;
42  const c = m[1].toLowerCase();
43  if (['gray', 'grey', 'silver', 'white', 'black', 'transparent', 'currentcolor'].includes(c)) return true;
44  const hex = c.match(/^#([0-9a-f]{2})([0-9a-f]{2})([0-9a-f]{2})$/);
45  if (hex) {
46    const [r, g, b] = [parseInt(hex[1], 16), parseInt(hex[2], 16), parseInt(hex[3], 16)];
47    return (Math.max(r, g, b) - Math.min(r, g, b)) < 30;
48  }
49  const shex = c.match(/^#([0-9a-f])([0-9a-f])([0-9a-f])$/);
50  if (shex) {
51    const [r, g, b] = [parseInt(shex[1] + shex[1], 16), parseInt(shex[2] + shex[2], 16), parseInt(shex[3] + shex[3], 16)];
52    return (Math.max(r, g, b) - Math.min(r, g, b)) < 30;
53  }
54  return false;
55}
56 
57const REGEX_MATCHERS = [
58  // --- Side-tab ---
59  { id: 'side-tab', regex: /\bborder-[lrse]-(\d+)\b/g,
60    test: (m, line) => { const n = +m[1]; return hasRounded(line) ? n >= 1 : n >= 4; },
61    fmt: (m) => m[0] },
62  { id: 'side-tab', regex: /border-(?:left|right)\s*:\s*(\d+)px\s+solid[^;]*/gi,
63    test: (m, line) => { if (isSafeElement(line)) return false; if (isNeutralBorderColor(m[0])) return false; const n = +m[1]; return hasBorderRadius(line) ? n >= 1 : n >= 3; },
64    fmt: (m) => m[0].replace(/\s*;?\s*$/, '') },
65  { id: 'side-tab', regex: /border-(?:left|right)-width\s*:\s*(\d+)px/gi,
66    test: (m, line) => !isSafeElement(line) && +m[1] >= 3,
67    fmt: (m) => m[0] },
68  { id: 'side-tab', regex: /border-inline-(?:start|end)\s*:\s*(\d+)px\s+solid/gi,
69    test: (m, line) => !isSafeElement(line) && +m[1] >= 3,
70    fmt: (m) => m[0] },
71  { id: 'side-tab', regex: /border-inline-(?:start|end)-width\s*:\s*(\d+)px/gi,
72    test: (m, line) => !isSafeElement(line) && +m[1] >= 3,
73    fmt: (m) => m[0] },
74  { id: 'side-tab', regex: /border(?:Left|Right)\s*[:=]\s*["'`](\d+)px\s+solid/g,
75    test: (m) => +m[1] >= 3,
76    fmt: (m) => m[0] },
77  // --- Border accent on rounded ---
78  { id: 'border-accent-on-rounded', regex: /\bborder-[tb]-(\d+)\b/g,
79    test: (m, line) => hasRounded(line) && +m[1] >= 1,
80    fmt: (m) => m[0] },
81  { id: 'border-accent-on-rounded', regex: /border-(?:top|bottom)\s*:\s*(\d+)px\s+solid/gi,
82    test: (m, line) => +m[1] >= 3 && hasBorderRadius(line),
83    fmt: (m) => m[0] },
84  // --- Overused font ---
85  { id: 'overused-font', regex: /font-family\s*:\s*['"]?(Inter|Roboto|Open Sans|Lato|Montserrat|Arial|Helvetica|Fraunces|Geist Sans|Geist Mono|Geist|Mona Sans|Plus Jakarta Sans|Space Grotesk|Recoleta|Instrument Sans|Instrument Serif)\b/gi,
86    test: () => true,
87    fmt: (m) => m[0] },
88  { id: 'overused-font', regex: /fonts\.googleapis\.com\/css2?\?family=(Inter|Roboto|Open\+Sans|Lato|Montserrat|Fraunces|Plus\+Jakarta\+Sans|Space\+Grotesk|Instrument\+Sans|Instrument\+Serif|Mona\+Sans|Geist)\b/gi,
89    test: () => true,
90    fmt: (m) => `Google Fonts: ${m[1].replace(/\+/g, ' ')}` },
91  // --- Gradient text ---
92  { id: 'gradient-text', regex: /background-clip\s*:\s*text|-webkit-background-clip\s*:\s*text/gi,
93    test: (m, line) => /gradient/i.test(line),
94    fmt: () => 'background-clip: text + gradient' },
95  // --- Gradient text (Tailwind) ---
96  { id: 'gradient-text', regex: /\bbg-clip-text\b/g,
97    test: (m, line) => /\bbg-gradient-to-/i.test(line),
98    fmt: () => 'bg-clip-text + bg-gradient' },
99  // --- Tailwind gray on colored bg ---
100  { id: 'gray-on-color', regex: /\btext-(?:gray|slate|zinc|neutral|stone)-(\d+)\b/g,
101    test: (m, line) => /\bbg-(?:red|orange|amber|yellow|lime|green|emerald|teal|cyan|sky|blue|indigo|violet|purple|fuchsia|pink|rose)-\d+\b/.test(line),
102    fmt: (m, line) => { const bg = line.match(/\bbg-(?:red|orange|amber|yellow|lime|green|emerald|teal|cyan|sky|blue|indigo|violet|purple|fuchsia|pink|rose)-\d+\b/); return `${m[0]} on ${bg?.[0] || '?'}`; } },
103  // --- Tailwind AI palette ---
104  { id: 'ai-color-palette', regex: /\btext-(?:purple|violet|indigo)-(\d+)\b/g,
105    test: (m, line) => /\btext-(?:[2-9]xl|[3-9]xl)\b|<h[1-3]/i.test(line),
106    fmt: (m) => `${m[0]} on heading` },
107  { id: 'ai-color-palette', regex: /\bfrom-(?:purple|violet|indigo)-(\d+)\b/g,
108    test: (m, line) => /\bto-(?:purple|violet|indigo|blue|cyan|pink|fuchsia)-\d+\b/.test(line),
109    fmt: (m) => `${m[0]} gradient` },
110  // --- Bounce/elastic easing ---
111  { id: 'bounce-easing', regex: /\banimate-bounce\b/g,
112    test: () => true,
113    fmt: () => 'animate-bounce (Tailwind)' },
114  { id: 'bounce-easing', regex: /animation(?:-name)?\s*:\s*([^;{}]*(?:bounce|elastic|wobble|jiggle|spring)[^;{}]*)/gi,
115    test: () => true,
116    fmt: (m) => {
117      const token = m[1]
118        .split(/[,\s]+/)
119        .find((part) => /bounce|elastic|wobble|jiggle|spring/i.test(part));
120      return `animation: ${token || m[1].trim()}`;
121    } },
122  { id: 'bounce-easing', regex: /cubic-bezier\(\s*([\d.-]+)\s*,\s*([\d.-]+)\s*,\s*([\d.-]+)\s*,\s*([\d.-]+)\s*\)/g,
123    test: (m) => {
124      const y1 = parseFloat(m[2]), y2 = parseFloat(m[4]);
125      return y1 < -0.1 || y1 > 1.1 || y2 < -0.1 || y2 > 1.1;
126    },
127    fmt: (m) => `cubic-bezier(${m[1]}, ${m[2]}, ${m[3]}, ${m[4]})` },
128  // --- Layout property transition ---
129  { id: 'layout-transition', regex: /transition\s*:\s*([^;{}]+)/gi,
130    test: (m) => {
131      const val = m[1].toLowerCase();
132      if (/\ball\b/.test(val)) return false;
133      return /\b(?:(?:max|min)-)?(?:width|height)\b|\bpadding\b|\bmargin\b/.test(val);
134    },
135    fmt: (m) => {
136      const found = m[1].match(/\b(?:(?:max|min)-)?(?:width|height)\b|\bpadding(?:-(?:top|right|bottom|left))?\b|\bmargin(?:-(?:top|right|bottom|left))?\b/gi);
137      return `transition: ${found ? found.join(', ') : m[1].trim()}`;
138    } },
139  { id: 'layout-transition', regex: /transition-property\s*:\s*([^;{}]+)/gi,
140    test: (m) => {
141      const val = m[1].toLowerCase();
142      if (/\ball\b/.test(val)) return false;
143      return /\b(?:(?:max|min)-)?(?:width|height)\b|\bpadding\b|\bmargin\b/.test(val);
144    },
145    fmt: (m) => {
146      const found = m[1].match(/\b(?:(?:max|min)-)?(?:width|height)\b|\bpadding(?:-(?:top|right|bottom|left))?\b|\bmargin(?:-(?:top|right|bottom|left))?\b/gi);
147      return `transition-property: ${found ? found.join(', ') : m[1].trim()}`;
148    } },
149  // --- Broken image: src="" or src="#" or src=" " ---
150  { id: 'broken-image', regex: /<img\b[^>]*?\bsrc\s*=\s*(?:""|''|"\s+"|'\s+'|"#"|'#')/gi,
151    test: () => true,
152    fmt: (m) => m[0].slice(0, 100) },
153  // --- Broken image: <img> with no src attribute at all ---
154  { id: 'broken-image', regex: /<img\b(?:(?!\bsrc\s*=)[^>])*>/gi,
155    test: (m) => !/\bsrc\s*=/i.test(m[0]),
156    fmt: (m) => m[0].slice(0, 100) },
157];
158 
159const REGEX_ANALYZERS = [
160  // Single font
161  (content, filePath) => {
162    const fontFamilyRe = /font-family\s*:\s*([^;}]+)/gi;
163    const fonts = new Set();
164    let m;
165    while ((m = fontFamilyRe.exec(content)) !== null) {
166      for (const f of m[1].split(',').map(f => f.trim().replace(/^['"]|['"]$/g, '').toLowerCase())) {
167        if (f && !GENERIC_FONTS.has(f)) fonts.add(f);
168      }
169    }
170    const gfRe = /fonts\.googleapis\.com\/css2?\?family=([^&"'\s]+)/gi;
171    while ((m = gfRe.exec(content)) !== null) {
172      for (const f of m[1].split('|').map(f => f.split(':')[0].replace(/\+/g, ' ').toLowerCase())) fonts.add(f);
173    }
174    if (fonts.size !== 1 || content.split('\n').length < 20) return [];
175    const name = [...fonts][0];
176    const lines = content.split('\n');
177    let line = 1;
178    for (let i = 0; i < lines.length; i++) { if (lines[i].toLowerCase().includes(name)) { line = i + 1; break; } }
179    return [finding('single-font', filePath, `only font used is ${name}`, line)];
180  },
181  // Flat type hierarchy
182  (content, filePath) => {
183    const sizes = new Set();
184    const REM = 16;
185    let m;
186    const sizeRe = /font-size\s*:\s*([\d.]+)(px|rem|em)\b/gi;
187    while ((m = sizeRe.exec(content)) !== null) {
188      const px = m[2] === 'px' ? +m[1] : +m[1] * REM;
189      if (px > 0 && px < 200) sizes.add(Math.round(px * 10) / 10);
190    }
191    const clampRe = /font-size\s*:\s*clamp\(\s*([\d.]+)(px|rem|em)\s*,\s*[^,]+,\s*([\d.]+)(px|rem|em)\s*\)/gi;
192    while ((m = clampRe.exec(content)) !== null) {
193      sizes.add(Math.round((m[2] === 'px' ? +m[1] : +m[1] * REM) * 10) / 10);
194      sizes.add(Math.round((m[4] === 'px' ? +m[3] : +m[3] * REM) * 10) / 10);
195    }
196    const TW = { 'text-xs': 12, 'text-sm': 14, 'text-base': 16, 'text-lg': 18, 'text-xl': 20, 'text-2xl': 24, 'text-3xl': 30, 'text-4xl': 36, 'text-5xl': 48, 'text-6xl': 60, 'text-7xl': 72, 'text-8xl': 96, 'text-9xl': 128 };
197    for (const [cls, px] of Object.entries(TW)) { if (new RegExp(`\\b${cls}\\b`).test(content)) sizes.add(px); }
198    if (sizes.size < 3) return [];
199    const sorted = [...sizes].sort((a, b) => a - b);
200    const ratio = sorted[sorted.length - 1] / sorted[0];
201    if (ratio >= 2.0) return [];
202    const lines = content.split('\n');
203    let line = 1;
204    for (let i = 0; i < lines.length; i++) { if (/font-size/i.test(lines[i]) || /\btext-(?:xs|sm|base|lg|xl|\d)/i.test(lines[i])) { line = i + 1; break; } }
205    return [finding('flat-type-hierarchy', filePath, `Sizes: ${sorted.map(s => s + 'px').join(', ')} (ratio ${ratio.toFixed(1)}:1)`, line)];
206  },
207  // Monotonous spacing (regex)
208  (content, filePath) => {
209    const vals = [];
210    let m;
211    const pxRe = /(?:padding|margin)(?:-(?:top|right|bottom|left))?\s*:\s*(\d+)px/gi;
212    while ((m = pxRe.exec(content)) !== null) { const v = +m[1]; if (v > 0 && v < 200) vals.push(v); }
213    const remRe = /(?:padding|margin)(?:-(?:top|right|bottom|left))?\s*:\s*([\d.]+)rem/gi;
214    while ((m = remRe.exec(content)) !== null) { const v = Math.round(parseFloat(m[1]) * 16); if (v > 0 && v < 200) vals.push(v); }
215    const gapRe = /gap\s*:\s*(\d+)px/gi;
216    while ((m = gapRe.exec(content)) !== null) vals.push(+m[1]);
217    const twRe = /\b(?:p|px|py|pt|pb|pl|pr|m|mx|my|mt|mb|ml|mr|gap)-(\d+)\b/g;
218    while ((m = twRe.exec(content)) !== null) vals.push(+m[1] * 4);
219    const rounded = vals.map(v => Math.round(v / 4) * 4);
220    if (rounded.length < 10) return [];
221    const counts = {};
222    for (const v of rounded) counts[v] = (counts[v] || 0) + 1;
223    const maxCount = Math.max(...Object.values(counts));
224    const pct = maxCount / rounded.length;
225    const unique = [...new Set(rounded)].filter(v => v > 0);
226    if (pct <= 0.6 || unique.length > 3) return [];
227    const dominant = Object.entries(counts).sort((a, b) => b[1] - a[1])[0][0];
228    return [finding('monotonous-spacing', filePath, `~${dominant}px used ${maxCount}/${rounded.length} times (${Math.round(pct * 100)}%)`)];
229  },
230  // Em-dash overuse: 5+ em-dashes or "--" in body text content
231  // (occasional em-dash use in prose is fine; the pattern fires only
232  // when count crosses into AI-cadence territory).
233  (content, filePath) => {
234    const text = stripHtmlToText(content);
235    let count = 0;
236    const re = /[—]|--(?=\S)/g;
237    while (re.exec(text) !== null) count++;
238    if (count < 5) return [];
239    return [finding('em-dash-overuse', filePath, `${count} em-dashes in body text`)];
240  },
241  // Marketing buzzwords: SaaS phrase list
242  (content, filePath) => {
243    const text = stripHtmlToText(content);
244    const lower = text.toLowerCase();
245    const BUZZWORDS = [
246      'streamline your', 'empower your', 'supercharge your',
247      'unleash your', 'unleash the power', 'leverage the power',
248      'built for the modern', 'trusted by leading', 'trusted by the world',
249      'best-in-class', 'industry-leading', 'world-class', 'enterprise-grade',
250      'next-generation', 'cutting-edge', 'transform your business',
251      'revolutionize', 'game-changer', 'game changing',
252      'mission-critical', 'best of breed', 'future-proof', 'future proof',
253      'seamless experience', 'seamlessly integrate',
254      'drive engagement', 'drive growth', 'drive results',
255      'harness the power',
256    ];
257    let count = 0;
258    let firstSample = '';
259    for (const phrase of BUZZWORDS) {
260      let from = 0;
261      while (true) {
262        const idx = lower.indexOf(phrase, from);
263        if (idx === -1) break;
264        count++;
265        if (!firstSample) {
266          firstSample = text.slice(Math.max(0, idx - 12), Math.min(text.length, idx + phrase.length + 12)).trim();
267        }
268        from = idx + phrase.length;
269      }
270    }
271    if (count === 0) return [];
272    return [finding('marketing-buzzword', filePath, `${count} buzzword phrase${count === 1 ? '' : 's'}: "${firstSample}"`)];
273  },
274  // Numbered section markers (01 / 02 / 03 ...)
275  (content, filePath) => {
276    const text = stripHtmlToText(content);
277    const re = /\b(0[1-9]|1[0-2])\b/g;
278    const seen = new Set();
279    let m;
280    while ((m = re.exec(text)) !== null) seen.add(m[1]);
281    if (seen.size < 3) return [];
282    const sorted = [...seen].sort();
283    let sequential = 0;
284    for (let i = 1; i < sorted.length; i++) {
285      if (parseInt(sorted[i], 10) === parseInt(sorted[i - 1], 10) + 1) sequential++;
286    }
287    if (sequential < 2) return [];
288    return [finding('numbered-section-markers', filePath, `Sequence: ${sorted.slice(0, 6).join(', ')}`)];
289  },
290  // Aphoristic cadence: manufactured-contrast + short-rebuttal
291  (content, filePath) => {
292    const text = stripHtmlToText(content);
293    const NOT_A_RE = /\bNot an? [a-z][^.!?]{1,40}[.!]\s+[A-Z][^.!?]{1,60}[.!]/g;
294    const SHORT_REBUTTAL_RE = /\b[A-Z][^.!?]{4,80}[.!]\s+(No|Just)\s+[a-z][^.!?]{2,60}[.!]/g;
295    let count = 0;
296    let firstSample = '';
297    let m;
298    NOT_A_RE.lastIndex = 0;
299    while ((m = NOT_A_RE.exec(text)) !== null) {
300      count++;
301      if (!firstSample) firstSample = m[0].trim().slice(0, 80);
302    }
303    SHORT_REBUTTAL_RE.lastIndex = 0;
304    while ((m = SHORT_REBUTTAL_RE.exec(text)) !== null) {
305      count++;
306      if (!firstSample) firstSample = m[0].trim().slice(0, 80);
307    }
308    if (count < 3) return [];
309    return [finding('aphoristic-cadence', filePath, `${count} aphoristic constructions: "${firstSample}"`)];
310  },
311  // Dark glow (page-level: dark bg + colored box-shadow with blur)
312  (content, filePath) => {
313    // Check if page has a dark background
314    const darkBgRe = /background(?:-color)?\s*:\s*(?:#(?:0[0-9a-f]|1[0-9a-f]|2[0-3])[0-9a-f]{4}\b|#(?:0|1)[0-9a-f]{2}\b|rgb\(\s*(\d{1,2})\s*,\s*(\d{1,2})\s*,\s*(\d{1,2})\s*\))/gi;
315    const twDarkBg = /\bbg-(?:gray|slate|zinc|neutral|stone)-(?:9\d{2}|800)\b/;
316    const hasDarkBg = darkBgRe.test(content) || twDarkBg.test(content);
317    if (!hasDarkBg) return [];
318 
319    // Check for colored box-shadow with blur > 4px
320    const shadowRe = /box-shadow\s*:\s*([^;{}]+)/gi;
321    let m;
322    while ((m = shadowRe.exec(content)) !== null) {
323      const val = m[1];
324      const colorMatch = val.match(/rgba?\(\s*(\d+)\s*,\s*(\d+)\s*,\s*(\d+)/);
325      if (!colorMatch) continue;
326      const [r, g, b] = [+colorMatch[1], +colorMatch[2], +colorMatch[3]];
327      if ((Math.max(r, g, b) - Math.min(r, g, b)) < 30) continue; // skip gray
328      // Check blur: look for pattern like "0 0 20px" (third number > 4)
329      const pxVals = [...val.matchAll(/(\d+)px|(?<![.\d])\b(0)\b(?![.\d])/g)].map(p => +(p[1] || p[2]));
330      if (pxVals.length >= 3 && pxVals[2] > 4) {
331        const lines = content.substring(0, m.index).split('\n');
332        return [finding('dark-glow', filePath, `Colored glow (rgb(${r},${g},${b})) on dark page`, lines.length)];
333      }
334    }
335    return [];
336  },
337];
338 
339// ---------------------------------------------------------------------------
340// Style block extraction (Vue/Svelte <style> blocks)
341// ---------------------------------------------------------------------------
342 
343function extractStyleBlocks(content, ext) {
344  ext = ext.toLowerCase();
345  if (ext !== '.vue' && ext !== '.svelte') return [];
346  const blocks = [];
347  const re = /<style[^>]*>([\s\S]*?)<\/style>/gi;
348  let m;
349  while ((m = re.exec(content)) !== null) {
350    const before = content.substring(0, m.index);
351    const startLine = before.split('\n').length + 1;
352    blocks.push({ content: m[1], startLine });
353  }
354  return blocks;
355}
356 
357// ---------------------------------------------------------------------------
358// CSS-in-JS extraction (styled-components, emotion)
359// ---------------------------------------------------------------------------
360 
361const CSS_IN_JS_EXTENSIONS = new Set(['.js', '.ts', '.jsx', '.tsx']);
362 
363function extractCSSinJS(content, ext) {
364  ext = ext.toLowerCase();
365  if (!CSS_IN_JS_EXTENSIONS.has(ext)) return [];
366  const blocks = [];
367  const re = /(?:styled(?:\.\w+|\([^)]+\))|css)\s*`([\s\S]*?)`/g;
368  let m;
369  while ((m = re.exec(content)) !== null) {
370    const before = content.substring(0, m.index);
371    const startLine = before.split('\n').length;
372    blocks.push({ content: m[1], startLine });
373  }
374  return blocks;
375}
376 
377function runRegexMatchers(lines, filePath, lineOffset = 0, blockContext = null, options = {}) {
378  const { profile, phase = 'regex-matchers' } = options || {};
379  const findings = [];
380  if (!profile) {
381    for (const matcher of REGEX_MATCHERS) {
382      for (let i = 0; i < lines.length; i++) {
383        const line = lines[i];
384        matcher.regex.lastIndex = 0;
385        let m;
386        while ((m = matcher.regex.exec(line)) !== null) {
387          // For extracted blocks, use nearby lines as context for multi-line CSS patterns
388          const context = blockContext
389            ? lines.slice(Math.max(0, i - 3), Math.min(lines.length, i + 4)).join(' ')
390            : line;
391          if (matcher.test(m, context)) {
392            findings.push(finding(matcher.id, filePath, matcher.fmt(m, context), i + 1 + lineOffset));
393          }
394        }
395      }
396    }
397    return findings;
398  }
399 
400  for (const matcher of REGEX_MATCHERS) {
401    const matcherFindings = profileFindings(profile, {
402      engine: 'regex',
403      phase,
404      ruleId: matcher.id,
405      target: filePath,
406    }, () => {
407      const matches = [];
408      for (let i = 0; i < lines.length; i++) {
409        const line = lines[i];
410        matcher.regex.lastIndex = 0;
411        let m;
412        while ((m = matcher.regex.exec(line)) !== null) {
413          // For extracted blocks, use nearby lines as context for multi-line CSS patterns
414          const context = blockContext
415            ? lines.slice(Math.max(0, i - 3), Math.min(lines.length, i + 4)).join(' ')
416            : line;
417          if (matcher.test(m, context)) {
418            matches.push(finding(matcher.id, filePath, matcher.fmt(m, context), i + 1 + lineOffset));
419          }
420        }
421      }
422      return matches;
423    });
424    findings.push(...matcherFindings);
425  }
426  return findings;
427}
428 
429/** Page-level analyzers that scan rendered text content (em-dash use,
430 *  buzzword phrases, numbered section markers, aphoristic cadence).
431 *  These are detector-agnostic — they work on any HTML/text source
432 *  and don't need a parsed DOM. Exported so detectHtml can call them
433 *  for `.html` files (which otherwise skip the regex engine). */
434const TEXT_CONTENT_ANALYZER_IDS = [
435  'em-dash-overuse',
436  'marketing-buzzword',
437  'numbered-section-markers',
438  'aphoristic-cadence',
439];
440 
441function runTextContentAnalyzers(content, filePath, options = {}) {
442  const profile = options?.profile;
443  if (!shouldRunPageAnalyzers(content, filePath)) return [];
444  // The 4 text-content analyzers are at indices 3-6 in REGEX_ANALYZERS.
445  const findings = [];
446  for (let i = 0; i < TEXT_CONTENT_ANALYZER_IDS.length; i++) {
447    const analyzer = REGEX_ANALYZERS[3 + i];
448    const ruleId = TEXT_CONTENT_ANALYZER_IDS[i];
449    findings.push(...profileFindings(profile, {
450      engine: 'regex',
451      phase: 'text-content',
452      ruleId,
453      target: filePath,
454    }, () => analyzer(content, filePath)));
455  }
456  return findings;
457}
458 
459function detectText(content, filePath, options = {}) {
460  const profile = options?.profile;
461  const findings = [];
462  const lines = content.split('\n');
463  const ext = extFromFilePath(filePath);
464 
465  // Run regex matchers on the full file content (catches Tailwind classes, inline styles)
466  // Enable block context for CSS files where related properties span multiple lines
467  const cssLike = new Set(['.css', '.scss', '.sass', '.less']);
468  findings.push(...runRegexMatchers(lines, filePath, 0, cssLike.has(ext) || null, {
469    profile,
470    phase: 'source',
471  }));
472 
473  // Extract and scan <style> blocks from Vue/Svelte SFCs
474  const styleBlocks = profile
475    ? profileStep(profile, {
476      engine: 'regex',
477      phase: 'extract',
478      ruleId: 'style-blocks',
479      target: filePath,
480    }, () => extractStyleBlocks(content, ext))
481    : extractStyleBlocks(content, ext);
482  for (const block of styleBlocks) {
483    const blockLines = block.content.split('\n');
484    findings.push(...runRegexMatchers(blockLines, filePath, block.startLine - 1, true, {
485      profile,
486      phase: 'style-block',
487    }));
488  }
489 
490  // Extract and scan CSS-in-JS template literals
491  const cssJsBlocks = profile
492    ? profileStep(profile, {
493      engine: 'regex',
494      phase: 'extract',
495      ruleId: 'css-in-js',
496      target: filePath,
497    }, () => extractCSSinJS(content, ext))
498    : extractCSSinJS(content, ext);
499  for (const block of cssJsBlocks) {
500    const blockLines = block.content.split('\n');
501    findings.push(...runRegexMatchers(blockLines, filePath, block.startLine - 1, true, {
502      profile,
503      phase: 'css-in-js',
504    }));
505  }
506 
507  if (options?.designSystem) {
508    findings.push(...profileFindings(profile, {
509      engine: 'regex',
510      phase: 'source',
511      ruleId: 'design-system',
512      target: filePath,
513    }, () => checkSourceDesignSystem(content, filePath, { designSystem: options.designSystem })));
514  }
515 
516  // Deduplicate findings (same antipattern + similar snippet, within 2 lines)
517  const deduped = [];
518  for (const f of findings) {
519    const isDupe = deduped.some(d =>
520      d.antipattern === f.antipattern &&
521      d.snippet === f.snippet &&
522      Math.abs(d.line - f.line) <= 2
523    );
524    if (!isDupe) deduped.push(f);
525  }
526 
527  // Page-level analyzers only run on full pages
528  if (shouldRunPageAnalyzers(content, filePath)) {
529    const analyzerIds = [
530      'single-font',
531      'flat-type-hierarchy',
532      'monotonous-spacing',
533      'em-dash-overuse',
534      'marketing-buzzword',
535      'numbered-section-markers',
536      'aphoristic-cadence',
537      'dark-glow',
538    ];
539    for (let i = 0; i < REGEX_ANALYZERS.length; i++) {
540      const analyzer = REGEX_ANALYZERS[i];
541      deduped.push(...profileFindings(profile, {
542        engine: 'regex',
543        phase: 'page-analyzer',
544        ruleId: analyzerIds[i] || `analyzer-${i + 1}`,
545        target: filePath,
546      }, () => analyzer(content, filePath)));
547    }
548  }
549 
550  return filterByProviders(deduped, options?.providers);
551}
552 
553export {
554  REGEX_MATCHERS,
555  REGEX_ANALYZERS,
556  TEXT_CONTENT_ANALYZER_IDS,
557  extractStyleBlocks,
558  extractCSSinJS,
559  runRegexMatchers,
560  runTextContentAnalyzers,
561  detectText,
562};
563