@redocly/recheck 0.14.0 → 2.58.0

package/lib/core/files.js

@@ -0,0 +1,184 @@
+import * as fs from 'fs/promises';
+import * as path from 'path';
+import picomatch from 'picomatch';
+import { parseMarkdown } from '../parser/index.js';
+import { getImageDestinations } from '../rules/token/helpers.js';
+const SKIP_DIRS = ['node_modules', 'dist', 'build'];
+export async function discoverMarkdownFiles(inputPath) {
+    async function walkDir(dir) {
+        const files = [];
+        try {
+            const entries = await fs.readdir(dir, { withFileTypes: true });
+            for (const entry of entries) {
+                const fullPath = path.join(dir, entry.name);
+                if (entry.isDirectory()) {
+                    if (entry.name.startsWith('.') || SKIP_DIRS.includes(entry.name)) {
+                        continue;
+                    }
+                    files.push(...(await walkDir(fullPath)));
+                }
+                else if (entry.isFile() &&
+                    picomatch.isMatch(entry.name, '*.{md,markdown}', { nocase: true })) {
+                    files.push(fullPath);
+                }
+            }
+        }
+        catch {
+            return [];
+        }
+        return files;
+    }
+    const stat = await fs.stat(inputPath);
+    if (stat.isFile()) {
+        return [inputPath];
+    }
+    else if (stat.isDirectory()) {
+        return walkDir(inputPath);
+    }
+    else {
+        return [];
+    }
+}
+/**
+ * Load list of changed files from either a file path or stdin
+ */
+export async function loadChangedFiles(changedListPath) {
+    if (changedListPath && changedListPath.length > 0) {
+        try {
+            const content = await fs.readFile(changedListPath, 'utf8');
+            return content
+                .split(/\r?\n/)
+                .map((s) => s.trim())
+                .filter(Boolean);
+        }
+        catch (_err) {
+            return [];
+        }
+    }
+    if (process.stdin.isTTY)
+        return [];
+    const chunks = [];
+    return await new Promise((resolve) => {
+        process.stdin.setEncoding('utf8');
+        process.stdin.on('data', (d) => chunks.push(String(d)));
+        process.stdin.on('end', () => {
+            const text = chunks.join('');
+            const lines = text
+                .split(/\r?\n/)
+                .map((s) => s.trim())
+                .filter(Boolean);
+            resolve(lines);
+        });
+        process.stdin.resume();
+    });
+}
+/**
+ * Runs `fn` over `items` with at most `limit` calls at once. Results are in the same order as
+ * `items`. Recheck is published on its own, so it cannot use @redocly/shared's promiseMapLimit.
+ */
+export async function mapLimit(items, limit, fn) {
+    const results = new Array(items.length);
+    let next = 0;
+    async function worker() {
+        while (next < items.length) {
+            const index = next++;
+            results[index] = await fn(items[index], index);
+        }
+    }
+    await Promise.all(Array.from({ length: Math.min(limit, items.length) }, worker));
+    return results;
+}
+/** Whether any rule uses `max-image-size`, which needs image metadata from disk. */
+export function needsImageMetadata(rules) {
+    return rules.some((rule) => Object.keys(rule.assertions).includes('max-image-size'));
+}
+/**
+ * Returns the local (not `http(s)://`) image paths in `content`, from inline and reference-style
+ * images. It uses the same `getImageDestinations` as `max-image-size`, so the paths match the
+ * keys that `loadImageMetadata` stores. Markdoc parsing is off here, because it does not change
+ * where images start or end.
+ */
+function extractImageReferences(content) {
+    return getImageDestinations(parseMarkdown(content))
+        .map(({ destination }) => destination)
+        .filter((destination) => !destination.startsWith('http://') && !destination.startsWith('https://'));
+}
+/**
+ * Most unique image refs checked per file, to limit the number of `fs.stat` calls. Refs past the
+ * limit are left out of the metadata instead of being marked as missing.
+ */
+export const MAX_IMAGE_REFS_PER_FILE = 1000;
+/** Whether `resolved` (an absolute path) is `root` itself or inside it. */
+function isInsideRoot(root, resolved) {
+    const relative = path.relative(root, resolved);
+    return (relative === '' ||
+        (relative !== '..' && !relative.startsWith(`..${path.sep}`) && !path.isAbsolute(relative)));
+}
+/**
+ * Loads size and existence for the images in a markdown file, keyed by the path as written.
+ *
+ * Refs are limited to `root` (default `process.cwd()`). A ref that leaves it, by `../` or an
+ * absolute path, is marked as missing without being checked, so a document cannot probe other
+ * files. Symlinks are followed only if they resolve inside the root.
+ *
+ * At most `MAX_IMAGE_REFS_PER_FILE` unique refs are checked. The rest are left out, so
+ * `max-image-size` skips them instead of reporting them as missing.
+ */
+export async function loadImageMetadata(file, content, root = process.cwd()) {
+    const images = new Map();
+    const baseDir = path.dirname(file);
+    const resolvedRoot = path.resolve(root);
+    const imageReferences = extractImageReferences(content);
+    // The root's real path, resolved once. `undefined` if the root does not exist.
+    let realRoot;
+    try {
+        realRoot = await fs.realpath(resolvedRoot);
+    }
+    catch {
+        realRoot = undefined;
+    }
+    for (const imagePath of imageReferences) {
+        if (images.has(imagePath)) {
+            continue;
+        }
+        if (images.size >= MAX_IMAGE_REFS_PER_FILE) {
+            break;
+        }
+        const fullPath = path.resolve(baseDir, imagePath);
+        if (!isInsideRoot(resolvedRoot, fullPath)) {
+            images.set(imagePath, { path: imagePath, size: 0, exists: false });
+            continue;
+        }
+        // Follow symlinks, and reject any that end up outside the root. If the path cannot be
+        // resolved, the `fs.stat` below reports it as missing.
+        if (realRoot !== undefined) {
+            try {
+                const realPath = await fs.realpath(fullPath);
+                if (!isInsideRoot(realRoot, realPath)) {
+                    images.set(imagePath, { path: imagePath, size: 0, exists: false });
+                    continue;
+                }
+            }
+            catch {
+                // Fall through to the stat below.
+            }
+        }
+        try {
+            const stats = await fs.stat(fullPath);
+            images.set(imagePath, {
+                path: imagePath,
+                size: stats.size,
+                exists: true,
+            });
+        }
+        catch {
+            images.set(imagePath, {
+                path: imagePath,
+                size: 0,
+                exists: false,
+            });
+        }
+    }
+    return images.size > 0 ? { images } : undefined;
+}
+//# sourceMappingURL=files.js.map

package/lib/core/files.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"files.js","sourceRoot":"","sources":["../../src/core/files.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,MAAM,aAAa,CAAC;AAClC,OAAO,KAAK,IAAI,MAAM,MAAM,CAAC;AAC7B,OAAO,SAAS,MAAM,WAAW,CAAC;AAElC,OAAO,EAAE,aAAa,EAAE,MAAM,oBAAoB,CAAC;AACnD,OAAO,EAAE,oBAAoB,EAAE,MAAM,2BAA2B,CAAC;AAIjE,MAAM,SAAS,GAAG,CAAC,cAAc,EAAE,MAAM,EAAE,OAAO,CAAC,CAAC;AAEpD,MAAM,CAAC,KAAK,UAAU,qBAAqB,CAAC,SAAiB;IAC3D,KAAK,UAAU,OAAO,CAAC,GAAW;QAChC,MAAM,KAAK,GAAa,EAAE,CAAC;QAE3B,IAAI,CAAC;YACH,MAAM,OAAO,GAAG,MAAM,EAAE,CAAC,OAAO,CAAC,GAAG,EAAE,EAAE,aAAa,EAAE,IAAI,EAAE,CAAC,CAAC;YAE/D,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;gBAC5B,MAAM,QAAQ,GAAG,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC;gBAE5C,IAAI,KAAK,CAAC,WAAW,EAAE,EAAE,CAAC;oBACxB,IAAI,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,SAAS,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC;wBACjE,SAAS;oBACX,CAAC;oBACD,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,MAAM,OAAO,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC;gBAC3C,CAAC;qBAAM,IACL,KAAK,CAAC,MAAM,EAAE;oBACd,SAAS,CAAC,OAAO,CAAC,KAAK,CAAC,IAAI,EAAE,iBAAiB,EAAE,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,EAClE,CAAC;oBACD,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;gBACvB,CAAC;YACH,CAAC;QACH,CAAC;QAAC,MAAM,CAAC;YACP,OAAO,EAAE,CAAC;QACZ,CAAC;QAED,OAAO,KAAK,CAAC;IACf,CAAC;IAED,MAAM,IAAI,GAAG,MAAM,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;IACtC,IAAI,IAAI,CAAC,MAAM,EAAE,EAAE,CAAC;QAClB,OAAO,CAAC,SAAS,CAAC,CAAC;IACrB,CAAC;SAAM,IAAI,IAAI,CAAC,WAAW,EAAE,EAAE,CAAC;QAC9B,OAAO,OAAO,CAAC,SAAS,CAAC,CAAC;IAC5B,CAAC;SAAM,CAAC;QACN,OAAO,EAAE,CAAC;IACZ,CAAC;AACH,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,KAAK,UAAU,gBAAgB,CAAC,eAAwB;IAC7D,IAAI,eAAe,IAAI,eAAe,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAClD,IAAI,CAAC;YACH,MAAM,OAAO,GAAG,MAAM,EAAE,CAAC,QAAQ,CAAC,eAAe,EAAE,MAAM,CAAC,CAAC;YAC3D,OAAO,OAAO;iBACX,KAAK,CAAC,OAAO,CAAC;iBACd,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;iBACpB,MAAM,CAAC,OAAO,CAAC,CAAC;QACrB,CAAC;QAAC,OAAO,IAAI,EAAE,CAAC;YACd,OAAO,EAAE,CAAC;QACZ,CAAC;IACH,CAAC;IACD,IAAI,OAAO,CAAC,KAAK,CAAC,KAAK;QAAE,OAAO,EAAE,CAAC;IACnC,MAAM,MAAM,GAAa,EAAE,CAAC;IAC5B,OAAO,MAAM,IAAI,OAAO,CAAW,CAAC,OAAO,EAAE,EAAE;QAC7C,OAAO,CAAC,KAAK,CAAC,WAAW,CAAC,MAAM,CAAC,CAAC;QAClC,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,MAAM,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QACxD,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,KAAK,EAAE,GAAG,EAAE;YAC3B,MAAM,IAAI,GAAG,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YAC7B,MAAM,KAAK,GAAG,IAAI;iBACf,KAAK,CAAC,OAAO,CAAC;iBACd,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;iBACpB,MAAM,CAAC,OAAO,CAAC,CAAC;YACnB,OAAO,CAAC,KAAK,CAAC,CAAC;QACjB,CAAC,CAAC,CAAC;QACH,OAAO,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC;IACzB,CAAC,CAAC,CAAC;AACL,CAAC;AAED;;;GAGG;AACH,MAAM,CAAC,KAAK,UAAU,QAAQ,CAC5B,KAAU,EACV,KAAa,EACb,EAA0C;IAE1C,MAAM,OAAO,GAAQ,IAAI,KAAK,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;IAC7C,IAAI,IAAI,GAAG,CAAC,CAAC;IACb,KAAK,UAAU,MAAM;QACnB,OAAO,IAAI,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;YAC3B,MAAM,KAAK,GAAG,IAAI,EAAE,CAAC;YACrB,OAAO,CAAC,KAAK,CAAC,GAAG,MAAM,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,KAAK,CAAC,CAAC;QACjD,CAAC;IACH,CAAC;IACD,MAAM,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE,KAAK,CAAC,MAAM,CAAC,EAAE,EAAE,MAAM,CAAC,CAAC,CAAC;IACjF,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,oFAAoF;AACpF,MAAM,UAAU,kBAAkB,CAAC,KAAuB;IACxD,OAAO,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,QAAQ,CAAC,gBAAgB,CAAC,CAAC,CAAC;AACvF,CAAC;AAED;;;;;GAKG;AACH,SAAS,sBAAsB,CAAC,OAAe;IAC7C,OAAO,oBAAoB,CAAC,aAAa,CAAC,OAAO,CAAC,CAAC;SAChD,GAAG,CAAC,CAAC,EAAE,WAAW,EAAE,EAAE,EAAE,CAAC,WAAW,CAAC;SACrC,MAAM,CACL,CAAC,WAAW,EAAE,EAAE,CAAC,CAAC,WAAW,CAAC,UAAU,CAAC,SAAS,CAAC,IAAI,CAAC,WAAW,CAAC,UAAU,CAAC,UAAU,CAAC,CAC3F,CAAC;AACN,CAAC;AAED;;;GAGG;AACH,MAAM,CAAC,MAAM,uBAAuB,GAAG,IAAI,CAAC;AAE5C,2EAA2E;AAC3E,SAAS,YAAY,CAAC,IAAY,EAAE,QAAgB;IAClD,MAAM,QAAQ,GAAG,IAAI,CAAC,QAAQ,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAC;IAC/C,OAAO,CACL,QAAQ,KAAK,EAAE;QACf,CAAC,QAAQ,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,UAAU,CAAC,KAAK,IAAI,CAAC,GAAG,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,QAAQ,CAAC,CAAC,CAC3F,CAAC;AACJ,CAAC;AAED;;;;;;;;;GASG;AACH,MAAM,CAAC,KAAK,UAAU,iBAAiB,CACrC,IAAY,EACZ,OAAe,EACf,OAAe,OAAO,CAAC,GAAG,EAAE;IAE5B,MAAM,MAAM,GAAG,IAAI,GAAG,EAAE,CAAC;IACzB,MAAM,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;IACnC,MAAM,YAAY,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;IACxC,MAAM,eAAe,GAAG,sBAAsB,CAAC,OAAO,CAAC,CAAC;IAExD,+EAA+E;IAC/E,IAAI,QAA4B,CAAC;IACjC,IAAI,CAAC;QACH,QAAQ,GAAG,MAAM,EAAE,CAAC,QAAQ,CAAC,YAAY,CAAC,CAAC;IAC7C,CAAC;IAAC,MAAM,CAAC;QACP,QAAQ,GAAG,SAAS,CAAC;IACvB,CAAC;IAED,KAAK,MAAM,SAAS,IAAI,eAAe,EAAE,CAAC;QACxC,IAAI,MAAM,CAAC,GAAG,CAAC,SAAS,CAAC,EAAE,CAAC;YAC1B,SAAS;QACX,CAAC;QACD,IAAI,MAAM,CAAC,IAAI,IAAI,uBAAuB,EAAE,CAAC;YAC3C,MAAM;QACR,CAAC;QAED,MAAM,QAAQ,GAAG,IAAI,CAAC,OAAO,CAAC,OAAO,EAAE,SAAS,CAAC,CAAC;QAClD,IAAI,CAAC,YAAY,CAAC,YAAY,EAAE,QAAQ,CAAC,EAAE,CAAC;YAC1C,MAAM,CAAC,GAAG,CAAC,SAAS,EAAE,EAAE,IAAI,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAC;YACnE,SAAS;QACX,CAAC;QAED,sFAAsF;QACtF,uDAAuD;QACvD,IAAI,QAAQ,KAAK,SAAS,EAAE,CAAC;YAC3B,IAAI,CAAC;gBACH,MAAM,QAAQ,GAAG,MAAM,EAAE,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;gBAC7C,IAAI,CAAC,YAAY,CAAC,QAAQ,EAAE,QAAQ,CAAC,EAAE,CAAC;oBACtC,MAAM,CAAC,GAAG,CAAC,SAAS,EAAE,EAAE,IAAI,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAC;oBACnE,SAAS;gBACX,CAAC;YACH,CAAC;YAAC,MAAM,CAAC;gBACP,kCAAkC;YACpC,CAAC;QACH,CAAC;QAED,IAAI,CAAC;YACH,MAAM,KAAK,GAAG,MAAM,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;YACtC,MAAM,CAAC,GAAG,CAAC,SAAS,EAAE;gBACpB,IAAI,EAAE,SAAS;gBACf,IAAI,EAAE,KAAK,CAAC,IAAI;gBAChB,MAAM,EAAE,IAAI;aACb,CAAC,CAAC;QACL,CAAC;QAAC,MAAM,CAAC;YACP,MAAM,CAAC,GAAG,CAAC,SAAS,EAAE;gBACpB,IAAI,EAAE,SAAS;gBACf,IAAI,EAAE,CAAC;gBACP,MAAM,EAAE,KAAK;aACd,CAAC,CAAC;QACL,CAAC;IACH,CAAC;IAED,OAAO,MAAM,CAAC,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC;AAClD,CAAC"}

package/lib/core/inline-code.d.ts

@@ -0,0 +1,33 @@
+/**
+ * Matches one inline code span: a run of N backticks, then text up to the next run of exactly N
+ * backticks. It stays on one line, so use it on single-line text.
+ */
+export declare const BACKTICK_SPAN_RE: RegExp;
+/** The character that `maskInlineCode` uses. Word tokenizers never match it. */
+export declare const INLINE_CODE_MASK_CHAR = "\0";
+/**
+ * Replaces each inline code span with the same number of `INLINE_CODE_MASK_CHAR`. The length stays
+ * the same, so positions in the masked text are valid in the original.
+ */
+export declare function maskInlineCode(text: string): string;
+/**
+ * Puts the code spans of `original` back into `transformed`, a masked copy that was changed
+ * without changing its length (for example, case changes).
+ */
+export declare function restoreInlineCode(original: string, transformed: string): string;
+/**
+ * The `[start, end)` ranges of the inline code spans in `text`.
+ *
+ * Rules that run a user regex (`swap`, `pattern`) use these ranges to drop matches inside code.
+ * Masking would not work for them, because the mask characters can change what the regex matches.
+ */
+export declare function inlineCodeRanges(text: string): Array<{
+    start: number;
+    end: number;
+}>;
+/** Whether `[start, end)` overlaps any of `ranges`. Touching at an end does not count. */
+export declare function overlapsAnyRange(start: number, end: number, ranges: Array<{
+    start: number;
+    end: number;
+}>): boolean;
+//# sourceMappingURL=inline-code.d.ts.map

package/lib/core/inline-code.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"inline-code.d.ts","sourceRoot":"","sources":["../../src/core/inline-code.ts"],"names":[],"mappings":"AAAA;;;GAGG;AACH,eAAO,MAAM,gBAAgB,QAAwB,CAAC;AAEtD,gFAAgF;AAChF,eAAO,MAAM,qBAAqB,OAAO,CAAC;AAE1C;;;GAGG;AACH,wBAAgB,cAAc,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAEnD;AAED;;;GAGG;AACH,wBAAgB,iBAAiB,CAAC,QAAQ,EAAE,MAAM,EAAE,WAAW,EAAE,MAAM,GAAG,MAAM,CAY/E;AAED;;;;;GAKG;AACH,wBAAgB,gBAAgB,CAAC,IAAI,EAAE,MAAM,GAAG,KAAK,CAAC;IAAE,KAAK,EAAE,MAAM,CAAC;IAAC,GAAG,EAAE,MAAM,CAAA;CAAE,CAAC,CAOpF;AAED,0FAA0F;AAC1F,wBAAgB,gBAAgB,CAC9B,KAAK,EAAE,MAAM,EACb,GAAG,EAAE,MAAM,EACX,MAAM,EAAE,KAAK,CAAC;IAAE,KAAK,EAAE,MAAM,CAAC;IAAC,GAAG,EAAE,MAAM,CAAA;CAAE,CAAC,GAC5C,OAAO,CAET"}

package/lib/core/inline-code.js

@@ -0,0 +1,50 @@
+/**
+ * Matches one inline code span: a run of N backticks, then text up to the next run of exactly N
+ * backticks. It stays on one line, so use it on single-line text.
+ */
+export const BACKTICK_SPAN_RE = /(`+)[^\n]*?\1(?!`)/g;
+/** The character that `maskInlineCode` uses. Word tokenizers never match it. */
+export const INLINE_CODE_MASK_CHAR = '\0';
+/**
+ * Replaces each inline code span with the same number of `INLINE_CODE_MASK_CHAR`. The length stays
+ * the same, so positions in the masked text are valid in the original.
+ */
+export function maskInlineCode(text) {
+    return text.replace(BACKTICK_SPAN_RE, (span) => INLINE_CODE_MASK_CHAR.repeat(span.length));
+}
+/**
+ * Puts the code spans of `original` back into `transformed`, a masked copy that was changed
+ * without changing its length (for example, case changes).
+ */
+export function restoreInlineCode(original, transformed) {
+    let result = '';
+    let cursor = 0;
+    for (const match of original.matchAll(BACKTICK_SPAN_RE)) {
+        const start = match.index ?? 0;
+        const end = start + match[0].length;
+        result += transformed.slice(cursor, start);
+        result += original.slice(start, end);
+        cursor = end;
+    }
+    result += transformed.slice(cursor);
+    return result;
+}
+/**
+ * The `[start, end)` ranges of the inline code spans in `text`.
+ *
+ * Rules that run a user regex (`swap`, `pattern`) use these ranges to drop matches inside code.
+ * Masking would not work for them, because the mask characters can change what the regex matches.
+ */
+export function inlineCodeRanges(text) {
+    const ranges = [];
+    for (const match of text.matchAll(BACKTICK_SPAN_RE)) {
+        const start = match.index ?? 0;
+        ranges.push({ start, end: start + match[0].length });
+    }
+    return ranges;
+}
+/** Whether `[start, end)` overlaps any of `ranges`. Touching at an end does not count. */
+export function overlapsAnyRange(start, end, ranges) {
+    return ranges.some((range) => start < range.end && range.start < end);
+}
+//# sourceMappingURL=inline-code.js.map

package/lib/core/inline-code.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"inline-code.js","sourceRoot":"","sources":["../../src/core/inline-code.ts"],"names":[],"mappings":"AAAA;;;GAGG;AACH,MAAM,CAAC,MAAM,gBAAgB,GAAG,qBAAqB,CAAC;AAEtD,gFAAgF;AAChF,MAAM,CAAC,MAAM,qBAAqB,GAAG,IAAI,CAAC;AAE1C;;;GAGG;AACH,MAAM,UAAU,cAAc,CAAC,IAAY;IACzC,OAAO,IAAI,CAAC,OAAO,CAAC,gBAAgB,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,qBAAqB,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC;AAC7F,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,iBAAiB,CAAC,QAAgB,EAAE,WAAmB;IACrE,IAAI,MAAM,GAAG,EAAE,CAAC;IAChB,IAAI,MAAM,GAAG,CAAC,CAAC;IACf,KAAK,MAAM,KAAK,IAAI,QAAQ,CAAC,QAAQ,CAAC,gBAAgB,CAAC,EAAE,CAAC;QACxD,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,IAAI,CAAC,CAAC;QAC/B,MAAM,GAAG,GAAG,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;QACpC,MAAM,IAAI,WAAW,CAAC,KAAK,CAAC,MAAM,EAAE,KAAK,CAAC,CAAC;QAC3C,MAAM,IAAI,QAAQ,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC;QACrC,MAAM,GAAG,GAAG,CAAC;IACf,CAAC;IACD,MAAM,IAAI,WAAW,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;IACpC,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,gBAAgB,CAAC,IAAY;IAC3C,MAAM,MAAM,GAA0C,EAAE,CAAC;IACzD,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,gBAAgB,CAAC,EAAE,CAAC;QACpD,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,IAAI,CAAC,CAAC;QAC/B,MAAM,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC;IACvD,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,0FAA0F;AAC1F,MAAM,UAAU,gBAAgB,CAC9B,KAAa,EACb,GAAW,EACX,MAA6C;IAE7C,OAAO,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,GAAG,KAAK,CAAC,GAAG,IAAI,KAAK,CAAC,KAAK,GAAG,GAAG,CAAC,CAAC;AACxE,CAAC"}

package/lib/core/line-endings.d.ts

@@ -0,0 +1,13 @@
+/** Matches one line ending: CRLF, lone CR or LF. */
+export declare const newLineRe: RegExp;
+/**
+ * Maps a 0-based offset in `text` to a 1-based line and column. CRLF, CR and LF each count as one
+ * line ending. An offset at the `\n` of a CRLF is reported as column 1 of the next line.
+ */
+export declare function offsetToLineColumn(text: string, offset: number): {
+    line: number;
+    column: number;
+};
+/** Returns the most common line ending in `input`, or the platform default if there is none. */
+export declare function getPreferredLineEnding(input: string): string;
+//# sourceMappingURL=line-endings.d.ts.map

package/lib/core/line-endings.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"line-endings.d.ts","sourceRoot":"","sources":["../../src/core/line-endings.ts"],"names":[],"mappings":"AAIA,oDAAoD;AACpD,eAAO,MAAM,SAAS,QAAc,CAAC;AAErC;;;GAGG;AACH,wBAAgB,kBAAkB,CAAC,IAAI,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,GAAG;IAAE,IAAI,EAAE,MAAM,CAAC;IAAC,MAAM,EAAE,MAAM,CAAA;CAAE,CAQjG;AAED,gGAAgG;AAChG,wBAAgB,sBAAsB,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,CAa5D"}

package/lib/core/line-endings.js

@@ -0,0 +1,39 @@
+// Based on markdownlint's `newlineRe` and `getPreferredLineEnding` (MIT, David Anson).
+// Always split lines with `newLineRe`, never '\n', or CRLF files keep a trailing '\r' on every line.
+import * as os from 'node:os';
+/** Matches one line ending: CRLF, lone CR or LF. */
+export const newLineRe = /\r\n?|\n/g;
+/**
+ * Maps a 0-based offset in `text` to a 1-based line and column. CRLF, CR and LF each count as one
+ * line ending. An offset at the `\n` of a CRLF is reported as column 1 of the next line.
+ */
+export function offsetToLineColumn(text, offset) {
+    let line = 1;
+    let lineStart = 0;
+    for (const match of text.slice(0, offset).matchAll(newLineRe)) {
+        line++;
+        lineStart = match.index + match[0].length;
+    }
+    return { line, column: offset - lineStart + 1 };
+}
+/** Returns the most common line ending in `input`, or the platform default if there is none. */
+export function getPreferredLineEnding(input) {
+    let cr = 0;
+    let lf = 0;
+    let crlf = 0;
+    const endings = input.match(newLineRe) ?? [];
+    for (const ending of endings) {
+        if (ending === '\r')
+            cr++;
+        else if (ending === '\n')
+            lf++;
+        else
+            crlf++;
+    }
+    if (!cr && !lf && !crlf)
+        return os.EOL;
+    if (lf >= crlf && lf >= cr)
+        return '\n';
+    return crlf >= cr ? '\r\n' : '\r';
+}
+//# sourceMappingURL=line-endings.js.map

package/lib/core/line-endings.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"line-endings.js","sourceRoot":"","sources":["../../src/core/line-endings.ts"],"names":[],"mappings":"AAAA,uFAAuF;AACvF,qGAAqG;AACrG,OAAO,KAAK,EAAE,MAAM,SAAS,CAAC;AAE9B,oDAAoD;AACpD,MAAM,CAAC,MAAM,SAAS,GAAG,WAAW,CAAC;AAErC;;;GAGG;AACH,MAAM,UAAU,kBAAkB,CAAC,IAAY,EAAE,MAAc;IAC7D,IAAI,IAAI,GAAG,CAAC,CAAC;IACb,IAAI,SAAS,GAAG,CAAC,CAAC;IAClB,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,MAAM,CAAC,CAAC,QAAQ,CAAC,SAAS,CAAC,EAAE,CAAC;QAC9D,IAAI,EAAE,CAAC;QACP,SAAS,GAAG,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IAC5C,CAAC;IACD,OAAO,EAAE,IAAI,EAAE,MAAM,EAAE,MAAM,GAAG,SAAS,GAAG,CAAC,EAAE,CAAC;AAClD,CAAC;AAED,gGAAgG;AAChG,MAAM,UAAU,sBAAsB,CAAC,KAAa;IAClD,IAAI,EAAE,GAAG,CAAC,CAAC;IACX,IAAI,EAAE,GAAG,CAAC,CAAC;IACX,IAAI,IAAI,GAAG,CAAC,CAAC;IACb,MAAM,OAAO,GAAG,KAAK,CAAC,KAAK,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC;IAC7C,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,IAAI,MAAM,KAAK,IAAI;YAAE,EAAE,EAAE,CAAC;aACrB,IAAI,MAAM,KAAK,IAAI;YAAE,EAAE,EAAE,CAAC;;YAC1B,IAAI,EAAE,CAAC;IACd,CAAC;IACD,IAAI,CAAC,EAAE,IAAI,CAAC,EAAE,IAAI,CAAC,IAAI;QAAE,OAAO,EAAE,CAAC,GAAG,CAAC;IACvC,IAAI,EAAE,IAAI,IAAI,IAAI,EAAE,IAAI,EAAE;QAAE,OAAO,IAAI,CAAC;IACxC,OAAO,IAAI,IAAI,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC;AACpC,CAAC"}

package/lib/core/markdoc-tags.d.ts

@@ -0,0 +1,48 @@
+import type { TokenTree } from '../parser/types.js';
+import type { Fix } from '../types/index.js';
+/**
+ * The character that replaces a Markdoc tag in prose text (see `scopes/extractor.ts`). It is a
+ * space, so word splitting and sentence splitting treat the tag as a gap. A user regex cannot rely
+ * on this, so regex rules check match ranges instead (`ScopedSegment.maskedRanges`,
+ * `nonProseRanges` in `rules/utils.ts`).
+ */
+export declare const MARKDOC_TAG_MASK_CHAR = " ";
+/**
+ * The source text of one Markdoc tag on one line. A tag that spans several lines gives one entry
+ * per line, because a `Fix` is always on a single line.
+ */
+export interface MarkdocTagSpan {
+    /** 1-based source line. */
+    line: number;
+    /** 1-based, inclusive. */
+    startColumn: number;
+    /** 1-based, exclusive. */
+    endColumn: number;
+    /** The source bytes between those columns — what must survive a fix. */
+    text: string;
+}
+/**
+ * The source span of every Markdoc tag in `content`, split by line. Empty when Markdoc parsing is
+ * off. Uses `filterByTypes` to skip tokens that repeat the position of their `htmlFlow` parent.
+ */
+export declare function markdocTagSpans(tree: TokenTree, content: string): MarkdocTagSpan[];
+export interface ProtectedFixes {
+    /** Fixes that are safe to apply. */
+    fixes: Fix[];
+    /** Fixes that could not be made safe. The caller reports them as skipped. */
+    dropped: Fix[];
+}
+/**
+ * Keeps `--fix` from changing Markdoc tags. Prose text has tags masked, so a fix that replaces a
+ * whole span would write the mask over the real tag and delete it. Every fix passes through here:
+ *
+ * - A fix that does not touch a tag is kept as is.
+ * - A fix that replaces text with the same number of characters keeps its place, and the original
+ *   tag text is copied back in.
+ * - Any other fix that touches a tag is dropped and reported as skipped.
+ *
+ * A fix that changes nothing once the tags are restored is dropped too. Otherwise it would be
+ * reported as fixed and proposed again on every pass of `runRulesUntilStable`.
+ */
+export declare function protectMarkdocTags(fixes: Fix[], spans: MarkdocTagSpan[], content: string): ProtectedFixes;
+//# sourceMappingURL=markdoc-tags.d.ts.map

package/lib/core/markdoc-tags.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"markdoc-tags.d.ts","sourceRoot":"","sources":["../../src/core/markdoc-tags.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,oBAAoB,CAAC;AACpD,OAAO,KAAK,EAAE,GAAG,EAAE,MAAM,mBAAmB,CAAC;AAG7C;;;;;GAKG;AACH,eAAO,MAAM,qBAAqB,MAAM,CAAC;AAEzC;;;GAGG;AACH,MAAM,WAAW,cAAc;IAC7B,2BAA2B;IAC3B,IAAI,EAAE,MAAM,CAAC;IACb,0BAA0B;IAC1B,WAAW,EAAE,MAAM,CAAC;IACpB,0BAA0B;IAC1B,SAAS,EAAE,MAAM,CAAC;IAClB,wEAAwE;IACxE,IAAI,EAAE,MAAM,CAAC;CACd;AAED;;;GAGG;AACH,wBAAgB,eAAe,CAAC,IAAI,EAAE,SAAS,EAAE,OAAO,EAAE,MAAM,GAAG,cAAc,EAAE,CAoBlF;AAED,MAAM,WAAW,cAAc;IAC7B,oCAAoC;IACpC,KAAK,EAAE,GAAG,EAAE,CAAC;IACb,6EAA6E;IAC7E,OAAO,EAAE,GAAG,EAAE,CAAC;CAChB;AAED;;;;;;;;;;;GAWG;AACH,wBAAgB,kBAAkB,CAChC,KAAK,EAAE,GAAG,EAAE,EACZ,KAAK,EAAE,cAAc,EAAE,EACvB,OAAO,EAAE,MAAM,GACd,cAAc,CAsDhB"}

package/lib/core/markdoc-tags.js

@@ -0,0 +1,102 @@
+import { filterByTypes } from '../parser/index.js';
+import { newLineRe } from './line-endings.js';
+/**
+ * The character that replaces a Markdoc tag in prose text (see `scopes/extractor.ts`). It is a
+ * space, so word splitting and sentence splitting treat the tag as a gap. A user regex cannot rely
+ * on this, so regex rules check match ranges instead (`ScopedSegment.maskedRanges`,
+ * `nonProseRanges` in `rules/utils.ts`).
+ */
+export const MARKDOC_TAG_MASK_CHAR = ' ';
+/**
+ * The source span of every Markdoc tag in `content`, split by line. Empty when Markdoc parsing is
+ * off. Uses `filterByTypes` to skip tokens that repeat the position of their `htmlFlow` parent.
+ */
+export function markdocTagSpans(tree, content) {
+    const tags = filterByTypes(tree, ['markdocTag']);
+    if (tags.length === 0)
+        return [];
+    const sourceLines = content.split(newLineRe);
+    const spans = [];
+    for (const tag of tags) {
+        for (let line = tag.startLine; line <= tag.endLine; line++) {
+            const sourceLine = sourceLines[line - 1] ?? '';
+            const startColumn = line === tag.startLine ? tag.startColumn : 1;
+            const endColumn = line === tag.endLine ? tag.endColumn : sourceLine.length + 1;
+            if (endColumn <= startColumn)
+                continue;
+            spans.push({
+                line,
+                startColumn,
+                endColumn,
+                text: sourceLine.slice(startColumn - 1, endColumn - 1),
+            });
+        }
+    }
+    return spans;
+}
+/**
+ * Keeps `--fix` from changing Markdoc tags. Prose text has tags masked, so a fix that replaces a
+ * whole span would write the mask over the real tag and delete it. Every fix passes through here:
+ *
+ * - A fix that does not touch a tag is kept as is.
+ * - A fix that replaces text with the same number of characters keeps its place, and the original
+ *   tag text is copied back in.
+ * - Any other fix that touches a tag is dropped and reported as skipped.
+ *
+ * A fix that changes nothing once the tags are restored is dropped too. Otherwise it would be
+ * reported as fixed and proposed again on every pass of `runRulesUntilStable`.
+ */
+export function protectMarkdocTags(fixes, spans, content) {
+    if (spans.length === 0 || fixes.length === 0)
+        return { fixes, dropped: [] };
+    const byLine = new Map();
+    for (const span of spans) {
+        const existing = byLine.get(span.line);
+        if (existing)
+            existing.push(span);
+        else
+            byLine.set(span.line, [span]);
+    }
+    const sourceLines = content.split(newLineRe);
+    const safe = [];
+    const dropped = [];
+    for (const fix of fixes) {
+        const lineSpans = byLine.get(fix.lineNumber);
+        if (lineSpans === undefined) {
+            safe.push(fix);
+            continue;
+        }
+        const deleteCount = fix.deleteCount ?? 0;
+        if (deleteCount === -1) {
+            dropped.push(fix);
+            continue;
+        }
+        const editColumn = fix.editColumn ?? 1;
+        const editEnd = editColumn + deleteCount;
+        const overlapped = lineSpans.filter((span) => editColumn < span.endColumn && span.startColumn < editEnd);
+        if (overlapped.length === 0) {
+            safe.push(fix);
+            continue;
+        }
+        const insertText = fix.insertText ?? '';
+        const splittable = insertText.length === deleteCount &&
+            overlapped.every((span) => span.startColumn >= editColumn && span.endColumn <= editEnd);
+        if (!splittable) {
+            dropped.push(fix);
+            continue;
+        }
+        let restored = insertText;
+        for (const span of overlapped) {
+            const at = span.startColumn - editColumn;
+            restored = restored.slice(0, at) + span.text + restored.slice(at + span.text.length);
+        }
+        const replaced = (sourceLines[fix.lineNumber - 1] ?? '').slice(editColumn - 1, editEnd - 1);
+        if (restored === replaced) {
+            dropped.push(fix);
+            continue;
+        }
+        safe.push({ ...fix, insertText: restored });
+    }
+    return { fixes: safe, dropped };
+}
+//# sourceMappingURL=markdoc-tags.js.map

package/lib/core/markdoc-tags.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"markdoc-tags.js","sourceRoot":"","sources":["../../src/core/markdoc-tags.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,aAAa,EAAE,MAAM,oBAAoB,CAAC;AAGnD,OAAO,EAAE,SAAS,EAAE,MAAM,mBAAmB,CAAC;AAE9C;;;;;GAKG;AACH,MAAM,CAAC,MAAM,qBAAqB,GAAG,GAAG,CAAC;AAiBzC;;;GAGG;AACH,MAAM,UAAU,eAAe,CAAC,IAAe,EAAE,OAAe;IAC9D,MAAM,IAAI,GAAG,aAAa,CAAC,IAAI,EAAE,CAAC,YAAY,CAAC,CAAC,CAAC;IACjD,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IACjC,MAAM,WAAW,GAAG,OAAO,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC;IAC7C,MAAM,KAAK,GAAqB,EAAE,CAAC;IACnC,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;QACvB,KAAK,IAAI,IAAI,GAAG,GAAG,CAAC,SAAS,EAAE,IAAI,IAAI,GAAG,CAAC,OAAO,EAAE,IAAI,EAAE,EAAE,CAAC;YAC3D,MAAM,UAAU,GAAG,WAAW,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;YAC/C,MAAM,WAAW,GAAG,IAAI,KAAK,GAAG,CAAC,SAAS,CAAC,CAAC,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,CAAC;YACjE,MAAM,SAAS,GAAG,IAAI,KAAK,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC,CAAC,UAAU,CAAC,MAAM,GAAG,CAAC,CAAC;YAC/E,IAAI,SAAS,IAAI,WAAW;gBAAE,SAAS;YACvC,KAAK,CAAC,IAAI,CAAC;gBACT,IAAI;gBACJ,WAAW;gBACX,SAAS;gBACT,IAAI,EAAE,UAAU,CAAC,KAAK,CAAC,WAAW,GAAG,CAAC,EAAE,SAAS,GAAG,CAAC,CAAC;aACvD,CAAC,CAAC;QACL,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AASD;;;;;;;;;;;GAWG;AACH,MAAM,UAAU,kBAAkB,CAChC,KAAY,EACZ,KAAuB,EACvB,OAAe;IAEf,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,KAAK,EAAE,OAAO,EAAE,EAAE,EAAE,CAAC;IAE5E,MAAM,MAAM,GAAG,IAAI,GAAG,EAA4B,CAAC;IACnD,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,QAAQ,GAAG,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACvC,IAAI,QAAQ;YAAE,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;;YAC7B,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,IAAI,CAAC,CAAC,CAAC;IACrC,CAAC;IACD,MAAM,WAAW,GAAG,OAAO,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC;IAE7C,MAAM,IAAI,GAAU,EAAE,CAAC;IACvB,MAAM,OAAO,GAAU,EAAE,CAAC;IAC1B,KAAK,MAAM,GAAG,IAAI,KAAK,EAAE,CAAC;QACxB,MAAM,SAAS,GAAG,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC;QAC7C,IAAI,SAAS,KAAK,SAAS,EAAE,CAAC;YAC5B,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACf,SAAS;QACX,CAAC;QACD,MAAM,WAAW,GAAG,GAAG,CAAC,WAAW,IAAI,CAAC,CAAC;QACzC,IAAI,WAAW,KAAK,CAAC,CAAC,EAAE,CAAC;YACvB,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YAClB,SAAS;QACX,CAAC;QACD,MAAM,UAAU,GAAG,GAAG,CAAC,UAAU,IAAI,CAAC,CAAC;QACvC,MAAM,OAAO,GAAG,UAAU,GAAG,WAAW,CAAC;QACzC,MAAM,UAAU,GAAG,SAAS,CAAC,MAAM,CACjC,CAAC,IAAI,EAAE,EAAE,CAAC,UAAU,GAAG,IAAI,CAAC,SAAS,IAAI,IAAI,CAAC,WAAW,GAAG,OAAO,CACpE,CAAC;QACF,IAAI,UAAU,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAC5B,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACf,SAAS;QACX,CAAC;QACD,MAAM,UAAU,GAAG,GAAG,CAAC,UAAU,IAAI,EAAE,CAAC;QACxC,MAAM,UAAU,GACd,UAAU,CAAC,MAAM,KAAK,WAAW;YACjC,UAAU,CAAC,KAAK,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,WAAW,IAAI,UAAU,IAAI,IAAI,CAAC,SAAS,IAAI,OAAO,CAAC,CAAC;QAC1F,IAAI,CAAC,UAAU,EAAE,CAAC;YAChB,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YAClB,SAAS;QACX,CAAC;QACD,IAAI,QAAQ,GAAG,UAAU,CAAC;QAC1B,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;YAC9B,MAAM,EAAE,GAAG,IAAI,CAAC,WAAW,GAAG,UAAU,CAAC;YACzC,QAAQ,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,GAAG,IAAI,CAAC,IAAI,GAAG,QAAQ,CAAC,KAAK,CAAC,EAAE,GAAG,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QACvF,CAAC;QACD,MAAM,QAAQ,GAAG,CAAC,WAAW,CAAC,GAAG,CAAC,UAAU,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,UAAU,GAAG,CAAC,EAAE,OAAO,GAAG,CAAC,CAAC,CAAC;QAC5F,IAAI,QAAQ,KAAK,QAAQ,EAAE,CAAC;YAC1B,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YAClB,SAAS;QACX,CAAC;QACD,IAAI,CAAC,IAAI,CAAC,EAAE,GAAG,GAAG,EAAE,UAAU,EAAE,QAAQ,EAAE,CAAC,CAAC;IAC9C,CAAC;IACD,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,OAAO,EAAE,CAAC;AAClC,CAAC"}

package/lib/core/prose-extract.d.ts

@@ -0,0 +1,6 @@
+import type { ScopedSegment } from '../scopes/types.js';
+/** Removes Markdoc tags and inline code. */
+export declare function stripNonProse(text: string): string;
+/** Builds the prose blocks a readability score reads, from `summary` segments. */
+export declare function extractProse(segments: ScopedSegment[]): string[];
+//# sourceMappingURL=prose-extract.d.ts.map

package/lib/core/prose-extract.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"prose-extract.d.ts","sourceRoot":"","sources":["../../src/core/prose-extract.ts"],"names":[],"mappings":"AAEA,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,oBAAoB,CAAC;AAOxD,4CAA4C;AAC5C,wBAAgB,aAAa,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAElD;AAyDD,kFAAkF;AAClF,wBAAgB,YAAY,CAAC,QAAQ,EAAE,aAAa,EAAE,GAAG,MAAM,EAAE,CAgBhE"}

package/lib/core/prose-extract.js

@@ -0,0 +1,71 @@
+import { BACKTICK_SPAN_RE } from './inline-code.js';
+// Markdoc tags. Non-greedy, so tags next to each other are removed one by one. Only used when
+// Markdoc parsing is off; otherwise tags are already masked in `segment.maskedRanges`.
+const MARKDOC_TAG_RE = /\{%-?[\s\S]*?-?%\}/g;
+/** Removes Markdoc tags and inline code. */
+export function stripNonProse(text) {
+    return text.replace(MARKDOC_TAG_RE, '').replace(BACKTICK_SPAN_RE, '');
+}
+// Cuts masked tags out of the text instead of leaving spaces, so `<code>{% x %}</code>` is one word.
+function spliceOutMaskedSpans(segment) {
+    const ranges = segment.maskedRanges;
+    if (ranges === undefined)
+        return segment.content;
+    let result = '';
+    let cursor = 0;
+    for (const range of ranges) {
+        result += segment.content.slice(cursor, Math.max(range.start, cursor));
+        cursor = Math.max(range.end, cursor);
+    }
+    return result + segment.content.slice(cursor);
+}
+function comparePosition(aLine, aColumn, bLine, bColumn) {
+    return aLine - bLine || aColumn - bColumn;
+}
+function spanContains(outer, inner) {
+    return (comparePosition(outer.startLine, outer.startColumn, inner.startLine, inner.startColumn) <= 0 &&
+        comparePosition(outer.endLine, outer.endColumn, inner.endLine, inner.endColumn) >= 0);
+}
+// Drops segments that sit inside another segment, such as a list in a blockquote, so words are
+// not counted twice. Sorted so that a container comes before what it contains.
+function dedupeBySpanContainment(segments) {
+    const ordered = [...segments].sort((a, b) => {
+        const byStart = comparePosition(a.startLine, a.startColumn, b.startLine, b.startColumn);
+        if (byStart !== 0)
+            return byStart;
+        return comparePosition(b.endLine, b.endColumn, a.endLine, a.endColumn);
+    });
+    const stack = [];
+    const kept = [];
+    for (const segment of ordered) {
+        while (stack.length > 0) {
+            const top = stack[stack.length - 1];
+            if (comparePosition(top.endLine, top.endColumn, segment.startLine, segment.startColumn) < 0) {
+                stack.pop();
+            }
+            else {
+                break;
+            }
+        }
+        const container = stack[stack.length - 1];
+        if (container && spanContains(container, segment))
+            continue;
+        kept.push(segment);
+        stack.push(segment);
+    }
+    return kept;
+}
+/** Builds the prose blocks a readability score reads, from `summary` segments. */
+export function extractProse(segments) {
+    // Headings are fragments, so readability tools skip them.
+    const proseSegments = segments.filter((segment) => !segment.sourceScope?.startsWith('heading.'));
+    const deduped = dedupeBySpanContainment(proseSegments);
+    // Back to source order.
+    deduped.sort((a, b) => a.startLine - b.startLine || a.startColumn - b.startColumn);
+    return deduped
+        .map((segment) => segment.maskedRanges !== undefined
+        ? spliceOutMaskedSpans(segment).replace(BACKTICK_SPAN_RE, '')
+        : stripNonProse(segment.content))
+        .filter((text) => text.trim() !== '');
+}
+//# sourceMappingURL=prose-extract.js.map

package/lib/core/prose-extract.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"prose-extract.js","sourceRoot":"","sources":["../../src/core/prose-extract.ts"],"names":[],"mappings":"AAGA,OAAO,EAAE,gBAAgB,EAAE,MAAM,kBAAkB,CAAC;AAEpD,8FAA8F;AAC9F,uFAAuF;AACvF,MAAM,cAAc,GAAG,qBAAqB,CAAC;AAE7C,4CAA4C;AAC5C,MAAM,UAAU,aAAa,CAAC,IAAY;IACxC,OAAO,IAAI,CAAC,OAAO,CAAC,cAAc,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,gBAAgB,EAAE,EAAE,CAAC,CAAC;AACxE,CAAC;AAED,qGAAqG;AACrG,SAAS,oBAAoB,CAAC,OAAsB;IAClD,MAAM,MAAM,GAAG,OAAO,CAAC,YAAY,CAAC;IACpC,IAAI,MAAM,KAAK,SAAS;QAAE,OAAO,OAAO,CAAC,OAAO,CAAC;IACjD,IAAI,MAAM,GAAG,EAAE,CAAC;IAChB,IAAI,MAAM,GAAG,CAAC,CAAC;IACf,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC3B,MAAM,IAAI,OAAO,CAAC,OAAO,CAAC,KAAK,CAAC,MAAM,EAAE,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC,CAAC;QACvE,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC;IACvC,CAAC;IACD,OAAO,MAAM,GAAG,OAAO,CAAC,OAAO,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;AAChD,CAAC;AAED,SAAS,eAAe,CAAC,KAAa,EAAE,OAAe,EAAE,KAAa,EAAE,OAAe;IACrF,OAAO,KAAK,GAAG,KAAK,IAAI,OAAO,GAAG,OAAO,CAAC;AAC5C,CAAC;AAED,SAAS,YAAY,CAAC,KAAoB,EAAE,KAAoB;IAC9D,OAAO,CACL,eAAe,CAAC,KAAK,CAAC,SAAS,EAAE,KAAK,CAAC,WAAW,EAAE,KAAK,CAAC,SAAS,EAAE,KAAK,CAAC,WAAW,CAAC,IAAI,CAAC;QAC5F,eAAe,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,SAAS,EAAE,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,SAAS,CAAC,IAAI,CAAC,CACrF,CAAC;AACJ,CAAC;AAED,+FAA+F;AAC/F,+EAA+E;AAC/E,SAAS,uBAAuB,CAAC,QAAyB;IACxD,MAAM,OAAO,GAAG,CAAC,GAAG,QAAQ,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE;QAC1C,MAAM,OAAO,GAAG,eAAe,CAAC,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC,WAAW,CAAC,CAAC;QACxF,IAAI,OAAO,KAAK,CAAC;YAAE,OAAO,OAAO,CAAC;QAClC,OAAO,eAAe,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC,SAAS,CAAC,CAAC;IACzE,CAAC,CAAC,CAAC;IAEH,MAAM,KAAK,GAAoB,EAAE,CAAC;IAClC,MAAM,IAAI,GAAoB,EAAE,CAAC;IACjC,KAAK,MAAM,OAAO,IAAI,OAAO,EAAE,CAAC;QAC9B,OAAO,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACxB,MAAM,GAAG,GAAG,KAAK,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;YACpC,IAAI,eAAe,CAAC,GAAG,CAAC,OAAO,EAAE,GAAG,CAAC,SAAS,EAAE,OAAO,CAAC,SAAS,EAAE,OAAO,CAAC,WAAW,CAAC,GAAG,CAAC,EAAE,CAAC;gBAC5F,KAAK,CAAC,GAAG,EAAE,CAAC;YACd,CAAC;iBAAM,CAAC;gBACN,MAAM;YACR,CAAC;QACH,CAAC;QAED,MAAM,SAAS,GAAG,KAAK,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;QAC1C,IAAI,SAAS,IAAI,YAAY,CAAC,SAAS,EAAE,OAAO,CAAC;YAAE,SAAS;QAE5D,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;QACnB,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IACtB,CAAC;IAED,OAAO,IAAI,CAAC;AACd,CAAC;AAED,kFAAkF;AAClF,MAAM,UAAU,YAAY,CAAC,QAAyB;IACpD,0DAA0D;IAC1D,MAAM,aAAa,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,WAAW,EAAE,UAAU,CAAC,UAAU,CAAC,CAAC,CAAC;IAEjG,MAAM,OAAO,GAAG,uBAAuB,CAAC,aAAa,CAAC,CAAC;IAEvD,wBAAwB;IACxB,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,SAAS,GAAG,CAAC,CAAC,SAAS,IAAI,CAAC,CAAC,WAAW,GAAG,CAAC,CAAC,WAAW,CAAC,CAAC;IAEnF,OAAO,OAAO;SACX,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CACf,OAAO,CAAC,YAAY,KAAK,SAAS;QAChC,CAAC,CAAC,oBAAoB,CAAC,OAAO,CAAC,CAAC,OAAO,CAAC,gBAAgB,EAAE,EAAE,CAAC;QAC7D,CAAC,CAAC,aAAa,CAAC,OAAO,CAAC,OAAO,CAAC,CACnC;SACA,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC;AAC1C,CAAC"}

package/lib/core/readability.d.ts

@@ -0,0 +1,17 @@
+export interface DocumentReadability {
+    words: number;
+    sentences: number;
+    /** Flesch reading ease; higher is easier. Null when there is no prose. */
+    fleschReadingEase: number | null;
+    /** Flesch-Kincaid grade level. Null when there is no prose. */
+    fleschKincaidGrade: number | null;
+    /**
+     * Automated Readability Index, a grade level based on character counts. Null when there is no
+     * prose.
+     */
+    automatedReadabilityIndex: number | null;
+}
+export declare function computeDocumentReadability(content: string, options?: {
+    markdoc?: boolean;
+}): DocumentReadability;
+//# sourceMappingURL=readability.d.ts.map

package/lib/core/readability.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"readability.d.ts","sourceRoot":"","sources":["../../src/core/readability.ts"],"names":[],"mappings":"AAKA,MAAM,WAAW,mBAAmB;IAClC,KAAK,EAAE,MAAM,CAAC;IACd,SAAS,EAAE,MAAM,CAAC;IAClB,0EAA0E;IAC1E,iBAAiB,EAAE,MAAM,GAAG,IAAI,CAAC;IACjC,+DAA+D;IAC/D,kBAAkB,EAAE,MAAM,GAAG,IAAI,CAAC;IAClC;;;OAGG;IACH,yBAAyB,EAAE,MAAM,GAAG,IAAI,CAAC;CAC1C;AAED,wBAAgB,0BAA0B,CACxC,OAAO,EAAE,MAAM,EACf,OAAO,GAAE;IAAE,OAAO,CAAC,EAAE,OAAO,CAAA;CAAO,GAClC,mBAAmB,CAiCrB"}