/**
 * Content body helpers — pure, dependency-free logic for reading and
 * shaping the Markdown sidecar that sits next to each record.
 *
 * Five concerns live here, all framework-agnostic so any renderer
 * (Astro today, future options) can compose them:
 *
 *   1. `resolveContentPath(contentPath, candidates?)`
 *      Locate a `ProjectRecord.content` path on disk. The candidates
 *      list mirrors the conventions a CLI-scaffolded project uses
 *      (`./content/...`, bare `content/...`, the legacy `apps/example`
 *      form for older workspaces). Returns the first existing absolute
 *      path, or null.
 *
 *   2. `readContentFile(contentPath, candidates?)`
 *      Combine resolve + read + frontmatter strip. Returns
 *      `{ body, frontmatter }` where `body` is the Markdown without its
 *      leading `---...---` block and `frontmatter` is the YAML text
 *      between the fences (or empty string when there's no frontmatter).
 *
 *   3. `stripFrontmatter(text)` — single source of truth for the
 *      leading-YAML-block rule. Used by `readContentFile`; also
 *      exported for consumers who already have the file contents in
 *      memory.
 *
 *   4. `extractToc(body, { maxDepth })` — pull h2/h3 headings out of a
 *      body so a record page can render a Table-of-Contents. IDs go
 *      through `slug.uniqueSlug` so they match the IDs the markdown
 *      renderer will emit (and so two TOC entries with the same
 *      label get `foo`, `foo-2`, `foo-3` rather than colliding).
 *
 *   5. `readingMetrics(body, { wpm })` — word count + minutes for the
 *      "X min read" pill on a detail page. Returns zeros (not throws)
 *      for empty input.
 *
 * Why a separate module instead of tacking these onto `markdown.ts`?
 * That file is the **awesome-list importer** (parses READMEs into
 * Grove records). Conflating record-side rendering helpers with
 * importer-side parsing helpers would re-create the
 * "what does this file do?" confusion that already exists between
 * `parseReadme.ts` and `markdown.ts`.
 */
/**
 * Resolve a record's `content` path against a list of candidate
 * roots. Returns the absolute path of the first existing candidate,
 * or null when none match.
 */
export declare function resolveContentPath(contentPath: string, candidates?: string[]): string | null;
/**
 * Strip a leading YAML frontmatter block (`---\n...\n---`) from a
 * Markdown string. Returns the body with the frontmatter removed;
 * callers that need the frontmatter contents should use
 * `readContentFile` instead.
 *
 * Restricts the search to the first 200 lines so a `---` later in
 * the document (e.g. as a horizontal rule) isn't mistaken for the
 * closing fence.
 */
export declare function stripFrontmatter(text: string): string;
export interface ReadContentFileResult {
    /** Markdown body with frontmatter stripped. */
    body: string;
    /** Raw YAML text between the frontmatter fences (no fences, no trim). */
    frontmatter: string;
    /** Absolute path the file was resolved from. */
    path: string;
}
/**
 * Read a record's content file from disk and split it into
 * frontmatter + body. Returns null when the file can't be located.
 */
export declare function readContentFile(contentPath: string, candidates?: string[]): ReadContentFileResult | null;
/**
 * Slug used for heading anchors. Unlike `slug.ts:slugify`, this is
 * GitHub-flavoured:
 *   - drops diacritics / smart quotes the same way `slug.ts` does
 *     (so README-style copy-paste gives predictable IDs)
 *   - replaces spaces with hyphens (rather than dropping everything
 *     that isn't `[a-z0-9]`)
 *   - collapses repeated hyphens
 *   - trims leading/trailing hyphens
 *   - no length cap (headings are short by nature)
 *
 * The collision counter (via `uniqueSlug`) gives `foo`, `foo-2`, …
 * so two headings labelled "Examples" in the same document anchor
 * to distinct IDs.
 *
 * Exported because the markdown renderer in `@grove-dev/astro` uses
 * this same slug rule so the IDs it emits on `<h2 id="…">` line up
 * with the IDs `extractToc` produces from the same body. Duplicating
 * the rule is the documented failure mode this comment warns against.
 */
export declare function headingSlug(text: string): string;
export interface TocEntry {
    /** Display text of the heading (Markdown syntax stripped). */
    text: string;
    /** Stable anchor id (kebab-case, collision-suffixed). */
    id: string;
    /** Heading depth — 2 or 3 by default. */
    depth: 2 | 3 | 4 | 5 | 6;
}
export interface ExtractTocOptions {
    /** Highest heading depth to include (default 2 → only `##` entries). */
    maxDepth?: 2 | 3 | 4 | 5 | 6;
}
/**
 * Pull a flat list of `{ text, id, depth }` out of a Markdown body,
 * skipping any leading frontmatter block. ID collisions are resolved
 * via `uniqueSlug` so the returned IDs match what a markdown→HTML
 * renderer (configured with the same slug rules) will emit.
 */
export declare function extractToc(body: string, options?: ExtractTocOptions): TocEntry[];
export interface ReadingMetrics {
    /** Whitespace-separated token count of the body. */
    wordCount: number;
    /** Reading time in minutes, rounded up, never less than 1. */
    minutes: number;
}
export interface ReadingMetricsOptions {
    /** Words-per-minute baseline (default 200). */
    wpm?: number;
}
/**
 * Word count + minutes for "X min read". Whitespace-only and empty
 * bodies return `{ wordCount: 0, minutes: 1 }` — the `minutes: 1`
 * floor prevents detail pages from rendering "0 min" for short
 * placeholder bodies.
 */
export declare function readingMetrics(body: string, options?: ReadingMetricsOptions): ReadingMetrics;
//# sourceMappingURL=content-body.d.ts.map