// INLINE CITATIONS — the one syntax for citing in markdown, and the numbers a
// reader sees.
//
//   [label](cite:<id>)
//
// anywhere a document's markdown is (a report's summary, a section's body, a
// claim's findings). `<id>` names a citation in the same document. A renderer
// replaces the link with a numbered marker that previews the citation and
// jumps to it; the citation's anchor is `#c-<id>` (`citationAnchor`), stable
// across edits because it is the id, not the number.
//
// NUMBERS ARE PER DOCUMENT, BY FIRST APPEARANCE: the first citation cited is
// [1], the next new one [2], and a citation cited again keeps its number. The
// document decides the order it is read in (lib/report/uses.ts walks a
// report); `numberCitations` only counts.
//
// Code is not citing: a `cite:` link inside an inline code span or a fenced
// block is text about the syntax, and is skipped.
//
// Pure, no imports: the export site's pages and the browser can use it.

export const CITE_SCHEME = "cite:";

export type CiteRef = {
  id: string;
  label: string;
  // Where the link starts in the markdown (a UTF-16 offset).
  offset: number;
};

// `[label](cite:id)`. The label may hold one level of balanced brackets — an
// editorial insertion in a quote, `[“told [the mayor] so”](cite:id)`, which
// markdown-to-jsx links too — but not a lone `]`; the id runs to the first `)`
// or space (an empty id is still a ref, so validation can name it).
const CITE_LINK_RE = /\[((?:[^[\]]|\[[^[\]]*\])*)\]\(\s*cite:([^)\s]*)\s*\)/g;

// Fenced blocks: a line opening with ``` or ~~~ (up to three spaces in) to
// the line closing it with at least as many of the same, or the end.
const FENCE_OPEN_RE = /^ {0,3}(`{3,}|~{3,})/;

// An inline code span: a run of backticks, to the next run of the same length.
const CODE_SPAN_RE = /(?<!`)(`+)(?!`)[\s\S]*?(?<!`)\1(?!`)/g;

const blank = (s: string) => s.replace(/[^\n]/g, " ");

// The markdown with every fenced block and code span blanked to spaces, so
// offsets still point into the original.
function blankCode(md: string): string {
  const lines = md.split("\n");
  let fence: string | null = null;
  for (let i = 0; i < lines.length; i++) {
    if (fence) {
      const close = new RegExp(`^ {0,3}${fence[0] === "`" ? "`" : "~"}{${fence.length},}\\s*$`);
      if (close.test(lines[i])) fence = null;
      lines[i] = blank(lines[i]);
      continue;
    }
    const open = FENCE_OPEN_RE.exec(lines[i]);
    if (open) {
      fence = open[1];
      lines[i] = blank(lines[i]);
    }
  }
  return lines.join("\n").replace(CODE_SPAN_RE, blank);
}

// Every `cite:` link in the markdown, in order.
export function extractCiteRefs(md: string | null | undefined): CiteRef[] {
  if (!md) return [];
  const scan = blankCode(md);
  const out: CiteRef[] = [];
  for (const m of scan.matchAll(CITE_LINK_RE)) {
    out.push({ id: m[2], label: md.slice(m.index + 1, m.index + 1 + m[1].length), offset: m.index });
  }
  return out;
}

// Numbers by first appearance: each id's number, from 1, in the order the ids
// are given; a repeated id keeps its first number.
export function numberCitations(ids: Iterable<string>): Map<string, number> {
  const out = new Map<string, number>();
  for (const id of ids) if (!out.has(id)) out.set(id, out.size + 1);
  return out;
}

// The href an inline citation is written with.
export function citeHref(id: string): string {
  return `${CITE_SCHEME}${id}`;
}

// The anchor (fragment id, without `#`) of a citation's entry in a document.
export function citationAnchor(id: string): string {
  return `c-${id}`;
}