// A /sweep REPORT (markdown) → `report.json` (kind "sweep").
//
// What /sweep writes (mcp/src/instructions.ts): `## sections` of findings,
// each cited as `[title @ mm:ss](<origin>/?v=<channel>%2F<id>&t=<seconds>)`
// for a video and `[post by <author>, <date>](<url>)` for a post — the url an
// archive post (`…&vm=post`) or the post on its platform. The engine
// (./convertShared.ts `structureMarkdown`) makes the document's first `#`
// heading the title, the text before the first section the summary, each
// section a section, and each list item or paragraph that cites a claim; every
// citing link becomes `[label](cite:<id>)`.
//
// A citation's quote is the quoted words nearest before its link (`"…"` or
// `“…”`, several joined by an ellipsis), else the claim's text — VERBATIM is
// the model's rule, and compose checks it against the cues. Its span is the
// cited second widened through the record's cues when the caller can read
// them (`ctx.cuesOf`), else the second plus `ctx.spanSeconds`. A post on its
// platform is cited only when `ctx.postChannelOf` finds the archive channel
// that keeps it; otherwise its link stays a plain link, with a warning.

import {
  CitationRegistry,
  emptyReport,
  finish,
  reportIdOf,
  structureMarkdown,
  type ConvertContext,
  type Converted,
} from "./convertShared";

export type SweepConvertOptions = ConvertContext & {
  // The report's id; default: the title's slug.
  id?: string;
  // The report's title; default: the markdown's first `#` heading.
  title?: string;
  // The one section of a sweep with no section headings.
  sectionTitle?: string;
};

export async function sweepToReport(md: string, opts: SweepConvertOptions = {}): Promise<Converted> {
  const warnings: string[] = [];
  const registry = new CitationRegistry(opts, warnings);
  const doc = await structureMarkdown(md, (link) => registry.link(link.href, link.label, link.quote), {
    defaultSection: opts.sectionTitle ?? "Findings",
  });
  const title = opts.title ?? doc.title ?? "Sweep";
  const report = emptyReport(reportIdOf(opts.id, title), "sweep", title);
  if (doc.summary) report.summary = doc.summary;
  report.citations = registry.citations;
  report.sections = doc.sections;
  if (!Object.keys(registry.citations).length) warnings.push("the markdown cites nothing the citation model knows");
  return finish(report, warnings);
}