import { sha256, simhash, type ConnectorMetadata, type NormalizedContent, type Observation, type SensorEndpoint } from "@websensor/core";
import { httpFetchWithRetry } from "./fetcher";
import { NormalizeError, type WebSensorConnector } from "./types";

/**
 * SEC EDGAR connector — company submissions API (`https://data.sec.gov/submissions/CIK##########.json`).
 * The API is columnar (`filings.recent.form[]`, `filings.recent.filingDate[]`…), which the generic
 * `jsonlist` connector cannot read; this connector rebuilds one item per filing keyed by accession
 * number, so a new 8-K / 10-Q / S-1 / 13D becomes a `filing` list event with a direct document link.
 *
 * Rules of the road (sec.gov/developer): a descriptive User-Agent with contact e-mail is mandatory,
 * ≤ 10 requests/second overall. The engine's per-host concurrency (2) plus tier B/C intervals keep us
 * far below that.
 *
 * Config: { forms?: ["8-K", "10-K", "10-Q", "S-1", "6-K", "SC 13D", "DEF 14A"], excludeForms?: ["4", "144", "3", "5"], maxItems?: 60 }
 * Default: all forms except insider forms 3/4/5 and 144 notices (noise).
 */
const DEFAULT_EXCLUDE = new Set(["3", "4", "5", "3/A", "4/A", "5/A", "144", "144/A"]);
const UA = process.env.WS_EDGAR_USER_AGENT ?? "WebSensor (contact@websensor.io)";

const FORM_LABEL: Record<string, string> = {
  "8-K": "Current report",
  "10-K": "Annual report",
  "10-Q": "Quarterly report",
  "20-F": "Annual report (foreign issuer)",
  "40-F": "Annual report (Canadian issuer)",
  "6-K": "Report of foreign issuer",
  "S-1": "Registration statement (IPO)",
  "S-3": "Shelf registration",
  "S-4": "Merger registration",
  "S-8": "Employee plan registration",
  "424B4": "Prospectus (final)",
  "424B5": "Prospectus supplement",
  "DEF 14A": "Proxy statement",
  "DEFA14A": "Additional proxy materials",
  "SC 13D": "Beneficial ownership >5% (active)",
  "SC 13G": "Beneficial ownership >5% (passive)",
  "13F-HR": "Institutional holdings",
  "SD": "Specialized disclosure",
  "11-K": "Employee plan annual report",
  "NT 10-K": "Late filing notice (10-K)",
  "NT 10-Q": "Late filing notice (10-Q)",
  "25-NSE": "Delisting notice",
  "8-K/A": "Current report (amended)",
  "10-K/A": "Annual report (amended)",
  "10-Q/A": "Quarterly report (amended)",
};

/** 8-K item numbers → human labels (Regulation S-K). */
const ITEM_LABEL: Record<string, string> = {
  "1.01": "Entry into a material agreement",
  "1.02": "Termination of a material agreement",
  "1.03": "Bankruptcy or receivership",
  "1.05": "Material cybersecurity incident",
  "2.01": "Completion of acquisition or disposition",
  "2.02": "Results of operations and financial condition",
  "2.03": "Creation of a direct financial obligation",
  "2.04": "Triggering events (acceleration of obligation)",
  "2.05": "Costs associated with exit or disposal activities",
  "2.06": "Material impairments",
  "3.01": "Delisting notice or failure to satisfy listing rule",
  "3.02": "Unregistered sales of equity securities",
  "3.03": "Material modification to rights of security holders",
  "4.01": "Changes in registrant's certifying accountant",
  "4.02": "Non-reliance on previously issued financial statements (restatement)",
  "5.01": "Changes in control of registrant",
  "5.02": "Departure/appointment of directors or officers; compensation",
  "5.03": "Amendments to articles/bylaws; change in fiscal year",
  "5.07": "Submission of matters to a vote of security holders",
  "7.01": "Regulation FD disclosure",
  "8.01": "Other events",
  "9.01": "Financial statements and exhibits",
};

export class EdgarConnector implements WebSensorConnector {
  mode = "list" as const;
  metadata(): ConnectorMetadata {
    return { key: "edgar", name: "SEC EDGAR filings", sensorTypes: ["REST_API", "JSON"], description: "Company submissions (data.sec.gov) → filings list keyed by accession number, 8-K items decoded", version: "1.0.0" };
  }
  async fetch(endpoint: SensorEndpoint): Promise<Observation> {
    const obs = await httpFetchWithRetry(endpoint.id, endpoint.url, { etag: endpoint.etag, lastModified: endpoint.lastModified, accept: "application/json", headers: { "user-agent": UA, host: "data.sec.gov" }, userAgent: UA, timeoutMs: 40_000, maxBytes: 30 * 1024 * 1024 });
    return obs;
  }
  async normalize(endpoint: SensorEndpoint, obs: Observation): Promise<NormalizedContent> {
    if (!obs.body) throw new NormalizeError("no_body", "Observation has no body");
    const text = obs.body.toString("utf8");
    let j: Record<string, unknown>;
    try {
      j = JSON.parse(text) as Record<string, unknown>;
    } catch {
      throw new NormalizeError("bad_json", "EDGAR response is not JSON");
    }
    const recent = ((j.filings as Record<string, unknown> | undefined)?.recent ?? {}) as Record<string, unknown[]>;
    const forms = (recent.form ?? []) as string[];
    if (!Array.isArray(forms)) throw new NormalizeError("bad_shape", "filings.recent.form missing");
    const cfg = endpoint.config as { forms?: string[]; excludeForms?: string[]; maxItems?: number };
    const include = cfg.forms ? new Set(cfg.forms.map((f) => f.toUpperCase())) : null;
    const exclude = new Set((cfg.excludeForms ?? [...DEFAULT_EXCLUDE]).map((f) => f.toUpperCase()));
    const cik = String(j.cik ?? "").padStart(10, "0");
    const cikNum = String(Number(cik));
    const company = String(j.name ?? "");
    const tickers = (j.tickers ?? []) as string[];
    const col = (k: string, i: number): string => String((recent[k] as unknown[] | undefined)?.[i] ?? "");
    const items: { key: string; [k: string]: unknown }[] = [];
    for (let i = 0; i < forms.length && items.length < (cfg.maxItems ?? 60); i++) {
      const form = forms[i]!.toUpperCase();
      if (exclude.has(form)) continue;
      if (include && !include.has(form)) continue;
      const acc = col("accessionNumber", i);
      const accPath = acc.replace(/-/g, "");
      const doc = col("primaryDocument", i);
      const itemsRaw = col("items", i);
      const itemLabels = itemsRaw
        .split(",")
        .map((s) => s.trim())
        .filter(Boolean)
        .map((n) => `${n} ${ITEM_LABEL[n] ?? ""}`.trim());
      const filingDate = col("filingDate", i);
      const accepted = col("acceptanceDateTime", i);
      const desc = col("primaryDocDescription", i);
      const label = FORM_LABEL[form] ?? form;
      const title = `${company}: ${form} — ${label}${itemLabels.length ? ` (${itemLabels.map((l) => l.split(" ")[0]).join(", ")})` : ""}`;
      const summary = [desc && desc !== form ? desc : "", itemLabels.length ? `Items: ${itemLabels.join("; ")}` : "", col("reportDate", i) ? `Period: ${col("reportDate", i)}` : ""].filter(Boolean).join(". ");
      items.push({
        key: acc,
        form,
        formLabel: label,
        items: itemsRaw,
        title,
        summary: summary.slice(0, 800),
        url: doc ? `https://www.sec.gov/Archives/edgar/data/${cikNum}/${accPath}/${doc}` : `https://www.sec.gov/Archives/edgar/data/${cikNum}/${accPath}/`,
        indexUrl: `https://www.sec.gov/Archives/edgar/data/${cikNum}/${accPath}/${acc}-index.htm`,
        filingDate,
        publishedAt: accepted && !Number.isNaN(new Date(accepted).getTime()) ? new Date(accepted).toISOString() : filingDate ? new Date(filingDate + "T12:00:00Z").toISOString() : null,
        size: Number(col("size", i)) || 0,
        xbrl: col("isXBRL", i) === "1",
      });
    }
    const canonical = items.map((i) => `${i.key}\t${String(i.form)}\t${String(i.items)}`).join("\n");
    const newest = items.map((i) => i.publishedAt as string | null).filter((x): x is string => Boolean(x)).sort().at(-1);
    return {
      mode: "list",
      items,
      compareFields: ["form", "items"],
      title: `${company} (${tickers.join(", ") || "CIK " + cikNum}) — EDGAR filings`,
      rawHash: sha256(text),
      canonicalHash: sha256(canonical),
      semanticHash: simhash(items.map((i) => String(i.title)).join("\n")),
      publishedAt: newest ? new Date(newest) : null,
      extractionConfidence: 1,
      extra: { cik: cikNum, company, tickers, exchanges: j.exchanges, sic: j.sic, sicDescription: j.sicDescription, fiscalYearEnd: j.fiscalYearEnd, stateOfIncorporation: j.stateOfIncorporation, totalRecent: forms.length, filings: items.length },
    };
  }
}