import { describe, expect, it } from "vitest";
import {
  LocalHeuristicOcrProvider,
  sanitizeOcrDisplayText,
} from "@/services/ocr";

describe("LocalHeuristicOcrProvider", () => {
  it("extracts CJK from buffer", async () => {
    const buf = Buffer.from("前置 卡转海 覆盖贴 后缀", "utf8");
    const r = await new LocalHeuristicOcrProvider().recognize(buf);
    expect(r.text).toMatch(/卡转海|覆盖贴/);
    expect(r.provider).toBe("local");
  });

  it("does not treat PDF binary noise as Chinese", async () => {
    const junk = Buffer.from("%PDF-1.4\n" + "4稍擁咥\n涅樺\n膽蟈籃\n", "utf8");
    const r = await new LocalHeuristicOcrProvider().recognize(junk, "application/pdf");
    expect(r.text).not.toMatch(/稍擁|涅樺|膽蟈/);
  });
});

describe("sanitizeOcrDisplayText", () => {
  it("strips 【OCR:file】 wrappers and garbage lines", () => {
    const raw = [
      "【OCR:WL103276-SBD1(1).pdf】",
      "4稍擁咥",
      "【OCR:WL103944-WL103940.pdf】",
      "请查收新增预报 柜号 ABCD1234567",
    ].join("\n");
    const out = sanitizeOcrDisplayText(raw);
    expect(out).not.toMatch(/【OCR:/);
    expect(out).not.toMatch(/稍擁咥/);
    expect(out).toMatch(/请查收新增预报/);
  });
});