feat(tools): 2.0 phase 5 wave 4 - office, ebooks, data, archives (14 tools) (#224)

tests/unit/ai/ssrf-prescan-regex.test.ts

@@ -0,0 +1,67 @@
+// Pure-regex unit test for the SSRF pre-scan in doc_html_pdf.py.
+// Runs python3 against the actual module regexes (no weasyprint needed).
+// Mirrors the step-3 verification matrix from the fix ticket.
+
+import { spawnSync } from "node:child_process";
+import { join } from "node:path";
+import { describe, expect, it } from "vitest";
+import { hasPython } from "../../helpers/python-gate.js";
+
+const SCRIPT_DIR = join(process.cwd(), "packages", "ai", "python");
+
+/** Run a python3 snippet that imports regexes from the actual module and tests a case. */
+function testRegex(html: string, patternName: string): boolean {
+  const code = [
+    "import sys, os",
+    `sys.path.insert(0, ${JSON.stringify(SCRIPT_DIR)})`,
+    "from doc_html_pdf import _REMOTE_REF_RE, _REMOTE_CSS_URL_RE",
+    `pat = ${patternName}`,
+    `m = pat.search(${JSON.stringify(html)})`,
+    "print('MATCH' if m else 'NO_MATCH')",
+  ].join("; ");
+  const res = spawnSync("python3", ["-c", code], { encoding: "utf8", timeout: 5000 });
+  if (res.status !== 0) throw new Error(`python3 failed: ${res.stderr}`);
+  return res.stdout.trim() === "MATCH";
+}
+
+describe.skipIf(!hasPython)("SSRF pre-scan regexes (doc_html_pdf.py)", () => {
+  describe("_REMOTE_REF_RE", () => {
+    it("matches double-slash http img", () => {
+      expect(testRegex('<img src="http://h/x.png">', "_REMOTE_REF_RE")).toBe(true);
+    });
+
+    it("matches single-slash http img (pandoc rewrite)", () => {
+      expect(testRegex('<img src="http:/h/x.png">', "_REMOTE_REF_RE")).toBe(true);
+    });
+
+    it("matches double-slash https img", () => {
+      expect(testRegex('<img src="https://h/x.png">', "_REMOTE_REF_RE")).toBe(true);
+    });
+
+    it("matches single-slash https img", () => {
+      expect(testRegex('<img src="https:/h/x.png">', "_REMOTE_REF_RE")).toBe(true);
+    });
+
+    it("does NOT match data: URIs", () => {
+      expect(testRegex('<img src="data:image/png;base64,AA==">', "_REMOTE_REF_RE")).toBe(false);
+    });
+
+    it("does NOT match relative paths", () => {
+      expect(testRegex('<a href="page2.xhtml">', "_REMOTE_REF_RE")).toBe(false);
+    });
+  });
+
+  describe("_REMOTE_CSS_URL_RE", () => {
+    it("matches single-slash CSS url()", () => {
+      expect(testRegex("url(https:/h/a.css)", "_REMOTE_CSS_URL_RE")).toBe(true);
+    });
+
+    it("matches double-slash CSS url()", () => {
+      expect(testRegex("url(https://h/a.css)", "_REMOTE_CSS_URL_RE")).toBe(true);
+    });
+
+    it("does NOT match data: CSS url()", () => {
+      expect(testRegex("url(data:image/png;base64,AA==)", "_REMOTE_CSS_URL_RE")).toBe(false);
+    });
+  });
+});

tests/unit/doc-engine/libreoffice.test.ts

@@ -0,0 +1,27 @@
+import { parseConvertTarget } from "@snapotter/doc-engine";
+import { describe, expect, it } from "vitest";
+
+describe("parseConvertTarget", () => {
+  it("returns a bare extension unchanged", () => {
+    const result = parseConvertTarget("pdf");
+    expect(result).toEqual({ ext: "pdf", convertTo: "pdf" });
+  });
+
+  it("splits on the first colon for qualified filter targets", () => {
+    const result = parseConvertTarget("docx:MS Word 2007 XML");
+    expect(result).toEqual({ ext: "docx", convertTo: "docx:MS Word 2007 XML" });
+  });
+
+  it("handles filter names with colons after the first", () => {
+    const result = parseConvertTarget("csv:Text - txt - csv (StarCalc):44,34,76,1");
+    expect(result).toEqual({
+      ext: "csv",
+      convertTo: "csv:Text - txt - csv (StarCalc):44,34,76,1",
+    });
+  });
+
+  it("handles single-character extensions", () => {
+    const result = parseConvertTarget("a:SomeFilter");
+    expect(result).toEqual({ ext: "a", convertTo: "a:SomeFilter" });
+  });
+});

tests/unit/doc-engine/pandoc.test.ts

@@ -0,0 +1,34 @@
+import { readFileSync } from "node:fs";
+import { mkdtemp, rm, writeFile } from "node:fs/promises";
+import { tmpdir } from "node:os";
+import { join } from "node:path";
+import { pandocAvailable, runPandoc } from "@snapotter/doc-engine";
+import { afterAll, beforeAll, describe, expect, it } from "vitest";
+
+describe("pandocAvailable", () => {
+  it("returns a boolean without throwing", () => {
+    expect(typeof pandocAvailable()).toBe("boolean");
+  });
+});
+
+describe.skipIf(!pandocAvailable())("runPandoc (requires pandoc)", () => {
+  let tmpDir: string;
+
+  beforeAll(async () => {
+    tmpDir = await mkdtemp(join(tmpdir(), "pandoc-test-"));
+  });
+
+  afterAll(async () => {
+    await rm(tmpDir, { recursive: true, force: true }).catch(() => {});
+  });
+
+  it("converts a trivial markdown file to html", async () => {
+    const mdPath = join(tmpDir, "test.md");
+    const htmlPath = join(tmpDir, "test.html");
+    await writeFile(mdPath, "# Hello\n\nWorld\n");
+    await runPandoc(mdPath, htmlPath, { extraArgs: ["--standalone"] });
+    const html = readFileSync(htmlPath, "utf8");
+    expect(html).toContain("Hello");
+    expect(html).toContain("<");
+  }, 30_000);
+});

tests/unit/shared/modality.test.ts

@@ -24,7 +24,11 @@ describe("modality metadata", () => {
     for (const tool of TOOLS) {
       expect(validModalities).toContain(tool.modality);
       expect(Array.isArray(tool.acceptedInputs)).toBe(true);
-      expect(tool.acceptedInputs.length).toBeGreaterThan(0);
+      // Empty acceptedInputs is valid for file-modality tools that accept any file
+      // (e.g. create-zip); the factory 415 gate skips when the list is empty.
+      if (tool.modality !== "file" || tool.acceptedInputs.length > 0) {
+        expect(tool.acceptedInputs.length).toBeGreaterThan(0);
+      }
       expect(["fast", "long"]).toContain(tool.executionHint);
     }
   });