Initial release of File Tools 0.1.0

2026-09-01 02:55:06 +02:00
commit 055f533cf1
62 changed files with 9050 additions and 0 deletions
@@ -0,0 +1,55 @@
+import { describe, expect, it } from "vitest";
+import { manifestCsv, manifestJson } from "../../src/file/manifest";
+import {
+  detectKnownSignature,
+  extensionMismatch,
+} from "../../src/file/signatures";
+import {
+  extractStrings,
+  looksLikeText,
+  shannonEntropy,
+} from "../../src/file/text";
+
+describe("file signatures", () => {
+  it("detects PNG without claiming parser validation", () =>
+    expect(
+      detectKnownSignature(
+        Uint8Array.of(0x89, 0x50, 0x4e, 0x47, 0x0d, 0x0a, 0x1a, 0x0a),
+      ),
+    ).toMatchObject({ extension: "png", mime: "image/png" }));
+  it("accounts for container extension aliases", () => {
+    expect(extensionMismatch("book.epub", "zip")).toBeUndefined();
+    expect(extensionMismatch("image.txt", "png")).toMatch(/does not match/u);
+  });
+});
+
+describe("bounded text helpers", () => {
+  it("extracts ASCII and UTF-16LE strings", () =>
+    expect(
+      extractStrings(
+        Uint8Array.from([65, 66, 67, 68, 0, 90, 0, 89, 0, 88, 0, 87, 0]),
+      ),
+    ).toEqual(expect.arrayContaining(["ABCD", "ZYXW"])));
+  it("distinguishes NUL-containing binary data", () =>
+    expect(looksLikeText(Uint8Array.of(65, 0, 66))).toBe(false));
+  it("measures a uniform byte distribution", () =>
+    expect(
+      shannonEntropy(Uint8Array.from({ length: 256 }, (_, index) => index)),
+    ).toBeCloseTo(8));
+});
+
+describe("manifests", () => {
+  const records = [
+    {
+      name: "=formula",
+      path: "=formula",
+      size: 1,
+      type: "text/plain",
+      lastModified: 0,
+    },
+  ];
+  it("is deterministic", () =>
+    expect(manifestJson(records)).toBe(manifestJson(records)));
+  it("neutralises spreadsheet-leading formulas", () =>
+    expect(manifestCsv(records)).toContain("'=formula"));
+});