Supercharge your AI agents with data from the web and beyond. Building the library for superintelligence. 🔥
TypeScript aiai-agentsai-crawlerai-scrapingai-searchcrawlerdata-extractionhtml-to-markdownllmmarkdownscraperscrapingweb-crawlerweb-dataweb-data-extractionweb-scraperweb-scrapingweb-searchwebscraping
import { describe, expect, it } from "vitest";
import {
imageContentTypeFromExtension,
imageExtensionFromContentType,
imageExtensionFromUrlPath,
sniffImageContentType,
sniffImageContentTypeFromBase64,
} from "./image-formats";
// The 12-byte JP2 signature box: length, "jP ", then the CR LF 0x87 LF check.
const JP2_SIGNATURE = [
0x00, 0x00, 0x00, 0x0c, 0x6a, 0x50, 0x20, 0x20, 0x0d, 0x0a, 0x87, 0x0a,
];
describe("image-formats", () => {
it("maps content types to extensions, ignoring parameters and case", () => {
expect(imageExtensionFromContentType("image/png")).toBe(".png");
expect(imageExtensionFromContentType("image/png;charset=UTF-8")).toBe(
".png",
);
expect(imageExtensionFromContentType("Image/JPEG")).toBe(".jpg");
expect(imageExtensionFromContentType("image/jpg")).toBe(".jpg");
expect(imageExtensionFromContentType("image/pjpeg")).toBe(".jpg");
expect(imageExtensionFromContentType("image/x-tiff")).toBe(".tif");
expect(imageExtensionFromContentType("image/tiff")).toBe(".tif");
expect(imageExtensionFromContentType("image/jp2")).toBe(".jp2");
expect(imageExtensionFromContentType("image/jpx")).toBe(".jp2");
expect(imageExtensionFromContentType("image/j2k")).toBe(".jp2");
expect(imageExtensionFromContentType("image/j2c")).toBe(".jp2");
expect(imageExtensionFromContentType("image/x-j2c")).toBe(".jp2");
expect(imageExtensionFromContentType("image/webp")).toBe(".webp");
expect(imageExtensionFromContentType("image/avif")).toBe(".avif");
expect(imageExtensionFromContentType("image/svg+xml")).toBeNull();
expect(imageExtensionFromContentType(undefined)).toBeNull();
});
it("maps extensions and their aliases back to content types", () => {
expect(imageContentTypeFromExtension(".jpg")).toBe("image/jpeg");
expect(imageContentTypeFromExtension(".JPEG")).toBe("image/jpeg");
expect(imageContentTypeFromExtension(".tiff")).toBe("image/tiff");
expect(imageContentTypeFromExtension(".tif")).toBe("image/tiff");
expect(imageContentTypeFromExtension(".png")).toBe("image/png");
expect(imageContentTypeFromExtension(".jp2")).toBe("image/jp2");
expect(imageContentTypeFromExtension(".JPX")).toBe("image/jp2");
expect(imageContentTypeFromExtension(".jpf")).toBe("image/jp2");
expect(imageContentTypeFromExtension(".j2k")).toBe("image/jp2");
expect(imageContentTypeFromExtension(".j2c")).toBe("image/jp2");
expect(imageContentTypeFromExtension(".webp")).toBe("image/webp");
expect(imageContentTypeFromExtension(".AVIF")).toBe("image/avif");
expect(imageContentTypeFromExtension(".svg")).toBeNull();
});
it("detects image extensions only at the end of a URL path", () => {
expect(imageExtensionFromUrlPath("/images/gallery/photo.jpg")).toBe(".jpg");
expect(imageExtensionFromUrlPath("/assets/scan.png")).toBe(".png");
expect(imageExtensionFromUrlPath("/scan.TIFF")).toBe(".tiff");
expect(imageExtensionFromUrlPath("/scans/page-0001.jp2")).toBe(".jp2");
expect(imageExtensionFromUrlPath("/tiles/region.j2k")).toBe(".j2k");
expect(imageExtensionFromUrlPath("/tiles/region.j2c")).toBe(".j2c");
expect(imageExtensionFromUrlPath("/photos/frame.jpf")).toBe(".jpf");
expect(imageExtensionFromUrlPath("/img.png/viewer")).toBeNull();
expect(imageExtensionFromUrlPath("/photo.webp")).toBe(".webp");
expect(imageExtensionFromUrlPath("/icons/logo.svg")).toBeNull();
expect(imageExtensionFromUrlPath("/paper.pdf")).toBeNull();
});
it("sniffs supported formats from magic bytes", () => {
expect(
sniffImageContentType(
Buffer.from([0x89, 0x50, 0x4e, 0x47, 0x0d, 0x0a, 0x1a, 0x0a]),
),
).toBe("image/png");
expect(sniffImageContentType(Buffer.from([0xff, 0xd8, 0xff, 0xe0]))).toBe(
"image/jpeg",
);
expect(sniffImageContentType(Buffer.from("GIF89a"))).toBe("image/gif");
expect(sniffImageContentType(Buffer.from([0x49, 0x49, 0x2a, 0x00]))).toBe(
"image/tiff",
);
expect(sniffImageContentType(Buffer.from([0x4d, 0x4d, 0x00, 0x2a]))).toBe(
"image/tiff",
);
expect(sniffImageContentType(Buffer.from("BM\x36\x00"))).toBe("image/bmp");
});
it("sniffs JPEG 2000 containers and raw codestreams", () => {
// Signature box followed by the file-type box header, as real encoders
// write it.
const jp2 = Buffer.concat([
Buffer.from(JP2_SIGNATURE),
Buffer.from("\x00\x00\x00\x14ftyp", "latin1"),
]);
expect(sniffImageContentType(jp2)).toBe("image/jp2");
// Raw codestream: SOC marker then SIZ marker.
expect(
sniffImageContentType(Buffer.from([0xff, 0x4f, 0xff, 0x51, 0x00, 0x2f])),
).toBe("image/jp2");
// The box length and type alone are not the full signature.
expect(
sniffImageContentType(Buffer.from(JP2_SIGNATURE.slice(0, 8))),
).toBeNull();
});
it("sniffs WebP and AVIF containers through their length fields", () => {
// RIFF header, a chunk size that differs per file, then the WEBP tag.
expect(
sniffImageContentType(Buffer.from("RIFF\x76\x11\x00\x00WEBPVP8L")),
).toBe("image/webp");
expect(
sniffImageContentType(Buffer.from("RIFF\x00\x00\x00\x00WEBPVP8 ")),
).toBe("image/webp");
// ftyp box of any size, still-image and image-sequence brands.
expect(
sniffImageContentType(Buffer.from("\x00\x00\x00\x20ftypavif\x00\x00")),
).toBe("image/avif");
expect(sniffImageContentType(Buffer.from("\x00\x00\x00\x1cftypavis"))).toBe(
"image/avif",
);
// The tag alone, without its container header, is not a match.
expect(sniffImageContentType(Buffer.from("WEBPVP8 "))).toBeNull();
expect(sniffImageContentType(Buffer.from("ftypavif"))).toBeNull();
});
it("rejects non-image and truncated payloads", () => {
expect(sniffImageContentType(Buffer.from("<!DOCTYPE html>"))).toBeNull();
expect(sniffImageContentType(Buffer.from("%PDF-1.7"))).toBeNull();
// A RIFF container that is not WebP (WAVE audio).
expect(
sniffImageContentType(Buffer.from("RIFF\x24\x00\x00\x00WAVEfmt ")),
).toBeNull();
// An ISO media file that is not AVIF (MP4 video).
expect(
sniffImageContentType(Buffer.from("\x00\x00\x00\x18ftypisom")),
).toBeNull();
expect(sniffImageContentType(Buffer.from([0x89, 0x50]))).toBeNull();
// Only the first six PNG bytes: not a file mupdf-style loaders accept.
expect(
sniffImageContentType(Buffer.from([0x89, 0x50, 0x4e, 0x47, 0x0d, 0x0a])),
).toBeNull();
expect(sniffImageContentType(Buffer.alloc(0))).toBeNull();
});
it("sniffs a base64 payload from its prefix alone", () => {
const png = Buffer.concat([
Buffer.from([0x89, 0x50, 0x4e, 0x47, 0x0d, 0x0a, 0x1a, 0x0a]),
Buffer.alloc(64, 1),
]).toString("base64");
expect(sniffImageContentTypeFromBase64(png)).toBe("image/png");
// The JP2 signature is the longest one; the prefix window must cover it.
const jp2 = Buffer.concat([
Buffer.from(JP2_SIGNATURE),
Buffer.alloc(64, 1),
]).toString("base64");
expect(sniffImageContentTypeFromBase64(jp2)).toBe("image/jp2");
expect(
sniffImageContentTypeFromBase64(Buffer.from("<html>").toString("base64")),
).toBeNull();
expect(sniffImageContentTypeFromBase64("")).toBeNull();
});
});