html2pdf-accessible 0.0.0-stage → 0.1.5

package/node_modules/@webgpu/types/package.json

@@ -0,0 +1,23 @@
+{
+  "name": "@webgpu/types",
+  "repository": "gpuweb/types",
+  "homepage": "https://github.com/gpuweb/types",
+  "bugs": "https://github.com/gpuweb/types/issues",
+  "version": "0.1.69",
+  "main": "",
+  "types": "dist/index.d.ts",
+  "license": "BSD-3-Clause",
+  "files": [
+    "dist/**/*"
+  ],
+  "scripts": {
+    "test": "for t in tests/*/ ; do echo \"\\n======== $t ========\" ; (cd \"$t\" && npm i && npm test); done",
+    "build-docs": "cd tsdoc-src && npm ci && npm run build",
+    "generate": "node generate.mjs",
+    "format": "prettier -w dist/index.d.ts"
+  },
+  "devDependencies": {
+    "bikeshed-to-ts": "file:third_party/bikeshed-to-ts",
+    "prettier": "^2.2.1"
+  }
+}

package/package.json

@@ -1,6 +1,127 @@
 {
   "name": "html2pdf-accessible",
-  "version": "0.0.0-stage",
-  "stub": true,
-  "description": "Temporary package placeholder for staged publishing"
-}
+  "description": "Accessible PDF generation using Skia. Supports vector graphics, text selection, and screen readers.",
+  "main": "lib/html2pdf-accessible.cjs.js",
+  "module": "lib/html2pdf-accessible.esm.mjs",
+  "typings": "lib/types/index.d.ts",
+  "browser": "lib/html2pdf-accessible.js",
+  "exports": {
+    ".": {
+      "types": "./lib/types/index.d.ts",
+      "import": "./lib/html2pdf-accessible.esm.mjs",
+      "require": "./lib/html2pdf-accessible.cjs.js"
+    },
+    "./lib/*": "./lib/*",
+    "./package.json": "./package.json"
+  },
+  "version": "0.1.5",
+  "author": {
+    "name": "Mathias Ciarlo",
+    "url": "https://github.com/MathiasCiarlo"
+  },
+  "license": "MIT",
+  "keywords": [
+    "html2pdf",
+    "pdf",
+    "html",
+    "accessibility",
+    "vector pdf",
+    "skia"
+  ],
+  "engines": {
+    "node": ">=22.0.0"
+  },
+  "files": [
+    "lib/",
+    "README.md",
+    "LICENSE",
+    "THIRD_PARTY_NOTICES.md",
+    "deps/canvaskit-pdf/package.json",
+    "deps/canvaskit-pdf/bin/canvaskit.js",
+    "deps/canvaskit-pdf/types",
+    "deps/canvaskit-pdf/LICENSE",
+    "deps/canvaskit-pdf/THIRD_PARTY_LICENSES.txt",
+    "deps/canvaskit-pdf/provenance.json",
+    "deps/canvaskit-pdf/README.md",
+    "deps/canvaskit-pdf/pdf_bindings.cpp",
+    "deps/canvaskit-pdf/CHROMIUM-LICENSE",
+    "deps/canvaskit-pdf/chromium-license-source.json",
+    "scripts/canvaskit/"
+  ],
+  "homepage": "https://github.com/MathiasCiarlo/html2pdf-accessible",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/MathiasCiarlo/html2pdf-accessible.git"
+  },
+  "devDependencies": {
+    "@napi-rs/canvas": "0.1.80",
+    "@types/jest": "^29.5.12",
+    "@types/node": "^20.14.12",
+    "@typescript-eslint/eslint-plugin": "^4.33.0",
+    "@typescript-eslint/parser": "^4.33.0",
+    "copy-webpack-plugin": "^13.0.0",
+    "eslint": "^7.32.0",
+    "eslint-config-prettier": "^9.1.0",
+    "eslint-plugin-prettier": "4.2.1",
+    "fast-xml-parser": "5.5.7",
+    "husky": "^8.0.0",
+    "jest": "^29.7.0",
+    "jest-environment-jsdom": "^29.7.0",
+    "lint-staged": "^15.2.7",
+    "mkdirp": "^3.0.1",
+    "path-browserify": "^1.0.1",
+    "pdfjs-dist": "5.4.296",
+    "playwright": "1.63.0",
+    "prettier": "^2.3.2",
+    "rimraf": "^6.0.1",
+    "terser-webpack-plugin": "^5.3.14",
+    "ts-jest": "^29.1.5",
+    "ts-loader": "^9.5.2",
+    "ts-node": "^10.9.2",
+    "tslib": "^2.3.0",
+    "typescript": "^5.5.4",
+    "webpack": "^5.100.0",
+    "webpack-cli": "^6.0.1"
+  },
+  "scripts": {
+    "prebuild": "npm run verify:canvaskit && rimraf lib/ && rimraf build/ && mkdirp lib && mkdirp build",
+    "build": "webpack --mode=production",
+    "prepublishOnly": "npm run build",
+    "build:dev": "webpack --mode=development",
+    "build:watch": "webpack --mode=development --watch",
+    "format": "prettier --write \"{src,tests,scripts}/**/*.ts\"",
+    "lint": "eslint src tests scripts *.js --ext .ts,.js,.cjs --max-warnings 0",
+    "test": "npm run unittest -- --runInBand && node --test tests/pdf/integration/*.test.cjs && npm run typecheck",
+    "unittest": "jest",
+    "test:pdf": "npm run build:dev && node scripts/pdf/run.cjs",
+    "test:pdf:serve": "node scripts/pdf/server.cjs",
+    "typecheck": "tsc --noEmit -p tests/tsconfig.json",
+    "test:pdf:visual": "node scripts/pdf/visual.cjs",
+    "test:pdf:validate": "node scripts/pdf/validate.cjs",
+    "test:pdf:update": "node scripts/pdf/visual.cjs --update",
+    "test:package": "node scripts/pdf/package.cjs",
+    "lint:changed": "node scripts/lint-changed.cjs",
+    "prebuild:dev": "npm run verify:canvaskit && rimraf lib/ && rimraf build/ && mkdirp lib && mkdirp build",
+    "build:canvaskit": "node scripts/canvaskit/build.cjs",
+    "test:canvaskit:rebuild": "node scripts/canvaskit/build.cjs --check",
+    "verify:canvaskit": "node scripts/canvaskit/verify.cjs"
+  },
+  "dependencies": {
+    "@html2pdf-skia/canvaskit-pdf": "file:deps/canvaskit-pdf",
+    "@pdf-lib/fontkit": "^1.1.1",
+    "@webgpu/types": "0.1.69",
+    "css-line-break": "^2.1.0",
+    "emoji-regex-xs": "^2.0.1",
+    "pdf-lib": "^1.17.1",
+    "text-segmentation": "^1.0.3"
+  },
+  "bundleDependencies": [
+    "@html2pdf-skia/canvaskit-pdf"
+  ],
+  "lint-staged": {
+    "{src,tests}/**/*.ts": [
+      "prettier --write",
+      "eslint --fix"
+    ]
+  }
+}

package/scripts/canvaskit/build.cjs

@@ -0,0 +1,74 @@
+const path = require("node:path");
+const fs = require("node:fs/promises");
+const { spawnSync } = require("node:child_process");
+const { verify } = require("./verify.cjs");
+const root = path.resolve(__dirname, "../..");
+const image =
+  "emscripten/emsdk:4.0.10@sha256:90b757eb11fa9a0e3ce4d2d9f76d932a56018e4accc37b5a28b2783751e60eb7";
+async function build() {
+  const result = spawnSync(
+    "docker",
+    [
+      "run",
+      "--rm",
+      "--mount",
+      "type=volume,source=html2pdf-skia-source-build,target=/build",
+      "--mount",
+      `type=bind,source=${root},target=/work`,
+      image,
+      "python3",
+      "/work/scripts/canvaskit/build.py",
+    ],
+    { stdio: "inherit" }
+  );
+  if (result.error) throw result.error;
+  if (result.status !== 0) throw new Error("CanvasKit source build failed");
+  const output = path.join(root, ".cache/canvaskit-built");
+  const target = path.join(root, "deps/canvaskit-pdf");
+  if (process.argv.includes("--check")) {
+    await verify(output);
+    for (const file of [
+      "canvaskit.js",
+      "canvaskit.wasm",
+      "google-types.d.ts",
+      "THIRD_PARTY_LICENSES.txt",
+      "provenance.json",
+    ]) {
+      const actual = await fs.readFile(path.join(output, file));
+      const expected = await fs.readFile(
+        path.join(
+          target,
+          file.endsWith(".wasm") || file.endsWith(".js")
+            ? "bin"
+            : file.endsWith(".d.ts")
+            ? "types"
+            : "",
+          file
+        )
+      );
+      if (!actual.equals(expected))
+        throw new Error(`Source rebuild differs: ${file}`);
+    }
+    console.log(
+      "Source-rebuilt binaries, types, licenses and provenance match included artifacts"
+    );
+  } else {
+    await fs.mkdir(path.join(target, "bin"), { recursive: true });
+    for (const file of ["canvaskit.js", "canvaskit.wasm"])
+      await fs.copyFile(
+        path.join(output, file),
+        path.join(target, "bin", file)
+      );
+    for (const file of ["provenance.json", "THIRD_PARTY_LICENSES.txt"])
+      await fs.copyFile(path.join(output, file), path.join(target, file));
+    await fs.copyFile(
+      path.join(output, "google-types.d.ts"),
+      path.join(target, "types/google-types.d.ts")
+    );
+    await verify();
+  }
+}
+build().catch((error) => {
+  console.error(error);
+  process.exitCode = 1;
+});

package/scripts/canvaskit/build.py

@@ -0,0 +1,143 @@
+#!/usr/bin/env python3
+"""Build our small PDF bridge with immutable Google Skia sources in Docker."""
+import concurrent.futures
+import hashlib
+import json
+import pathlib
+import runpy
+import shutil
+import subprocess
+import time
+
+ROOT = pathlib.Path("/work")
+SKIA = pathlib.Path("/build/skia")
+REVISION = "e2def78fa232884d4aedcf3ad92d4a0dd673b898"
+IMAGE = "emscripten/emsdk:4.0.10@sha256:90b757eb11fa9a0e3ce4d2d9f76d932a56018e4accc37b5a28b2783751e60eb7"
+
+
+def run(*args, **kwargs):
+    subprocess.run(args, check=True, **kwargs)
+
+
+def digest(path):
+    return hashlib.sha256(path.read_bytes()).hexdigest()
+
+
+if not (SKIA / ".git").exists():
+    run("git", "clone", "--depth", "1", "--branch", "canvaskit/0.42.0",
+        "https://skia.googlesource.com/skia.git", str(SKIA))
+head = subprocess.check_output(["git", "rev-parse", "HEAD"], cwd=SKIA, text=True).strip()
+if head != REVISION:
+    raise RuntimeError("Unexpected Skia revision: " + head)
+run("git", "config", "--global", "http.version", "HTTP/1.1")
+
+# Check cached sources before executing any upstream build tooling.
+run("git", "restore", "--source", REVISION, "--", "modules/canvaskit/BUILD.gn",
+    "modules/canvaskit/compile.sh", cwd=SKIA)
+if subprocess.check_output(["git", "diff", "HEAD", "--name-only"], cwd=SKIA, text=True).strip():
+    raise RuntimeError("Unexpected modified Google Skia source")
+
+# Use Google's checkout implementation, but fetch only libraries needed by the
+# CPU/PDF build and limit parallelism to avoid transient server transport errors.
+sync = runpy.run_path(str(SKIA / "tools/git-sync-deps"))
+deps = sync["parse_file_to_dict"](str(SKIA / "DEPS"))["deps"]
+needed = {"buildtools", "brotli", "freetype", "harfbuzz", "highway", "libjpeg-turbo",
+          "libpng", "libwebp", "wuffs", "zlib", "perfetto", "libgrapheme", "expat", "partition_alloc", "icu"}
+selected = {path: value for path, value in deps.items()
+            if path.split("/")[-1] in needed}
+if len(selected) != len(needed):
+    raise RuntimeError("Skia dependency list changed")
+
+
+def checkout(entry):
+    path, value = entry
+    repo, revision = value.split("@", 1)
+    for attempt in range(3):
+        try:
+            sync["git_checkout_to_directory"]("git", repo, revision,
+                str(SKIA / path), True, True)
+            actual = subprocess.check_output(["git", "rev-parse", "HEAD"],
+                                            cwd=SKIA / path, text=True).strip()
+            expected = subprocess.check_output(["git", "rev-parse", revision + "^{}"],
+                                               cwd=SKIA / path, text=True).strip()
+            if actual != expected:
+                raise RuntimeError("Dependency revision mismatch: " + path)
+            dirty = subprocess.check_output(["git", "status", "--porcelain", "--untracked-files=no"],
+                                            cwd=SKIA / path, text=True).strip()
+            if dirty:
+                raise RuntimeError("Modified dependency source: " + path)
+            return
+        except subprocess.CalledProcessError:
+            if attempt == 2:
+                raise
+            time.sleep(2 ** attempt)
+
+
+with concurrent.futures.ThreadPoolExecutor(max_workers=4) as pool:
+    list(pool.map(checkout, selected.items()))
+
+# Restore exactly the files our overlay modifies, even in a cached build volume.
+run("git", "restore", "--source", REVISION, "--", "modules/canvaskit/BUILD.gn",
+    "modules/canvaskit/compile.sh", cwd=SKIA)
+if subprocess.check_output(["git", "diff", "--name-only"], cwd=SKIA, text=True).strip():
+    raise RuntimeError("Unexpected modified Google Skia source")
+module = SKIA / "modules/canvaskit"
+bridge = ROOT / "deps/canvaskit-pdf/pdf_bindings.cpp"
+shutil.copyfile(bridge, module / "pdf_bindings.cpp")
+build = (module / "BUILD.gn").read_text()
+if build.count('    "canvaskit_bindings.cpp",') != 1:
+    raise RuntimeError("CanvasKit build layout changed")
+build = build.replace('    "canvaskit_bindings.cpp",',
+                      '    "canvaskit_bindings.cpp",\n    "pdf_bindings.cpp",')
+(module / "BUILD.gn").write_text(build)
+compile_script = (module / "compile.sh").read_text()
+for expected in ["skia_enable_pdf=false", '--args="is_debug=', "-k 10"]:
+    if compile_script.count(expected) != 1:
+        raise RuntimeError("CanvasKit compile script changed")
+compile_script = compile_script.replace("skia_enable_pdf=false", "skia_enable_pdf=true")
+compile_script = compile_script.replace('--args="is_debug=',
+    '--args="skia_emsdk_dir=\\\"/emsdk\\\" is_debug=')
+compile_script = compile_script.replace("-k 10", "-k 1 -j 8")
+(module / "compile.sh").write_text(compile_script)
+run("bash", str(module / "compile.sh"), "cpu", "no_skottie",
+    "client_unicode", "no_embedded_font", "no_woff2", "force_tracing", cwd=SKIA)
+
+out = ROOT / ".cache/canvaskit-built"
+out.mkdir(parents=True, exist_ok=True)
+for filename in ["canvaskit.js", "canvaskit.wasm"]:
+    shutil.copyfile(SKIA / "out/canvaskit_wasm" / filename, out / filename)
+shutil.copyfile(module / "npm_build/types/index.d.ts", out / "google-types.d.ts")
+chromium_license = ROOT / "deps/canvaskit-pdf/CHROMIUM-LICENSE"
+chromium_source = json.loads((ROOT / "deps/canvaskit-pdf/chromium-license-source.json").read_text())
+if digest(chromium_license) != chromium_source["sha256"]:
+    raise RuntimeError("Chromium license checksum mismatch")
+notices = []
+for library in sorted(needed - {"buildtools"}):
+    folder = SKIA / ("buildtools" if library == "buildtools" else "third_party/externals/" + library)
+    files = sorted(set(path for pattern in ["LICENSE*", "COPYING*", "NOTICE*", "README.ijg", "docs/FTL.TXT"]
+                       for path in folder.glob(pattern) if path.is_file()))
+    if library == "partition_alloc":
+        notices.append("=== partition_alloc: " + chromium_source["url"] + " ===\n" + chromium_license.read_text())
+        continue
+    if not files:
+        raise RuntimeError("Missing license notice: " + library)
+    for path in files:
+        notices.append("=== " + str(path.relative_to(SKIA)) + " ===\n" + path.read_text(errors="replace"))
+(out / "THIRD_PARTY_LICENSES.txt").write_text("\n\n".join(notices) + "\n")
+manifest = {
+    "skiaRevision": REVISION,
+    "skiaRelease": "canvaskit/0.42.0",
+    "source": "https://skia.googlesource.com/skia.git",
+    "buildImage": IMAGE,
+    "bridgeSha256": digest(bridge),
+    "buildScriptSha256": digest(pathlib.Path(__file__)),
+    "depsSha256": digest(SKIA / "DEPS"),
+    "typeDefinitionsSha256": digest(out / "google-types.d.ts"),
+    "licensesSha256": digest(out / "THIRD_PARTY_LICENSES.txt"),
+    "dependencies": selected,
+    "chromiumLicenseSource": chromium_source,
+    "artifacts": {filename: digest(out / filename)
+                  for filename in ["canvaskit.js", "canvaskit.wasm"]},
+}
+(out / "provenance.json").write_text(json.dumps(manifest, indent=2) + "\n")
+print("Built CanvasKit PDF from Google Skia " + REVISION, flush=True)

package/scripts/canvaskit/verify.cjs

@@ -0,0 +1,82 @@
+const fs = require("node:fs/promises");
+const path = require("node:path");
+const { createHash } = require("node:crypto");
+const assert = require("node:assert/strict");
+const root = path.resolve(__dirname, "../..");
+const backend = path.join(root, "deps/canvaskit-pdf");
+const digest = (buffer) => createHash("sha256").update(buffer).digest("hex");
+async function verify(output) {
+  const manifest = JSON.parse(
+    await fs.readFile(path.join(output || backend, "provenance.json"), "utf8")
+  );
+  assert.equal(
+    manifest.skiaRevision,
+    "e2def78fa232884d4aedcf3ad92d4a0dd673b898"
+  );
+  assert.equal(manifest.source, "https://skia.googlesource.com/skia.git");
+  assert.equal(
+    manifest.buildImage,
+    "emscripten/emsdk:4.0.10@sha256:90b757eb11fa9a0e3ce4d2d9f76d932a56018e4accc37b5a28b2783751e60eb7"
+  );
+  assert.equal(
+    digest(await fs.readFile(path.join(backend, "pdf_bindings.cpp"))),
+    manifest.bridgeSha256,
+    "PDF bridge changed without a source rebuild"
+  );
+  assert.equal(
+    digest(await fs.readFile(path.join(__dirname, "build.py"))),
+    manifest.buildScriptSha256,
+    "Build script changed without a source rebuild"
+  );
+  assert.equal(
+    digest(
+      await fs.readFile(
+        path.join(output || path.join(backend, "types"), "google-types.d.ts")
+      )
+    ),
+    manifest.typeDefinitionsSha256,
+    "Google type definitions differ"
+  );
+  assert.equal(
+    digest(
+      await fs.readFile(
+        path.join(output || backend, "THIRD_PARTY_LICENSES.txt")
+      )
+    ),
+    manifest.licensesSha256,
+    "License notices differ"
+  );
+  const chromium = JSON.parse(
+    await fs.readFile(
+      path.join(backend, "chromium-license-source.json"),
+      "utf8"
+    )
+  );
+  assert.deepEqual(chromium, manifest.chromiumLicenseSource);
+  assert.equal(
+    digest(await fs.readFile(path.join(backend, "CHROMIUM-LICENSE"))),
+    chromium.sha256
+  );
+  for (const file of ["canvaskit.js", "canvaskit.wasm"]) {
+    let artifact = path.join(output || path.join(backend, "bin"), file);
+    if (!output && file.endsWith(".wasm")) {
+      try {
+        await fs.access(artifact);
+      } catch {
+        artifact = path.join(root, "lib/wasm/canvaskit-pdf.wasm");
+      }
+    }
+    assert.equal(
+      digest(await fs.readFile(artifact)),
+      manifest.artifacts[file],
+      `Artifact checksum mismatch: ${file}`
+    );
+  }
+  console.log("CanvasKit source manifest and artifact checksums verified");
+}
+module.exports = { verify };
+if (require.main === module)
+  verify().catch((error) => {
+    console.error(error);
+    process.exitCode = 1;
+  });