OpenAI's Codex Security CLI and TypeScript SDK for finding, validating, and fixing security vulnerabilities. npm: https://www.npmjs.com/package/@openai/codex-security
TypeScript ai-securityapplication-securityclicode-scanningcodexcodex-securitycybersecuritydevsecopsnodejsnpmopenaisdksecuritytypescriptvulnerability-scanning
fix(sdk): honor cancellation during PDF knowledge-base extraction (#1529)
Peter73abb734f72b
diff --git a/sdk/typescript/src/knowledge-base.ts b/sdk/typescript/src/knowledge-base.tsindex 2df9c138b..307b7cac6 100644--- a/sdk/typescript/src/knowledge-base.ts+++ b/sdk/typescript/src/knowledge-base.ts@@ -97,7 +97,7 @@ export async function readKnowledgeBaseSnapshot(const extension = extname(document).toLowerCase();const text =extension === ".pdf"- ? await extractPdf(document, bytes)+ ? await extractPdf(document, bytes, signal): extension === ".docx"? extractDocx(document, bytes): decodeText(document, bytes);@@ -251,7 +251,11 @@ function decodeText(path: string, bytes: Uint8Array): string {}}-async function extractPdf(path: string, bytes: Uint8Array): Promise<string> {+async function extractPdf(+ path: string,+ bytes: Uint8Array,+ signal?: AbortSignal,+): Promise<string> {try {const { getDocument, VerbosityLevel } =await import("pdfjs-dist/legacy/build/pdf.mjs");@@ -264,7 +268,9 @@ async function extractPdf(path: string, bytes: Uint8Array): Promise<string> {const document = await loadingTask.promise;const pages: string[] = [];for (let number = 1; number <= document.numPages; number++) {+ signal?.throwIfAborted();const content = await (await document.getPage(number)).getTextContent();+ signal?.throwIfAborted();pages.push(content.items.map((item) => ("str" in item ? item.str : ""))@@ -276,6 +282,7 @@ async function extractPdf(path: string, bytes: Uint8Array): Promise<string> {await loadingTask.destroy();}} catch (error) {+ if (signal?.aborted) throw error;throw new Error(`Cannot extract text from knowledge base PDF: ${path}`, {cause: error,});diff --git a/sdk/typescript/tests-ts/knowledge-base.test.ts b/sdk/typescript/tests-ts/knowledge-base.test.tsindex 71eff27be..9786f6419 100644--- a/sdk/typescript/tests-ts/knowledge-base.test.ts+++ b/sdk/typescript/tests-ts/knowledge-base.test.ts@@ -12,7 +12,7 @@ import {import * as filesystem from "node:fs/promises";import * as os from "node:os";import { join } from "node:path";-import { afterEach, describe, expect, spyOn, test } from "bun:test";+import { afterEach, describe, expect, mock, spyOn, test } from "bun:test";import { strToU8, zipSync } from "fflate";import { ConfigurationError } from "../src/errors.js";import {@@ -521,6 +521,44 @@ describe("scan knowledge bases", () => {},);+ test("cancels PDF extraction when the scan is aborted", async () => {+ const root = await temporaryDirectory();+ const source = join(root, "large.pdf");+ await writeFile(source, pdf("Cancellable content"));+ const pdfjs = await import("pdfjs-dist/legacy/build/pdf.mjs");+ const controller = new AbortController();+ const reason = new Error("Synthetic cancellation.");+ let pagesRead = 0;+ const destroy = mock(async () => {});+ const parser = spyOn(pdfjs, "getDocument").mockImplementation(+ () =>+ ({+ promise: Promise.resolve({+ numPages: 3,+ getPage: async () => {+ pagesRead += 1;+ return {+ getTextContent: async () => {+ controller.abort(reason);+ return { items: [{ str: "page" }] };+ },+ };+ },+ }),+ destroy,+ }) as never,+ );++ try {+ const prepared = prepareKnowledgeBase([source], controller.signal);+ await expect(prepared).rejects.toBe(reason);+ expect(pagesRead).toBe(1);+ expect(destroy).toHaveBeenCalledTimes(1);+ } finally {+ parser.mockRestore();+ }+ });+test("preserves the local origin and cause when snapshot staging fails", async () => {const root = await temporaryDirectory();const parent = join(root, "not-a-directory");