External/SnapOtter
1
0
Fork 0
mirror of https://github.com/snapotter-hq/SnapOtter.git synced 2026-08-03 07:46:42 +02:00
Code Issues Packages Projects Releases Wiki Activity

fix: make OCR portable and reliable across AMD64 and ARM64 (#519)

Browse Source 
* fix: make OCR portable and reliable

* fix: harden OCR installation portability

* fix: pin OCR partials across downloads

* fix: make OCR execution reliably asynchronous

* fix: harden OCR portability and docs routes

* fix: preserve decoder and docs safeguards
This commit is contained in:
SnapOtter
2026-07-15 03:34:24 +08:00
committed by GitHub
parent 58121f205f
commit 991c981529
409 changed files with 67151 additions and 8076 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/integration/tools/document/ocr-pdf.test.ts
+139 -58
View File
@@ -1,14 +1,12 @@
/**
* Integration tests for the ocr-pdf tool (/api/v1/tools/pdf/ocr-pdf).
*
* The OCR bundle (PaddleOCR / Tesseract) is 3-4 GB and not installed locally
* or in any verification environment this wave. The 501 gate is always hit.
* The bundle-gated happy path lives in a skipped describe for future
* in-container-after-install runs. The 501 contract + python-side conventions
* carry the verification.
* Fast PDF OCR is built in. Balanced and Best are optional accurate-pack
* capabilities and are rejected before enqueue when no healthy v3 runtime is
* active.
*/
import { afterAll, beforeAll, describe, expect, it } from "vitest";
import { afterAll, beforeAll, beforeEach, describe, expect, it, vi } from "vitest";
import { fixtures, readFixture } from "../../../fixtures/index.js";
import {
buildTestApp,
@@ -19,6 +17,27 @@ import {
const PDF = readFixture(fixtures.document.pdf3);
const mocks = vi.hoisted(() => ({
enqueueToolJob: vi.fn(),
getOcrRuntimeCapability: vi.fn(),
}));
vi.mock("@snapotter/ai", async (importOriginal) => {
const actual = await importOriginal<typeof import("@snapotter/ai")>();
return {
...actual,
getOcrRuntimeCapability: mocks.getOcrRuntimeCapability,
};
});
vi.mock("../../../../apps/api/src/jobs/enqueue.js", async (importOriginal) => {
const actual = await importOriginal<typeof import("../../../../apps/api/src/jobs/enqueue.js")>();
return {
...actual,
enqueueToolJob: mocks.enqueueToolJob,
};
});
let testApp: TestApp;
let app: TestApp["app"];
let adminToken: string;
@@ -33,10 +52,19 @@ afterAll(async () => {
await testApp.cleanup();
}, 10_000);
describe("ocr-pdf", () => {
// -- 501 gate (always fires locally: OCR bundle never installed) --
beforeEach(() => {
mocks.enqueueToolJob.mockReset();
mocks.enqueueToolJob.mockResolvedValue(undefined);
mocks.getOcrRuntimeCapability.mockReset();
mocks.getOcrRuntimeCapability.mockReturnValue({
available: false,
qualities: [],
providers: [],
});
});
it("returns 501 FEATURE_NOT_INSTALLED when bundle is absent", async () => {
describe("ocr-pdf", () => {
it("enqueues built-in Fast OCR when the accurate pack is absent", async () => {
const { body, contentType } = createMultipartPayload([
{
name: "file",
@@ -44,7 +72,70 @@ describe("ocr-pdf", () => {
contentType: "application/pdf",
content: PDF,
},
{ name: "settings", content: JSON.stringify({}) },
{ name: "settings", content: JSON.stringify({ quality: "fast", pages: "1" }) },
]);
const res = await app.inject({
method: "POST",
url: "/api/v1/tools/pdf/ocr-pdf",
headers: {
authorization: `Bearer ${adminToken}`,
"content-type": contentType,
},
body,
});
expect(res.statusCode).toBe(202);
const json = JSON.parse(res.body);
expect(json.jobId).toBeDefined();
expect(json.async).toBe(true);
expect(mocks.enqueueToolJob).toHaveBeenCalledWith(
expect.objectContaining({
toolId: "ocr-pdf",
pool: "ai",
settings: expect.objectContaining({ quality: "fast", pages: "1" }),
}),
);
});
it("defaults to Fast when the accurate pack is absent", async () => {
const { body, contentType } = createMultipartPayload([
{
name: "file",
filename: "test.pdf",
contentType: "application/pdf",
content: PDF,
},
{ name: "settings", content: JSON.stringify({ pages: "1" }) },
]);
const res = await app.inject({
method: "POST",
url: "/api/v1/tools/pdf/ocr-pdf",
headers: {
authorization: `Bearer ${adminToken}`,
"content-type": contentType,
},
body,
});
expect(res.statusCode).toBe(202);
expect(mocks.enqueueToolJob).toHaveBeenCalledWith(
expect.objectContaining({
settings: expect.objectContaining({ quality: "fast", pages: "1" }),
}),
);
});
it("maps the legacy paddleocr engine to Balanced like batch and pipeline ingress", async () => {
const { body, contentType } = createMultipartPayload([
{
name: "file",
filename: "test.pdf",
contentType: "application/pdf",
content: PDF,
},
{ name: "settings", content: JSON.stringify({ engine: "paddleocr", pages: "1" }) },
]);
const res = await app.inject({
@@ -58,11 +149,40 @@ describe("ocr-pdf", () => {
});
expect(res.statusCode).toBe(501);
const json = JSON.parse(res.body);
expect(json.code).toBe("FEATURE_NOT_INSTALLED");
expect(json.feature).toBe("ocr");
expect(json.featureName).toBe("OCR");
expect(json.estimatedSize).toBeDefined();
expect(JSON.parse(res.body)).toMatchObject({
code: "FEATURE_NOT_INSTALLED",
feature: "ocr",
requestedQuality: "balanced",
});
});
it("returns 501 FEATURE_NOT_INSTALLED for an unavailable accurate tier", async () => {
const { body, contentType } = createMultipartPayload([
{
name: "file",
filename: "test.pdf",
contentType: "application/pdf",
content: PDF,
},
{ name: "settings", content: JSON.stringify({ quality: "best" }) },
]);
const res = await app.inject({
method: "POST",
url: "/api/v1/tools/pdf/ocr-pdf",
headers: {
authorization: `Bearer ${adminToken}`,
"content-type": contentType,
},
body,
});
expect(res.statusCode).toBe(501);
expect(JSON.parse(res.body)).toMatchObject({
code: "FEATURE_NOT_INSTALLED",
feature: "ocr",
requestedQuality: "best",
});
});
// -- Auth gate --
@@ -88,9 +208,7 @@ describe("ocr-pdf", () => {
expect(res.statusCode).toBe(401);
});
// -- Validation (501 fires before settings parse, so bad quality also 501s) --
it("returns 501 even with invalid quality (gate fires first)", async () => {
it("validates settings before checking the optional pack", async () => {
const { body, contentType } = createMultipartPayload([
{
name: "file",
@@ -114,46 +232,9 @@ describe("ocr-pdf", () => {
body,
});
// 501 because the bundle gate fires before settings validation
expect(res.statusCode).toBe(501);
expect(res.statusCode).toBe(400);
const json = JSON.parse(res.body);
expect(json.code).toBe("FEATURE_NOT_INSTALLED");
});
// -- Bundle-gated happy path (skipped: OCR bundle is 3-4 GB) --
// The OCR bundle is not installed in any verification environment this wave.
// These tests exist for future in-container runs after bundle install.
// The 501 contract + python-side conventions carry the verification.
describe.skip("with ocr bundle installed", () => {
it("extracts text from PDF (202 + async)", async () => {
const { body, contentType } = createMultipartPayload([
{
name: "file",
filename: "test-3page.pdf",
contentType: "application/pdf",
content: PDF,
},
{
name: "settings",
content: JSON.stringify({ quality: "fast", pages: "1" }),
},
]);
const res = await app.inject({
method: "POST",
url: "/api/v1/tools/pdf/ocr-pdf",
headers: {
authorization: `Bearer ${adminToken}`,
"content-type": contentType,
},
body,
});
expect(res.statusCode).toBe(202);
const json = JSON.parse(res.body);
expect(json.jobId).toBeDefined();
expect(json.async).toBe(true);
}, 300_000);
expect(json.error).toBe("Invalid settings");
expect(mocks.getOcrRuntimeCapability).not.toHaveBeenCalled();
});
});