import { describe, it, expect, vi, afterEach } from "vitest";
import { extractText, terminateOcr, searchText } from "../../src/services/ocr.js";
// Mock tesseract.js with v7 API structure (blocks -> paragraphs -> lines -> words)
vi.mock("tesseract.js", () => ({
createWorker: vi.fn().mockResolvedValue({
recognize: vi.fn().mockResolvedValue({
data: {
blocks: [
{
paragraphs: [
{
lines: [
{
words: [
{
text: "Hello",
confidence: 95,
bbox: { x0: 10, y0: 20, x1: 100, y1: 50 },
},
{
text: "World",
confidence: 87,
bbox: { x0: 110, y0: 20, x1: 200, y1: 50 },
},
],
},
],
},
],
},
],
},
}),
terminate: vi.fn(),
}),
}));
describe("OCR Service", () => {
afterEach(async () => {
await terminateOcr();
vi.clearAllMocks();
});
describe("extractText", () => {
it("extracts words with bounds from image", async () => {
const results = await extractText("/fake/path.png");
expect(results).toHaveLength(2);
expect(results[0]).toEqual({
text: "Hello",
confidence: 0.95,
bounds: { x0: 10, y0: 20, x1: 100, y1: 50 },
});
expect(results[1]).toEqual({
text: "World",
confidence: 0.87,
bounds: { x0: 110, y0: 20, x1: 200, y1: 50 },
});
});
it("filters out empty text results", async () => {
const tesseract = await import("tesseract.js");
vi.mocked(tesseract.createWorker).mockResolvedValueOnce({
recognize: vi.fn().mockResolvedValue({
data: {
blocks: [
{
paragraphs: [
{
lines: [
{
words: [
{ text: "Valid", confidence: 90, bbox: { x0: 0, y0: 0, x1: 10, y1: 10 } },
{ text: " ", confidence: 80, bbox: { x0: 20, y0: 0, x1: 30, y1: 10 } },
{ text: "", confidence: 70, bbox: { x0: 40, y0: 0, x1: 50, y1: 10 } },
],
},
],
},
],
},
],
},
}),
terminate: vi.fn(),
} as any);
// Force new worker creation
await terminateOcr();
const results = await extractText("/fake/path.png");
expect(results).toHaveLength(1);
expect(results[0].text).toBe("Valid");
});
it("normalizes confidence to 0-1 range", async () => {
const results = await extractText("/fake/path.png");
expect(results[0].confidence).toBe(0.95);
expect(results[0].confidence).toBeLessThanOrEqual(1);
expect(results[0].confidence).toBeGreaterThanOrEqual(0);
});
it("reuses worker across multiple calls", async () => {
const tesseract = await import("tesseract.js");
await extractText("/fake/path1.png");
await extractText("/fake/path2.png");
// Worker should only be created once
expect(tesseract.createWorker).toHaveBeenCalledTimes(1);
});
it("handles empty blocks gracefully", async () => {
const tesseract = await import("tesseract.js");
vi.mocked(tesseract.createWorker).mockResolvedValueOnce({
recognize: vi.fn().mockResolvedValue({
data: {
blocks: null,
},
}),
terminate: vi.fn(),
} as any);
await terminateOcr();
const results = await extractText("/fake/path.png");
expect(results).toHaveLength(0);
});
});
describe("terminateOcr", () => {
it("terminates worker when called", async () => {
const tesseract = await import("tesseract.js");
const mockWorker = await tesseract.createWorker("eng");
await extractText("/fake/path.png");
await terminateOcr();
expect(mockWorker.terminate).toHaveBeenCalled();
});
});
describe("searchText", () => {
it("finds text containing search term (case-insensitive)", () => {
const ocrResults = [
{ text: "Chobani", confidence: 0.95, bounds: { x0: 10, y0: 20, x1: 100, y1: 50 } },
{ text: "High", confidence: 0.90, bounds: { x0: 110, y0: 20, x1: 160, y1: 50 } },
{ text: "Protein", confidence: 0.88, bounds: { x0: 170, y0: 20, x1: 250, y1: 50 } },
];
const results = searchText(ocrResults, "chobani");
expect(results).toHaveLength(1);
expect(results[0].text).toBe("Chobani");
});
it("returns elements with center coordinates", () => {
const ocrResults = [
{ text: "Button", confidence: 0.95, bounds: { x0: 100, y0: 200, x1: 200, y1: 250 } },
];
const results = searchText(ocrResults, "button");
expect(results[0].center).toEqual({ x: 150, y: 225 });
});
it("returns elements with formatted bounds string", () => {
const ocrResults = [
{ text: "Test", confidence: 0.90, bounds: { x0: 10, y0: 20, x1: 100, y1: 50 } },
];
const results = searchText(ocrResults, "test");
expect(results[0].bounds).toBe("[10,20][100,50]");
});
it("finds partial matches (contains)", () => {
const ocrResults = [
{ text: "Chobani High Protein Drinks", confidence: 0.92, bounds: { x0: 10, y0: 20, x1: 300, y1: 50 } },
];
const results = searchText(ocrResults, "protein");
expect(results).toHaveLength(1);
expect(results[0].text).toBe("Chobani High Protein Drinks");
});
it("returns empty array when no matches", () => {
const ocrResults = [
{ text: "Hello", confidence: 0.95, bounds: { x0: 10, y0: 20, x1: 100, y1: 50 } },
];
const results = searchText(ocrResults, "goodbye");
expect(results).toHaveLength(0);
});
it("includes confidence in results", () => {
const ocrResults = [
{ text: "Target", confidence: 0.88, bounds: { x0: 10, y0: 20, x1: 100, y1: 50 } },
];
const results = searchText(ocrResults, "target");
expect(results[0].confidence).toBe(0.88);
});
it("assigns sequential indices to results", () => {
const ocrResults = [
{ text: "First Match", confidence: 0.95, bounds: { x0: 10, y0: 20, x1: 100, y1: 50 } },
{ text: "Second Match", confidence: 0.90, bounds: { x0: 10, y0: 60, x1: 100, y1: 90 } },
];
const results = searchText(ocrResults, "match");
expect(results[0].index).toBe(0);
expect(results[1].index).toBe(1);
});
});
});
