|
| 1 | +// npx vitest services/code-index/embedders/__tests__/fastembed.spec.ts |
| 2 | + |
| 3 | +import { describe, it, expect, beforeEach, vi } from "vitest" |
| 4 | +import { FastEmbedEmbedder } from "../fastembed" |
| 5 | + |
| 6 | +// Mock TelemetryService |
| 7 | +vi.mock("@roo-code/telemetry", () => ({ |
| 8 | + TelemetryService: { |
| 9 | + instance: { |
| 10 | + captureEvent: vi.fn(), |
| 11 | + }, |
| 12 | + }, |
| 13 | +})) |
| 14 | + |
| 15 | +// Mock i18n |
| 16 | +vi.mock("../../../i18n", () => ({ |
| 17 | + t: vi.fn((key: string, params?: any) => { |
| 18 | + if (key === "embeddings:fastembed.modelNotSupported") { |
| 19 | + return `Model "${params?.model}" not supported. Available models: ${params?.availableModels}` |
| 20 | + } |
| 21 | + if (key === "embeddings:fastembed.embeddingFailed") { |
| 22 | + return `Failed to create embeddings with FastEmbed: ${params?.message}` |
| 23 | + } |
| 24 | + if (key === "embeddings:fastembed.noValidTexts") { |
| 25 | + return "No valid texts to embed" |
| 26 | + } |
| 27 | + if (key === "embeddings:fastembed.invalidResponseFormat") { |
| 28 | + return "Invalid response format from FastEmbed" |
| 29 | + } |
| 30 | + if (key === "embeddings:fastembed.invalidEmbeddingFormat") { |
| 31 | + return "Invalid embedding format from FastEmbed" |
| 32 | + } |
| 33 | + return key |
| 34 | + }), |
| 35 | +})) |
| 36 | + |
| 37 | +// Mock getModelQueryPrefix |
| 38 | +vi.mock("../../../shared/embeddingModels", () => ({ |
| 39 | + getModelQueryPrefix: vi.fn(() => null), |
| 40 | +})) |
| 41 | + |
| 42 | +// Mock @mastra/fastembed |
| 43 | +vi.mock("@mastra/fastembed", () => ({ |
| 44 | + fastembed: { |
| 45 | + small: { |
| 46 | + doEmbed: vi.fn(), |
| 47 | + maxEmbeddingsPerCall: 256, |
| 48 | + }, |
| 49 | + base: { |
| 50 | + doEmbed: vi.fn(), |
| 51 | + maxEmbeddingsPerCall: 256, |
| 52 | + }, |
| 53 | + }, |
| 54 | +})) |
| 55 | + |
| 56 | +describe("FastEmbedEmbedder", () => { |
| 57 | + let embedder: FastEmbedEmbedder |
| 58 | + let mockSmallDoEmbed: any |
| 59 | + let mockBaseDoEmbed: any |
| 60 | + |
| 61 | + beforeEach(() => { |
| 62 | + vi.clearAllMocks() |
| 63 | + |
| 64 | + // Get references to the mocked functions |
| 65 | + const { fastembed } = require("@mastra/fastembed") |
| 66 | + mockSmallDoEmbed = fastembed.small.doEmbed |
| 67 | + mockBaseDoEmbed = fastembed.base.doEmbed |
| 68 | + }) |
| 69 | + |
| 70 | + describe("constructor", () => { |
| 71 | + it("should initialize with default model (bge-small-en-v1.5)", () => { |
| 72 | + embedder = new FastEmbedEmbedder({}) |
| 73 | + expect(embedder.embedderInfo.name).toBe("fastembed") |
| 74 | + }) |
| 75 | + |
| 76 | + it("should initialize with specified model", () => { |
| 77 | + embedder = new FastEmbedEmbedder({ fastEmbedModel: "bge-base-en-v1.5" }) |
| 78 | + expect(embedder.embedderInfo.name).toBe("fastembed") |
| 79 | + }) |
| 80 | + |
| 81 | + it("should use fallback model for unsupported model", () => { |
| 82 | + const consoleSpy = vi.spyOn(console, "warn").mockImplementation(() => {}) |
| 83 | + |
| 84 | + embedder = new FastEmbedEmbedder({ fastEmbedModel: "unsupported-model" }) |
| 85 | + |
| 86 | + expect(consoleSpy).toHaveBeenCalledWith(expect.stringContaining('Model "unsupported-model" not available')) |
| 87 | + |
| 88 | + consoleSpy.mockRestore() |
| 89 | + }) |
| 90 | + }) |
| 91 | + |
| 92 | + describe("createEmbeddings", () => { |
| 93 | + beforeEach(() => { |
| 94 | + embedder = new FastEmbedEmbedder({}) |
| 95 | + }) |
| 96 | + |
| 97 | + it("should create embeddings for single text using small model", async () => { |
| 98 | + const mockEmbeddings = [[0.1, 0.2, 0.3, 0.4]] |
| 99 | + mockSmallDoEmbed.mockResolvedValue(mockEmbeddings) |
| 100 | + |
| 101 | + const result = await embedder.createEmbeddings(["test text"]) |
| 102 | + |
| 103 | + expect(mockSmallDoEmbed).toHaveBeenCalledWith({ values: ["test text"] }) |
| 104 | + expect(result).toEqual({ |
| 105 | + embeddings: mockEmbeddings, |
| 106 | + }) |
| 107 | + }) |
| 108 | + |
| 109 | + it("should create embeddings for multiple texts using small model", async () => { |
| 110 | + const mockEmbeddings = [ |
| 111 | + [0.1, 0.2, 0.3, 0.4], |
| 112 | + [0.5, 0.6, 0.7, 0.8], |
| 113 | + ] |
| 114 | + mockSmallDoEmbed.mockResolvedValue(mockEmbeddings) |
| 115 | + |
| 116 | + const result = await embedder.createEmbeddings(["text 1", "text 2"]) |
| 117 | + |
| 118 | + expect(mockSmallDoEmbed).toHaveBeenCalledWith({ values: ["text 1", "text 2"] }) |
| 119 | + expect(result).toEqual({ |
| 120 | + embeddings: mockEmbeddings, |
| 121 | + }) |
| 122 | + }) |
| 123 | + |
| 124 | + it("should create embeddings using base model when specified", async () => { |
| 125 | + embedder = new FastEmbedEmbedder({ fastEmbedModel: "bge-base-en-v1.5" }) |
| 126 | + const mockEmbeddings = [[0.1, 0.2, 0.3, 0.4]] |
| 127 | + mockBaseDoEmbed.mockResolvedValue(mockEmbeddings) |
| 128 | + |
| 129 | + const result = await embedder.createEmbeddings(["test text"]) |
| 130 | + |
| 131 | + expect(mockBaseDoEmbed).toHaveBeenCalledWith({ values: ["test text"] }) |
| 132 | + expect(result).toEqual({ |
| 133 | + embeddings: mockEmbeddings, |
| 134 | + }) |
| 135 | + }) |
| 136 | + |
| 137 | + it("should handle empty input", async () => { |
| 138 | + const result = await embedder.createEmbeddings([]) |
| 139 | + |
| 140 | + expect(mockSmallDoEmbed).not.toHaveBeenCalled() |
| 141 | + expect(result).toEqual({ |
| 142 | + embeddings: [], |
| 143 | + }) |
| 144 | + }) |
| 145 | + |
| 146 | + it("should handle FastEmbed API errors", async () => { |
| 147 | + const error = new Error("FastEmbed API error") |
| 148 | + mockSmallDoEmbed.mockRejectedValue(error) |
| 149 | + |
| 150 | + await expect(embedder.createEmbeddings(["test text"])).rejects.toThrow( |
| 151 | + "Failed to create embeddings with FastEmbed: FastEmbed API error", |
| 152 | + ) |
| 153 | + }) |
| 154 | + |
| 155 | + it("should process large batches correctly", async () => { |
| 156 | + const texts = Array.from({ length: 150 }, (_, i) => `text ${i}`) |
| 157 | + const mockEmbeddings = texts.map((_, i) => [i * 0.1, i * 0.2, i * 0.3, i * 0.4]) |
| 158 | + mockSmallDoEmbed.mockResolvedValue(mockEmbeddings) |
| 159 | + |
| 160 | + const result = await embedder.createEmbeddings(texts) |
| 161 | + |
| 162 | + expect(mockSmallDoEmbed).toHaveBeenCalledWith({ values: texts }) |
| 163 | + expect(result.embeddings).toHaveLength(150) |
| 164 | + }) |
| 165 | + }) |
| 166 | + |
| 167 | + describe("validateConfiguration", () => { |
| 168 | + beforeEach(() => { |
| 169 | + embedder = new FastEmbedEmbedder({}) |
| 170 | + }) |
| 171 | + |
| 172 | + it("should validate successfully with small model", async () => { |
| 173 | + const mockEmbeddings = [[0.1, 0.2, 0.3, 0.4]] |
| 174 | + mockSmallDoEmbed.mockResolvedValue(mockEmbeddings) |
| 175 | + |
| 176 | + const result = await embedder.validateConfiguration() |
| 177 | + |
| 178 | + expect(mockSmallDoEmbed).toHaveBeenCalledWith({ values: ["test"] }) |
| 179 | + expect(result).toEqual({ valid: true }) |
| 180 | + }) |
| 181 | + |
| 182 | + it("should validate successfully with base model", async () => { |
| 183 | + embedder = new FastEmbedEmbedder({ fastEmbedModel: "bge-base-en-v1.5" }) |
| 184 | + const mockEmbeddings = [[0.1, 0.2, 0.3, 0.4]] |
| 185 | + mockBaseDoEmbed.mockResolvedValue(mockEmbeddings) |
| 186 | + |
| 187 | + const result = await embedder.validateConfiguration() |
| 188 | + |
| 189 | + expect(mockBaseDoEmbed).toHaveBeenCalledWith({ values: ["test"] }) |
| 190 | + expect(result).toEqual({ valid: true }) |
| 191 | + }) |
| 192 | + |
| 193 | + it("should return invalid when FastEmbed fails", async () => { |
| 194 | + const error = new Error("FastEmbed validation error") |
| 195 | + mockSmallDoEmbed.mockRejectedValue(error) |
| 196 | + |
| 197 | + const result = await embedder.validateConfiguration() |
| 198 | + |
| 199 | + expect(result).toEqual({ |
| 200 | + valid: false, |
| 201 | + error: "FastEmbed validation failed: FastEmbed validation error", |
| 202 | + }) |
| 203 | + }) |
| 204 | + |
| 205 | + it("should handle unexpected validation errors", async () => { |
| 206 | + mockSmallDoEmbed.mockRejectedValue("Unexpected error") |
| 207 | + |
| 208 | + const result = await embedder.validateConfiguration() |
| 209 | + |
| 210 | + expect(result).toEqual({ |
| 211 | + valid: false, |
| 212 | + error: "FastEmbed validation failed: Unexpected error", |
| 213 | + }) |
| 214 | + }) |
| 215 | + }) |
| 216 | + |
| 217 | + describe("embedderInfo", () => { |
| 218 | + it("should return correct embedder info", () => { |
| 219 | + embedder = new FastEmbedEmbedder({}) |
| 220 | + expect(embedder.embedderInfo).toEqual({ |
| 221 | + name: "fastembed", |
| 222 | + }) |
| 223 | + }) |
| 224 | + }) |
| 225 | + |
| 226 | + describe("model selection", () => { |
| 227 | + it("should use small model by default", () => { |
| 228 | + embedder = new FastEmbedEmbedder({}) |
| 229 | + // We can't directly test the private property, but we can test the behavior |
| 230 | + expect(() => embedder).not.toThrow() |
| 231 | + }) |
| 232 | + |
| 233 | + it("should use base model when specified", () => { |
| 234 | + embedder = new FastEmbedEmbedder({ fastEmbedModel: "bge-base-en-v1.5" }) |
| 235 | + expect(() => embedder).not.toThrow() |
| 236 | + }) |
| 237 | + |
| 238 | + it("should use small model when explicitly specified", () => { |
| 239 | + embedder = new FastEmbedEmbedder({ fastEmbedModel: "bge-small-en-v1.5" }) |
| 240 | + expect(() => embedder).not.toThrow() |
| 241 | + }) |
| 242 | + }) |
| 243 | +}) |
0 commit comments