Skip to content

赵月辉 / fastgpt-migrated

Go to a project
Commit af081457 by Octopus Committed by GitHub
Options

fix(embedding): decode base64 embedding responses before vector processing (#6807) 

* fix(embedding): decode base64 embedding responses before vector processing

When a model's extra body config includes `encoding_format: "base64"`, the
embedding API returns a base64-encoded IEEE 754 little-endian float32 array
instead of a `number[]`. The previous code passed this raw string directly
to `formatVectors`, which called `.reduce()` on it and threw:
  TypeError: a.reduce is not a function

Add `decodeEmbedding()` that detects base64 strings and decodes them to
`number[]` via `Buffer → Float32Array → Array.from()`, then use it in
`getVectorsByText` before calling `formatVectors`.

Fixes #6769

* perf: test

---------

Co-authored-by: octo-patch <octo-patch@github.com>
Co-authored-by: archer <545436317@qq.com>
parent 39d74852
Showing with 13 additions and 210 deletions
packages/service/core/ai/embedding/index.ts
......@@ -91,7 +91,9 @@ export async function getVectorsByText({ model, input, type, headers }: GetVecto
return tokens.reduce((sum, item) => sum + item, 0);
})(),
Promise.all(
res.data.map((item) => formatVectors(item.embedding, model.normalization))
res.data.map((item) =>
formatVectors(decodeEmbedding(item.embedding), model.normalization)
)
)
]);
......@@ -121,6 +123,16 @@ export async function getVectorsByText({ model, input, type, headers }: GetVecto
}
}
export function decodeEmbedding(embedding: number[] | string): number[] {
if (typeof embedding === 'string') {
// base64-encoded IEEE 754 little-endian float32 array
const buf = Buffer.from(embedding, 'base64');
const floats = new Float32Array(buf.buffer, buf.byteOffset, buf.byteLength / 4);
return Array.from(floats);
}
return embedding;
}
export function formatVectors(vector: number[], normalization = false) {
// normalization processing
function normalizationVector(vector: number[]) {
......
test/cases/service/core/ai/embedding.test.ts deleted 100644 → 0
import { formatVectors } from '@fastgpt/service/core/ai/embedding/index';
import { describe, expect, it, vi } from 'vitest';
describe('formatVectors function test', () => {
// Helper function to create a normalized vector (L2 norm = 1)
const createNormalizedVector = (length: number): number[] => {
const vector = Array.from({ length }, (_, i) => (i + 1) / length);
const norm = Math.sqrt(vector.reduce((sum, val) => sum + val * val, 0));
return vector.map((val) => val / norm);
};
// Helper function to create an unnormalized vector
const createUnnormalizedVector = (length: number): number[] => {
return Array.from({ length }, (_, i) => (i + 1) * 10);
};
// Helper function to calculate L2 norm
const calculateNorm = (vector: number[]): number => {
return Math.sqrt(vector.reduce((sum, val) => sum + val * val, 0));
};
// Helper function to check if vector is normalized (L2 norm H 1)
const isNormalized = (vector: number[]): boolean => {
const norm = calculateNorm(vector);
return Math.abs(norm - 1) < 1e-10;
};
describe('1536 dimension vectors', () => {
it('should handle normalized 1536-dim vector with normalization=true', () => {
const inputVector = createNormalizedVector(1536);
const result = formatVectors(inputVector, true);
expect(result).toHaveLength(1536);
expect(isNormalized(result)).toBe(true);
// Since input is already normalized, result should be very similar
expect(result).toEqual(
expect.arrayContaining(inputVector.map((val) => expect.closeTo(val, 10)))
);
});
it('should handle normalized 1536-dim vector with normalization=false', () => {
const inputVector = createNormalizedVector(1536);
const result = formatVectors(inputVector, false);
expect(result).toHaveLength(1536);
expect(result).toEqual(inputVector);
expect(isNormalized(result)).toBe(true);
});
it('should handle unnormalized 1536-dim vector with normalization=true', () => {
const inputVector = createUnnormalizedVector(1536);
const result = formatVectors(inputVector, true);
expect(result).toHaveLength(1536);
expect(isNormalized(result)).toBe(true);
// Result should be different from input (normalized)
expect(result).not.toEqual(inputVector);
});
it('should handle unnormalized 1536-dim vector with normalization=false', () => {
const inputVector = createUnnormalizedVector(1536);
const result = formatVectors(inputVector, false);
expect(result).toHaveLength(1536);
expect(result).toEqual(inputVector);
expect(isNormalized(result)).toBe(false);
});
});
describe('Greater than 1536 dimension vectors', () => {
it('should handle normalized >1536-dim vector with normalization=true', () => {
const inputVector = createNormalizedVector(2048);
const result = formatVectors(inputVector, true);
expect(result).toHaveLength(1536);
expect(isNormalized(result)).toBe(true);
// Should be truncated to first 1536 elements and then normalized
expect(result).toEqual(
expect.arrayContaining(inputVector.slice(0, 1536).map((val) => expect.any(Number)))
);
});
it('should handle normalized >1536-dim vector with normalization=false', () => {
const inputVector = createNormalizedVector(2048);
const result = formatVectors(inputVector, true); // Always normalized for >1536 dims
expect(result).toHaveLength(1536);
expect(isNormalized(result)).toBe(true);
// Should be truncated and normalized regardless of normalization flag
});
it('should handle unnormalized >1536-dim vector with normalization=true', () => {
const inputVector = createUnnormalizedVector(2048);
const result = formatVectors(inputVector, true);
expect(result).toHaveLength(1536);
expect(isNormalized(result)).toBe(true);
// Should be truncated to first 1536 elements and then normalized
});
it('should handle unnormalized >1536-dim vector with normalization=false', () => {
const inputVector = createUnnormalizedVector(2048);
const result = formatVectors(inputVector, false); // Always normalized for >1536 dims
expect(result).toHaveLength(1536);
expect(isNormalized(result)).toBe(true);
// Should be truncated and normalized regardless of normalization flag
});
});
describe('Less than 1536 dimension vectors', () => {
it('should handle normalized <1536-dim vector with normalization=true', () => {
const inputVector = createNormalizedVector(512);
const result = formatVectors(inputVector, true);
expect(result).toHaveLength(1536);
expect(isNormalized(result)).toBe(true);
// First 512 elements should match input, rest should be 0
expect(result.slice(0, 512)).toEqual(
expect.arrayContaining(inputVector.map((val) => expect.any(Number)))
);
expect(result.slice(512)).toEqual(new Array(1024).fill(0));
});
it('should handle normalized <1536-dim vector with normalization=false', () => {
const inputVector = createNormalizedVector(512);
const result = formatVectors(inputVector, false);
expect(result).toHaveLength(1536);
// First 512 elements should match input exactly, rest should be 0
expect(result.slice(0, 512)).toEqual(inputVector);
expect(result.slice(512)).toEqual(new Array(1024).fill(0));
// The result remains normalized because adding zeros doesn't change the L2 norm
expect(isNormalized(result)).toBe(true);
});
it('should handle unnormalized <1536-dim vector with normalization=true', () => {
const inputVector = createUnnormalizedVector(512);
const result = formatVectors(inputVector, true);
expect(result).toHaveLength(1536);
expect(isNormalized(result)).toBe(true);
// Should be padded with zeros and then normalized
expect(result.slice(512)).toEqual(new Array(1024).fill(0));
});
it('should handle unnormalized <1536-dim vector with normalization=false', () => {
const inputVector = createUnnormalizedVector(512);
const result = formatVectors(inputVector, false);
expect(result).toHaveLength(1536);
// First 512 elements should match input exactly, rest should be 0
expect(result.slice(0, 512)).toEqual(inputVector);
expect(result.slice(512)).toEqual(new Array(1024).fill(0));
expect(isNormalized(result)).toBe(false);
});
it('should demonstrate that padding preserves normalization status', () => {
// Create a vector that becomes unnormalized after some scaling
const baseVector = [3, 4]; // norm = 5, not normalized
const result = formatVectors(baseVector, false);
expect(result).toHaveLength(1536);
expect(result[0]).toBe(3);
expect(result[1]).toBe(4);
expect(result.slice(2)).toEqual(new Array(1534).fill(0));
expect(isNormalized(result)).toBe(false);
expect(calculateNorm(result)).toBeCloseTo(5, 10);
});
});
describe('Edge cases', () => {
it('should handle zero vector', () => {
const inputVector = new Array(1536).fill(0);
const result = formatVectors(inputVector, true);
expect(result).toHaveLength(1536);
expect(result).toEqual(inputVector); // Zero vector remains zero after normalization
});
it('should handle single element vector', () => {
const inputVector = [5.0];
const result = formatVectors(inputVector, true);
expect(result).toHaveLength(1536);
expect(result[0]).toBeCloseTo(1.0, 10); // Normalized single element should be 1
expect(result.slice(1)).toEqual(new Array(1535).fill(0));
});
it('should handle exactly 1536 dimension vector', () => {
const inputVector = createNormalizedVector(1536);
const result = formatVectors(inputVector, true);
expect(result).toHaveLength(1536);
expect(isNormalized(result)).toBe(true);
});
it('should handle vector with negative values', () => {
const inputVector = [-1, -2, -3];
const result = formatVectors(inputVector, true);
expect(result).toHaveLength(1536);
expect(isNormalized(result)).toBe(true);
expect(result[0]).toBeLessThan(0); // Should preserve negative values
expect(result[1]).toBeLessThan(0);
expect(result[2]).toBeLessThan(0);
});
});
});
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or sign in to comment