import { Configuration, OpenAIApi } from 'openai'
import * as cosineSimilarity from 'cosine-similarity'

// 配置OpenAI API
const configuration = new Configuration({
  apiKey: process.env.OPENAI_API_KEY,
})

const openai = new OpenAIApi(configuration)

// 使用 OpenAI 的 text-embedding-3-large 模型来将文本转为向量
async function getEmbedding(text: string): Promise<number[]> {
  const response = await openai.createEmbedding({
    model: 'text-embedding-3-large',
    input: text,
  })

  const embedding = response.data.data[0].embedding
  return embedding
}

// 计算两个向量的余弦相似度
function calculateCosineSimilarity(vec1: number[], vec2: number[]): number {
  return cosineSimilarity(vec1, vec2)
}

// 将多个文本转换为向量并存储
async function embedTexts(texts: string[]): Promise<number[][]> {
  const embeddings = []
  for (const text of texts) {
    const embedding = await getEmbedding(text)
    embeddings.push(embedding)
  }
  return embeddings
}

// 获取与问题最相似的前10条文本资料
async function findTopMatches(
  question: string,
  texts: string[],
  embeddings: number[][]
): Promise<string[]> {
  const questionEmbedding = await getEmbedding(question)

  const similarities = embeddings.map((embedding, index) => ({
    text: texts[index],
    similarity: calculateCosineSimilarity(questionEmbedding, embedding),
  }))

  // 按相似度排序并返回前10个
  similarities.sort((a, b) => b.similarity - a.similarity)

  return similarities.slice(0, 10).map((item) => item.text)
}

// 主逻辑：将用户问题与文本资料匹配
async function handleUserQuestion(
  question: string,
  texts: string[]
): Promise<string[]> {
  const embeddings = await embedTexts(texts)
  return await findTopMatches(question, texts, embeddings)
}

// 示例使用
;(async () => {
  const texts = [
    '这是关于人工智能的资料。',
    '这是关于机器学习的资料。',
    '这是关于自然语言处理的资料。',
  ]

  const question = '什么是人工智能？'
  const matches = await handleUserQuestion(question, texts)

  console.log('匹配到的文本资料：', matches)
})()