演示

大型語言模型：摘要內容

• 取得API金鑰：https://platform.openai.com/api-keys
• 添加串流服務器 endpoint

  export const POST: RequestHandler = async ({ request }) => {
    const { model, messages } = await request.json()
    const response = await fetch('https://api.openai.com/v1/chat/completions', {
      method: 'POST',
      headers: {
        'Content-Type': 'application/json',
        Authorization: `Bearer ${OPENAI_API_KEY}`,
      },
      body: JSON.stringify({
        model,
        messages,
        stream: true,
        temperature: 0,
      }),
    })
    return new Response(response.body, { headers: { 'Content-Type': 'text/event-stream;charset=utf-8' } })
  }
  export const POST: RequestHandler = async ({ request }) => {
    const { model, messages } = await request.json()
    const response = await fetch('https://api.openai.com/v1/chat/completions', {
      method: 'POST',
      headers: {
        'Content-Type': 'application/json',
        Authorization: `Bearer ${OPENAI_API_KEY}`,
      },
      body: JSON.stringify({
        model,
        messages,
        stream: true,
        temperature: 0,
      }),
    })
    return new Response(response.body, { headers: { 'Content-Type': 'text/event-stream;charset=utf-8' } })
  }

• 在前端顯示摘要

Google 自然語言處理: 語法分析

斷詞問題:

• 我.個人.認為
• 十個.人認為

Google 自然語言處理: 語法分析

斷詞問題:

• 我.個人.認為
• 十個.人認為

Google 自然語言處理: 語法分析

• 取得API金鑰: https://cloud.google.com/natural-language
• 添加服務器 endpoint

  const languageServiceClient = new LanguageServiceClient({ credentials: CREDENTIALS, projectId: CREDENTIALS.project_id })
  
  export const POST: RequestHandler = async ({ request }) => {
    const { content, language } = await request.json()
    const [syntax] = await languageServiceClient.analyzeSyntax({
      document: {
        content,
        language, // 'en' or 'zh'
        type: 'PLAIN_TEXT',
      },
      encodingType: 'UTF8',
    })
    return json(syntax)
  }
  const languageServiceClient = new LanguageServiceClient({ credentials: CREDENTIALS, projectId: CREDENTIALS.project_id })
  
  export const POST: RequestHandler = async ({ request }) => {
    const { content, language } = await request.json()
    const [syntax] = await languageServiceClient.analyzeSyntax({
      document: {
        content,
        language, // 'en' or 'zh'
        type: 'PLAIN_TEXT',
      },
      encodingType: 'UTF8',
    })
    return json(syntax)
  }

• 在前端顯示分析

Google 翻譯

Google 翻譯: 安裝

• 取得API金鑰: https://cloud.google.com/translate
• 添加服務器 endpoint

const translationClient = new TranslationServiceClient({ credentials: CREDENTIALS, projectId: CREDENTIALS.project_id })

export const POST: RequestHandler = async ({ request }) => {
  const { text, sourceLanguageCode, targetLanguageCode } = await request.json()
  const [response] = await translationClient.translateText({
    parent: `projects/${CREDENTIALS.project_id}/locations/global`,
    contents: [text],
    mimeType: 'text/plain',
    sourceLanguageCode,
    targetLanguageCode,
  })
  const translation = response.translations.map(t => t.translatedText).join('\n')
  return json(translation)
}
const translationClient = new TranslationServiceClient({ credentials: CREDENTIALS, projectId: CREDENTIALS.project_id })

export const POST: RequestHandler = async ({ request }) => {
  const { text, sourceLanguageCode, targetLanguageCode } = await request.json()
  const [response] = await translationClient.translateText({
    parent: `projects/${CREDENTIALS.project_id}/locations/global`,
    contents: [text],
    mimeType: 'text/plain',
    sourceLanguageCode,
    targetLanguageCode,
  })
  const translation = response.translations.map(t => t.translatedText).join('\n')
  return json(translation)
}

另外一個功能

• 根據上下文解釋一個單字
• 根據上下文提出問題

• Slides: http://github.com/jacob-8/gdg23-languages
• Code: http://github.com/jacob-8/poly-tutor
• Language tutor: https://tutor.polylingual.dev
• X: @jacobbowdoin
• 了解 JavaScript 高雄社團

Google 語法分析