baidu-ai-search

{ "cells": [ { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "# 文本生成\n", "\n", "## 说明\n", "本文展示了如何使用appbuilder内置组件以及如何快速构建全新的文本生成组件。以下是一个在营销场景下生成各种文本内容的样例。\n", "\n", "## 概览\n", "具体地，本样例按照以下流程进行文本生成：\n", "1. 使用**空模板**构建一个全新的文本生成组件：商品信息生成组件。\n", "2. 输入一个商品，使用新构建的商品信息生成组件生成该商品的商品信息。\n", "3. 基于生成的商品信息，使用内置的**问答对挖掘**组件生成问答对。生成的问答对可用于客服等。\n", "4. 基于生成的商品信息，使用内置的**风格写作**组件为该商品生成一份【小红书】文案。\n", "5. 使用内置的**标签抽取**组件对生成的文案进行标签抽取。抽取出的标签可用于检索文案。\n", "\n", "以下我们会演示如何实现我们自己的定制化能力。" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "## 流程演示\n", "### Step 0: 安装Python SDK\n", "我们的appbuilder支持使用pip安装（要求Python >= 3.8）：" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "vscode": { "languageId": "shellscript" } }, "outputs": [], "source": [ "pip install appbuilder-sdk" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "代码中需要配置用户的`APPBUILDER_TOKEN`。" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "import os\n", "\n", "# 设置环境变量\n", "os.environ['APPBUILDER_TOKEN'] = '...'" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "安装成功后，我们就可以搭建我们自己的能力了。" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "### Step 1: 商品信息生成\n", "基于appbuilder提供的**空模板**，用户可以自定义文本生成组件以解决多样化的需求。这里我们构建一个**商品信息生成**组件以生成商品信息：" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "from appbuilder import Message, Playground\n", "\n", "# 输入到大模型中的prompt的模板\n", "prompt_template = \\\n", "'''输入商品名，我需要你为我生成该商品的商品信息。\n", "\n", "要求：\n", "- 你生成的商品信息应该包含多方面信息。\n", "- 开头和结尾不需要有其他与商品信息无关的内容。\n", "\n", "商品名：{product_name}\n", "商品信息：\n", "'''\n", "# 创建商品信息生成组件\n", "product_information_generation = Playground(prompt_template=prompt_template, model='Qianfan-Agent-Speed-8K')\n", "\n", "# 获取商品信息\n", "# 填充prompt_template参数的参数映射表，需要与prompt_template对应\n", "prompt_template_kwargs = {\n", " 'product_name': '特斯拉Model Y'\n", "}\n", "response = product_information_generation(Message(prompt_template_kwargs), stream=False, temperature=0.5)\n", "product_information = response.content\n", "print(f'商品信息：\\n{product_information}')" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "输出结果：\n", "\n", "```\n", "商品信息：\n", "特斯拉Model Y是一款电动汽车，它拥有许多独特的功能和特性。以下是它的主要特点和优势：\n", "\n", "特斯拉Model Y是一款电动汽车，它采用纯电力驱动，没有排放和噪音污染，具有零启动加速、零延迟、零维护的特性。\n", "特斯拉Model Y拥有先进的自动驾驶技术，可以自动识别交通状况并做出相应的反应，大大提高了驾驶安全性。\n", "特斯拉Model Y的电池寿命长，充电效率高，可以在短时间内将电池充满，同时支持快速充电和慢充两种方式。\n", "特斯拉Model Y的外观设计独特，线条流畅，充满未来感，内部空间宽敞，乘坐舒适。\n", "特斯拉Model Y的价格相对较高，但性价比高，是一款非常值得购买的汽车。\n", "\n", "总之，特斯拉Model Y是一款集未来科技、安全性、舒适性和实用性于一体的电动汽车。\n", "```" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "### Step 2: 问答对生成\n", "基于Step 1生成的商品信息我们生成一些问答对，这里我们使用内置的**问答对生成**组件。" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "from appbuilder import QAPairMining\n", "\n", "# 初始化问答对生成组件\n", "qa_pair_mining = QAPairMining(model='Qianfan-Agent-Speed-8K')\n", "\n", "# 获取问答对\n", "response = qa_pair_mining(Message(product_information), stream=False, temperature=1e-10)\n", "qa_pairs = response.content\n", "print(f'问答对：\\n{qa_pairs}')" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "输出结果：\n", "\n", "```\n", "问答对：\n", "问题：特斯拉Model Y有什么特点？\n", "答案：特斯拉Model Y是一款电动汽车，它采用纯电力驱动，没有排放和噪音污染，具有零启动加速、零延迟、零维护的特性。\n", "\n", "问题：特斯拉Model Y有什么优势？\n", "答案：特斯拉Model Y拥有先进的自动驾驶技术，可以自动识别交通状况并做出相应的反应，大大提高了驾驶安全性。\n", "\n", "问题：特斯拉Model Y的电池寿命如何？\n", "答案：特斯拉Model Y的电池寿命长，充电效率高，可以在短时间内将电池充满，同时支持快速充电和慢充两种方式。\n", "\n", "问题：特斯拉Model Y的外观设计如何？\n", "答案：特斯拉Model Y的外观设计独特，线条流畅，充满未来感，内部空间宽敞，乘坐舒适。\n", "\n", "问题：特斯拉Model Y的价格是多少？\n", "答案：特斯拉Model Y的价格相对较高，但性价比高，是一款非常值得购买的汽车。\n", "```" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "### Step 3: 营销文案生成\n", "基于Step 1生成的商品信息生成【小红书】文案，这里我们使用内置的**风格写作**组件。" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "from appbuilder import StyleWriting\n", "\n", "# 构造query\n", "query = \\\n", "f'''请你基于以下商品信息生成文案：\n", "{product_information}\n", "'''\n", "\n", "# 初始化风格写作组件\n", "style_writing = StyleWriting(model='Qianfan-Agent-Speed-8K')\n", "\n", "# 获取小红书文案\n", "response = style_writing(Message(query), style_query='小红书', length=300)\n", "copywriting = response.content\n", "print(f'文案：\\n{copywriting}')" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "输出结果：\n", "\n", "```\n", "【特斯拉Model Y，一款值得拥有的未来科技之车！🚀】\n", "\n", "🚗Model Y作为一款电动汽车，具有许多独特的功能和特性。它采用纯电力驱动，没有排放和噪音污染，具有零启动加速、零延迟、零维护的特性。在城市道路上行驶，它就像是一阵清风，让你轻松应对各种交通状况。💨\n", "\n", "🚗特斯拉Model Y还拥有先进的自动驾驶技术，可以自动识别交通状况并做出相应的反应，大大提高了驾驶安全性。在行驶过程中，它就像是一位智能助手，为你提供安全、舒适的驾驶体验。🛡️\n", "\n", "🚗Model Y的电池寿命长，充电效率高，可以在短时间内将电池充满，同时支持快速充电和慢充两种方式。无论是长途旅行还是城市代步，它都能满足你的需求。⚡\n", "\n", "🚗Model Y的外观设计独特，线条流畅，充满未来感，内部空间宽敞，乘坐舒适。无论是驾驶还是乘坐，都能感受到它的豪华与舒适。🌟\n", "\n", "💰虽然价格相对较高，但性价比高，是一款非常值得购买的汽车。它不仅是一款车，更是一种生活方式的体现。🚗\n", "\n", "#未来科技 #特斯拉 #安全性 #舒适性 #实用性 #电动汽车\n", "```" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "### Step 4: 标签抽取\n", "基于Step 3生成的文案进行标签抽取，这里我们使用内置的**标签抽取**组件。" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "from appbuilder import TagExtraction\n", "\n", "# 初始化标签抽取组件\n", "tag_extraction = TagExtraction(model='Qianfan-Agent-Speed-8K')\n", "\n", "# 获取标签\n", "response = tag_extraction(Message(copywriting), stream=False, temperature=1e-10)\n", "tags = response.content\n", "print(f'标签：\\n{tags}')" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "输出结果：\n", "\n", "```\n", "标签：\n", "1.特斯拉\n", "2.Model Y\n", "3.未来科技\n", "4.纯电力驱动\n", "5.自动驾驶\n", "6.电池寿命\n", "7.充电效率\n", "8.外观设计\n", "9.内部空间\n", "10.性价比\n", "```" ] } ], "metadata": { "kernelspec": { "display_name": "base", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.8.8" }, "orig_nbformat": 4, "vscode": { "interpreter": { "hash": "bf58028eab835c8b5fcb0ca40b311fc21e6cba5ccf47f1b42d31ca2545cde380" } } }, "nbformat": 4, "nbformat_minor": 2 }