流式新版

.github/ISSUE_TEMPLATE/document.md

@@ -1,26 +0,0 @@
----
-name: 文档改善
-about: 欢迎分享您的文档改善建议！
-title: "📝 文档改善"
-labels: ["documentation"]
----
-
-# 文档改善建议 📝
-
-欢迎在此分享您对文档的改善建议，我们期待听到您的想法和建议。
-
-## 您的建议是什么？ 🤔
-
-请简要描述您的文档改善建议，包括您的目标和想法。
-
-如果您的建议是解决某个特定问题的，请尽可能提供更多的上下文和细节。
-
-## 您的建议有哪些优势？ 🌟
-
-请简要描述您的建议的优势和特点，比如：
-
-- 是否可以提高文档的可读性和易用性？
-- 是否可以使文档更加详细和准确？
-- 是否可以让文档更好地反映项目的实际情况？
-
-感谢您的分享和支持！🙏

.github/ISSUE_TEMPLATE/enhancement.md

@@ -1,19 +0,0 @@
----
-name: 功能改善
-about: 欢迎分享您的改善建议！
-title: "🚀 功能改善"
-labels: ["enhancement"]
----
-
-# 功能改善建议 🚀
-
-欢迎在此分享您对功能的改善建议，我们期待听到您的想法和建议。
-
-## 您的建议是什么？ 🤔
-
-请简要描述您的功能改善建议，包括您的目标和想法。
-
-如果您的建议是解决某个特定问题的，请尽可能提供更多的上下文和细节。
-
-
-感谢您的分享和支持！🙏

.github/ISSUE_TEMPLATE/error.md

@@ -1,26 +0,0 @@
----
-name: 错误报告
-about: 提出关于此项目的错误报告
-title: "🐞 错误报告"
-labels: ["bug"]
----
-
-# 错误报告 🐞
-
-如果您在使用此项目时遇到了错误，请在此报告，我们会尽快解决此问题。
-
-## 错误描述 🤔
-
-请详细地描述您遇到的问题，包括出现问题的环境和步骤，以及您已经尝试过的解决方法。
-
-另外，如果您在解决问题时已经查看过其他 GitHub Issue，请务必在文本中说明并引用相关信息。
-
-## 附加信息 📝
-
-请提供以下信息以帮助我们更快地解决问题：
-
-- 输出日志，包括错误信息和堆栈跟踪
-- 相关的代码片段或文件
-- 您的操作系统、软件版本等环境信息
-
-感谢您的反馈！🙏

.github/ISSUE_TEMPLATE/maintenance.md

@@ -1,29 +0,0 @@
----
-name: 项目维护
-about: 欢迎提交您的项目维护问题和建议！
-title: "🔧 项目维护"
-labels: ["maintenance"]
----
-
-# 项目维护问题和建议 🔧
-
-欢迎在此分享您对项目维护的问题和建议，我们期待听到您的想法和建议。
-
-## 您的问题或建议是什么？ 🤔
-
-请简要描述您遇到的项目维护问题或者您的项目维护建议，包括您的目标和想法。
-
-注意：如果您的建议涉及到以下方面，请在描述中加以说明，以帮助我们更好地理解您的意见。
-
-- 代码重构
-- 设计模式加强
-- 优化算法
-- 依赖升级
-
-## 您期望的解决方案是什么？ 💡
-
-请简要描述您期望的解决方案，包括您的期望和想法。
-
-如果您期望的解决方案是解决某个特定问题的，请尽可能提供更多的上下文和细节。
-
-感谢您的分享和支持！🙏

.github/ISSUE_TEMPLATE/question.md

@@ -1,26 +0,0 @@
----
-name: 部署问题反馈
-about: 如果您在部署中遇到任何问题，欢迎在这里与我们交流。
-title: "🚰 部署问题反馈"
-labels: ["question"]
----
-
-# 问题交流 💬
-
-欢迎在此提交您遇到的问题，我们会尽快回复您并提供帮助。
-
-## 问题描述 🤔
-
-请详细描述您遇到的问题，包括出现问题的环境和步骤，以及您已经尝试过的解决方法。
-
-如果您在解决问题时已经查看过其他 GitHub Issue，请务必在文本中说明并引用相关信息。
-
-## 附加信息 📝
-
-为了更好地了解您遇到的问题，我们需要您提供以下信息：
-
-- 输出日志，包括错误信息和堆栈跟踪。
-- 相关的代码片段或文件。
-- 操作系统、golang 版本等环境信息。
-
-感谢您的反馈和支持！🙏

.github/pull-request-template.md

@@ -1,6 +0,0 @@
-<!-- 请务必在创建PR前，在右侧 Labels 选项中加上label的其中一个: [feature]、[fix]、[documentation]、[dependencies]、[test] 。以便于Actions自动生成Releases时自动对PR进行归类。-->
-## 描述
-请简要描述此Pull Request中的更改。
-
-## 相关问题
-- [问题编号]（问题链接）

.github/release-drafter.yml

@@ -1,43 +0,0 @@
-# Configuration for Release Drafter: https://github.com/toolmantim/release-drafter
-name-template: 'v$NEXT_PATCH_VERSION 🌈'
-tag-template: 'v$NEXT_PATCH_VERSION'
-version-template: $MAJOR.$MINOR.$PATCH
-# Emoji reference: https://gitmoji.carloscuesta.me/
-categories:
-  - title: 🚀 Features
-    labels:
-      - 'feature'
-      - 'enhancement'
-      - 'kind/feature'
-  - title: 🚑️ Bug Fixes
-    labels:
-      - 'fix'
-      - 'bugfix'
-      - 'bug'
-      - 'regression'
-      - 'kind/bug'
-  - title: 📝 Documentation updates
-    labels:
-      - 'doc'
-      - 'documentation'
-      - 'kind/doc'
-  - title: 👷 Maintenance
-    labels:
-      - refactor
-      - chore
-      - dependencies
-      - 'kind/chore'
-      - 'kind/dep'
-  - title: 🚦 Tests
-    labels:
-      - test
-      - tests
-exclude-labels:
-  - reverted
-  - no-changelog
-  - skip-changelog
-  - invalid
-change-template: '* $TITLE (#$NUMBER) @$AUTHOR'
-template: |
-  ## What’s Changed
-  $CHANGES

.github/workflows/binary-release.yml

@@ -1,39 +0,0 @@
-name: Build Release
-
-on:
-  release:
-    types: [created,published]
-
-permissions:
-  contents: read
-
-jobs:
-  build-go-binary:
-    permissions:
-      contents: write  # for build-go-binary
-    runs-on: ubuntu-latest
-    strategy:
-      matrix:
-        goos: [ linux, windows, darwin ] # 需要打包的系统
-        goarch: [ amd64, arm64 ] # 需要打包的架构
-        exclude: # 排除某些平台和架构
-          - goarch: arm64
-            goos: windows
-    steps:
-      - name: Checkout the code
-        uses: actions/checkout@v2
-      - name: Create version file
-        run: echo ${{ github.event.release.tag_name }} > VERSION
-      - name: Parallel build
-        uses: wangyoucao577/go-release-action@v1.30
-        with:
-          github_token: ${{ secrets.GITHUB_TOKEN }}
-          goos: ${{ matrix.goos }}
-          goarch: ${{ matrix.goarch }}
-          goversion: 1.18
-          pre_command: export CGO_ENABLED=0 && export GODEBUG=http2client=0
-          executable_compression: "upx -9"
-          md5sum: false
-          project_path: "./code"
-          binary_name: "feishu-chatgpt"
-          extra_files: ./code/config.example.yaml readme.md LICENSE ./code/role_list.yaml

.github/workflows/docker-publish.yml

@@ -1,59 +0,0 @@
-name: build docker image
-
-# release 事件触发 + Master 分支触发 + 手动触发
-on:
-#  push:
-#    branches:
-#      - master
-  release:
-    types: [created,published]
-
-  # 可以手动触发
-  workflow_dispatch:
-    inputs:
-      logLevel:
-        description: 'Log level'
-        required: true
-        default: 'warning'
-      tags:
-        description: 'Test scenario tags'
-
-jobs:
-  buildx:
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v2
-
-      - name: Inject slug/short variables
-        uses: rlespinasse/github-slug-action@v4
-
-      - name: Set up QEMU
-        uses: docker/setup-qemu-action@v1
-
-      - name: Set up Docker Buildx
-        id: buildx
-        uses: docker/setup-buildx-action@v1
-
-      - name: Available platforms
-        run: echo ${{ steps.buildx.outputs.platforms }}
-
-      - name: Login to DockerHub
-        uses: docker/login-action@v1
-        with:
-          username: ${{ secrets.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_TOKEN }}
-
-      - name: Build and push
-        uses: docker/build-push-action@v2
-        with:
-          context: .
-          file: ./Dockerfile
-          # 所需要的体系结构，可以在 Available platforms 步骤中获取所有的可用架构
-          platforms: linux/amd64,linux/arm64/v8
-          # 镜像推送时间
-          push: ${{ github.event_name != 'pull_request' }}
-          # 给清单打上多个标签
-          tags: |
-            ${{ secrets.DOCKERHUB_USERNAME }}/${{ env.GITHUB_REPOSITORY_NAME_PART }}:${{ env.GITHUB_REF_NAME }}
-            ${{ secrets.DOCKERHUB_USERNAME }}/${{ env.GITHUB_REPOSITORY_NAME_PART }}:latest

.github/workflows/docker.yml

@@ -0,0 +1,42 @@
+name: Docker
+
+on:
+  push:
+    branches:
+      - main
+    tags:
+      - v*
+
+env:
+  IMAGE_NAME: oapi-feishu
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: 检出代码库
+        uses: actions/checkout@v3
+
+      - name: 构建镜像
+        run: docker build . --file Dockerfile --tag $IMAGE_NAME
+
+      - name: 登录到镜像仓库
+        run: echo "${{ secrets.ACCESS_TOKEN }}" | docker login -u woodchen --password-stdin
+
+      - name: 推送镜像
+        run: |
+          IMAGE_ID=woodchen/$IMAGE_NAME
+          IMAGE_ID=$(echo $IMAGE_ID | tr '[A-Z]' '[a-z]')
+
+          # 从 GitHub 事件负载中获取分支名
+          BRANCH_NAME=$(echo "${{ github.ref }}" | sed -e 's,.*/\(.*\),\1,')
+
+          # 对于除了 "main" 分支和标签以外的分支，使用 "latest" 版本号
+          VERSION=$(if [ "$BRANCH_NAME" == "main" ]; then echo "latest"; else echo $BRANCH_NAME; fi)
+
+          echo IMAGE_ID=$IMAGE_ID
+          echo VERSION=$VERSION
+
+          docker tag $IMAGE_NAME $IMAGE_ID:$VERSION
+          docker push $IMAGE_ID:$VERSION

.github/workflows/release-draft.yml

@@ -1,38 +0,0 @@
-name: Release Drafter
-
-on:
-  push:
-    # branches to consider in the event; optional, defaults to all
-    branches:
-      - master
-  # pull_request event is required only for autolabeler
-  pull_request:
-    # Only following types are handled by the action, but one can default to all as well
-    types: [opened, reopened, synchronize]
-  # pull_request_target event is required for autolabeler to support PRs from forks
-  # pull_request_target:
-  #   types: [opened, reopened, synchronize]
-
-permissions:
-  contents: read
-
-jobs:
-  update_release_draft:
-    permissions:
-      contents: write  # for release-drafter/release-drafter to create a github release
-      pull-requests: write  # for release-drafter/release-drafter to add label to PR
-    runs-on: ubuntu-latest
-    steps:
-      # (Optional) GitHub Enterprise requires GHE_HOST variable set
-      #- name: Set GHE_HOST
-      #  run: |
-      #    echo "GHE_HOST=${GITHUB_SERVER_URL##https:\/\/}" >> $GITHUB_ENV
-
-      # Drafts your next Release notes as Pull Requests are merged into "master"
-      - uses: release-drafter/release-drafter@v5
-        # (Optional) specify config name to use, relative to .github/. Default: release-drafter.yml
-        # with:
-        #   config-name: my-config.yml
-        #   disable-autolabeler: true
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

.gitignore

@@ -1,40 +0,0 @@
-### Go template
-# If you prefer the allow list template instead of the deny list, see community template:
-# https://github.com/github/gitignore/blob/main/community/Golang/Go.AllowList.gitignore
-#
-# Binaries for programs and plugins
-*.exe
-*.exe~
-*.dll
-*.so
-*.dylib
-
-# Test binary, built with `go test -c`
-*.test
-
-# Output of the go coverage tool, specifically when used with LiteIDE
-*.out
-
-# Dependency directories (remove the comment below to include it)
-# vendor/
-
-# Go workspace file
-go.work
-./code/target
-.idea
-.vscode
-.s
-
-config.yaml
-
-
-
-/code/target/
-start-feishubot
-.env
-
-docker.md
-# Mac OS
-.DS_Store
-**/.DS_Store
-*.pem

code/config.example.yaml

@@ -1,10 +1,12 @@
+# 是否启用日志。
+ENABLE_LOG: true
 #  飞书
 APP_ID: cli_axxx
 APP_SECRET: xxx
 APP_ENCRYPT_KEY: xxx
 APP_VERIFICATION_TOKEN: xxx
-# 请确保和飞书应用管理平台中的设置一致
+# 请确保和飞书应用管理平台中的设置一致。这里建议直接用 Feishu-OpenAI-Stream-Chatbot 作为机器人名称，这样的话，如果你有多个bot就好区分
-BOT_NAME: chatGpt
+BOT_NAME: xxx
 # openAI key 支持负载均衡 可以填写多个key 用逗号分隔
 OPENAI_KEY: sk-xxx,sk-xxx,sk-xxx
 # 服务器配置
@@ -17,11 +19,20 @@ KEY_FILE: key.pem
 API_URL: https://oapi.czl.net
 # 代理设置, 例如 "http://127.0.0.1:7890", ""代表不使用代理
 HTTP_PROXY: ""
+# 访问OpenAi的 普通 Http请求的超时时间，单位秒，不配置的话默认为 550 秒
+OPENAI_HTTP_CLIENT_TIMEOUT:
+# openai 指定模型, 更多见 https://platform.openai.com/docs/models/model-endpoint-compatibility 中 /v1/chat/completions
+OPENAI_MODEL: gpt-3.5-turbo
 
 # AZURE OPENAI
-AZURE_ON: false # set true to use Azure rather than OpenAI
+AZURE_ON: false # set to true to use Azure rather than OpenAI
 AZURE_API_VERSION: 2023-03-15-preview # 2023-03-15-preview or 2022-12-01 refer https://learn.microsoft.com/en-us/azure/cognitive-services/openai/reference#completions
 AZURE_RESOURCE_NAME: xxxx   # you can find in endpoint url. Usually looks like https://{RESOURCE_NAME}.openai.azure.com
 AZURE_DEPLOYMENT_NAME: xxxx # usually looks like ...openai.azure.com/openai/deployments/{DEPLOYMENT_NAME}/chat/completions.
 AZURE_OPENAI_TOKEN: xxxx  # Authentication key. We can use Azure Active Directory Authentication(TBD).
 
+## 访问控制
+# 是否启用访问控制。默认不启用。
+ACCESS_CONTROL_ENABLE: false
+# 每个用户每天最多问多少个问题。默认为不限制. 配置成为小于等于0表示不限制。
+ACCESS_CONTROL_MAX_COUNT_PER_USER_PER_DAY: 0

code/go.mod

@@ -8,10 +8,12 @@ require (
 	github.com/duke-git/lancet/v2 v2.1.17
 	github.com/gin-gonic/gin v1.8.2
 	github.com/google/uuid v1.3.0
+	github.com/k0kubun/pp/v3 v3.2.0
 	github.com/larksuite/oapi-sdk-gin v1.0.0
 	github.com/pandodao/tokenizer-go v0.2.0
 	github.com/patrickmn/go-cache v2.1.0+incompatible
 	github.com/pion/opus v0.0.0-20230123082803-1052c3e89e58
+	github.com/sashabaranov/go-openai v1.7.0
 	github.com/spf13/pflag v1.0.5
 	github.com/spf13/viper v1.14.0
 	gopkg.in/yaml.v2 v2.4.0
@@ -33,6 +35,7 @@ require (
 	github.com/json-iterator/go v1.1.12 // indirect
 	github.com/leodido/go-urn v1.2.1 // indirect
 	github.com/magiconair/properties v1.8.7 // indirect
+	github.com/mattn/go-colorable v0.1.13 // indirect
 	github.com/mattn/go-isatty v0.0.17 // indirect
 	github.com/mitchellh/mapstructure v1.5.0 // indirect
 	github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd // indirect
@@ -51,5 +54,6 @@ require (
 	golang.org/x/text v0.8.0 // indirect
 	google.golang.org/protobuf v1.28.1 // indirect
 	gopkg.in/ini.v1 v1.67.0 // indirect
+	gopkg.in/natefinch/lumberjack.v2 v2.2.1 // indirect
 	gopkg.in/yaml.v3 v3.0.1 // indirect
 )

code/go.sum

@@ -170,6 +170,8 @@ github.com/json-iterator/go v1.1.12 h1:PV8peI4a0ysnczrg+LtxykD8LfKY9ML6u2jnxaEnr
 github.com/json-iterator/go v1.1.12/go.mod h1:e30LSqwooZae/UwlEbR2852Gd8hjQvJoHmT4TnhNGBo=
 github.com/jstemmer/go-junit-report v0.0.0-20190106144839-af01ea7f8024/go.mod h1:6v2b51hI/fHJwM22ozAgKL4VKDeJcHhJFhtBdhmNjmU=
 github.com/jstemmer/go-junit-report v0.9.1/go.mod h1:Brl9GWCQeLvo8nXZwPNNblvFj/XSXhF0NWZEnDohbsk=
+github.com/k0kubun/pp/v3 v3.2.0 h1:h33hNTZ9nVFNP3u2Fsgz8JXiF5JINoZfFq4SvKJwNcs=
+github.com/k0kubun/pp/v3 v3.2.0/go.mod h1:ODtJQbQcIRfAD3N+theGCV1m/CBxweERz2dapdz1EwA=
 github.com/kisielk/gotool v1.0.0/go.mod h1:XhKaO+MFFWcvkIS/tQcRk01m1F5IRFswLeQ+oQHNcck=
 github.com/kr/fs v0.1.0/go.mod h1:FFnZGqtBN9Gxj7eW1uZ42v5BccTP0vu6NEaFoC2HwRg=
 github.com/kr/pretty v0.1.0/go.mod h1:dAy3ld7l9f0ibDNOQOHHMYYIIbhfbHSm3C4ZsoJORNo=
@@ -188,6 +190,9 @@ github.com/leodido/go-urn v1.2.1 h1:BqpAaACuzVSgi/VLzGZIobT2z4v53pjosyNd9Yv6n/w=
 github.com/leodido/go-urn v1.2.1/go.mod h1:zt4jvISO2HfUBqxjfIshjdMTYS56ZS/qv49ictyFfxY=
 github.com/magiconair/properties v1.8.7 h1:IeQXZAiQcpL9mgcAe1Nu6cX9LLw6ExEHKjN0VQdvPDY=
 github.com/magiconair/properties v1.8.7/go.mod h1:Dhd985XPs7jluiymwWYZ0G4Z61jb3vdS329zhj2hYo0=
+github.com/mattn/go-colorable v0.1.13 h1:fFA4WZxdEF4tXPZVKMLwD8oUnCTTo08duU7wxecdEvA=
+github.com/mattn/go-colorable v0.1.13/go.mod h1:7S9/ev0klgBDR4GtXTXX8a3vIGJpMovkB8vQcUbaXHg=
+github.com/mattn/go-isatty v0.0.16/go.mod h1:kYGgaQfpe5nmfYZH+SKPsOc2e4SrIfOl2e/yFXSvRLM=
 github.com/mattn/go-isatty v0.0.17 h1:BTarxUcIeDqL27Mc+vyvdWYSL28zpIhv3RoTdsLMPng=
 github.com/mattn/go-isatty v0.0.17/go.mod h1:kYGgaQfpe5nmfYZH+SKPsOc2e4SrIfOl2e/yFXSvRLM=
 github.com/mitchellh/mapstructure v1.5.0 h1:jeMsZIYE/09sWLaz43PL7Gy6RuMjD2eJVyuac5Z2hdY=
@@ -217,6 +222,10 @@ github.com/rogpeppe/go-internal v1.3.0/go.mod h1:M8bDsm7K2OlrFYOpmOWEs/qY81heoFR
 github.com/rogpeppe/go-internal v1.6.1/go.mod h1:xXDCJY+GAPziupqXw64V24skbSoqbTEfhy4qGm1nDQc=
 github.com/rogpeppe/go-internal v1.8.0 h1:FCbCCtXNOY3UtUuHUYaghJg4y7Fd14rXifAYUAtL9R8=
 github.com/rogpeppe/go-internal v1.8.0/go.mod h1:WmiCO8CzOY8rg0OYDC4/i/2WRWAB6poM+XZ2dLUbcbE=
+github.com/sashabaranov/go-openai v1.7.0 h1:D1dBXoZhtf/aKNu6WFf0c7Ah2NM30PZ/3Mqly6cZ7fk=
+github.com/sashabaranov/go-openai v1.7.0/go.mod h1:lj5b/K+zjTSFxVLijLSTDZuP7adOgerWeFyZLUhAKRg=
+github.com/sashabaranov/go-openai v1.9.0 h1:NoiO++IISxxJ1pRc0n7uZvMGMake0G+FJ1XPwXtprsA=
+github.com/sashabaranov/go-openai v1.9.0/go.mod h1:lj5b/K+zjTSFxVLijLSTDZuP7adOgerWeFyZLUhAKRg=
 github.com/spf13/afero v1.9.3 h1:41FoI0fD7OR7mGcKE/aOiLkGreyf8ifIOQmJANWogMk=
 github.com/spf13/afero v1.9.3/go.mod h1:iUV7ddyEEZPO5gA3zD4fJt6iStLlL+Lg4m2cihcDf8Y=
 github.com/spf13/cast v1.5.0 h1:rj3WzYc11XZaIZMPKmwP96zkFEnnAmV8s6XbB2aY32w=
@@ -570,6 +579,8 @@ gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c/go.mod h1:JHkPIbrfpd72SG/EV
 gopkg.in/errgo.v2 v2.1.0/go.mod h1:hNsd1EY+bozCKY1Ytp96fpM3vjJbqLJn88ws8XvfDNI=
 gopkg.in/ini.v1 v1.67.0 h1:Dgnx+6+nfE+IfzjUEISNeydPJh9AXNNsWbGP9KzCsOA=
 gopkg.in/ini.v1 v1.67.0/go.mod h1:pNLf8WUiyNEtQjuu5G5vTm06TEv9tsIgeAvK8hOrP4k=
+gopkg.in/natefinch/lumberjack.v2 v2.2.1 h1:bBRl1b0OH9s/DuPhuXpNl+VtCaJXFZ5/uEFST95x9zc=
+gopkg.in/natefinch/lumberjack.v2 v2.2.1/go.mod h1:YD8tP3GAjkrDg1eZH7EGmyESg/lsYskCTPBJVb9jqSc=
 gopkg.in/yaml.v2 v2.2.2/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
 gopkg.in/yaml.v2 v2.4.0 h1:D8xgwECY7CYvx+Y2n4sBz93Jn9JRvxdiyyo8CTfuKaY=
 gopkg.in/yaml.v2 v2.4.0/go.mod h1:RDklbk79AGWmwhnvt/jBztapEOGDOx6ZbXqjP6csGnQ=

code/handlers/card_ai_mode_action.go

@@ -1,38 +0,0 @@
-package handlers
-
-import (
-	"context"
-
-	"start-feishubot/services"
-	"start-feishubot/services/openai"
-
-	larkcard "github.com/larksuite/oapi-sdk-go/v3/card"
-)
-
-// AIModeChooseKind is the kind of card action for choosing AI mode
-func NewAIModeCardHandler(cardMsg CardMsg,
-	m MessageHandler) CardHandlerFunc {
-	return func(ctx context.Context, cardAction *larkcard.CardAction) (interface{}, error) {
-
-		if cardMsg.Kind == AIModeChooseKind {
-			newCard, err, done := CommonProcessAIMode(cardMsg, cardAction,
-				m.sessionCache)
-			if done {
-				return newCard, err
-			}
-			return nil, nil
-		}
-		return nil, ErrNextHandler
-	}
-}
-
-// CommonProcessAIMode is the common process for choosing AI mode
-func CommonProcessAIMode(msg CardMsg, cardAction *larkcard.CardAction,
-	cache services.SessionServiceCacheInterface) (interface{},
-	error, bool) {
-	option := cardAction.Action.Option
-	replyMsg(context.Background(), "已选择AI模式:"+option,
-		&msg.MsgId)
-	cache.SetAIMode(msg.SessionId, openai.AIModeMap[option])
-	return nil, nil, true
-}

code/handlers/card_clear_action.go

@@ -2,10 +2,8 @@ package handlers
 
 import (
 	"context"
-
-	"start-feishubot/services"
-
 	larkcard "github.com/larksuite/oapi-sdk-go/v3/card"
+	"start-feishubot/services"
 )
 
 func NewClearCardHandler(cardMsg CardMsg, m MessageHandler) CardHandlerFunc {

code/handlers/card_common_action.go

@@ -4,7 +4,6 @@ import (
 	"context"
 	"encoding/json"
 	"fmt"
-
 	larkcard "github.com/larksuite/oapi-sdk-go/v3/card"
 )
 
@@ -23,16 +22,13 @@ func NewCardHandler(m MessageHandler) CardHandlerFunc {
 		NewPicModeChangeHandler,
 		NewRoleTagCardHandler,
 		NewRoleCardHandler,
-		NewAIModeCardHandler,
 	}
 
 	return func(ctx context.Context, cardAction *larkcard.CardAction) (interface{}, error) {
 		var cardMsg CardMsg
 		actionValue := cardAction.Action.Value
 		actionValueJson, _ := json.Marshal(actionValue)
-		if err := json.Unmarshal(actionValueJson, &cardMsg); err != nil {
+		json.Unmarshal(actionValueJson, &cardMsg)
-			return nil, err
-		}
 		//pp.Println(cardMsg)
 		for _, handler := range handlers {
 			h := handler(cardMsg, m)

code/handlers/card_pic_action.go

@@ -2,10 +2,8 @@ package handlers
 
 import (
 	"context"
-
-	"start-feishubot/services"
-
 	larkcard "github.com/larksuite/oapi-sdk-go/v3/card"
+	"start-feishubot/services"
 )
 
 func NewPicResolutionHandler(cardMsg CardMsg, m MessageHandler) CardHandlerFunc {
@@ -30,7 +28,6 @@ func NewPicModeChangeHandler(cardMsg CardMsg, m MessageHandler) CardHandlerFunc
 		return nil, ErrNextHandler
 	}
 }
-
 func NewPicTextMoreHandler(cardMsg CardMsg, m MessageHandler) CardHandlerFunc {
 	return func(ctx context.Context, cardAction *larkcard.CardAction) (interface{}, error) {
 		if cardMsg.Kind == PicTextMoreKind {

code/handlers/card_role_action.go

@@ -2,12 +2,10 @@ package handlers
 
 import (
 	"context"
-
+	larkcard "github.com/larksuite/oapi-sdk-go/v3/card"
 	"start-feishubot/initialization"
 	"start-feishubot/services"
 	"start-feishubot/services/openai"
-
-	larkcard "github.com/larksuite/oapi-sdk-go/v3/card"
 )
 
 func NewRoleTagCardHandler(cardMsg CardMsg,

code/handlers/common.go

@@ -15,108 +15,9 @@ func msgFilter(msg string) string {
 	return regex.ReplaceAllString(msg, "")
 
 }
-
+func parseContent(content string) string {
-// Parse rich text json to text
-func parsePostContent(content string) string {
-	/*
-		{
-		    "title":"我是一个标题",
-		    "content":[
-		        [
-		            {
-		                "tag":"text",
-		                "text":"第一行 :",
-		                "style": ["bold", "underline"]
-		            },
-		            {
-		                "tag":"a",
-		                "href":"http://www.feishu.cn",
-		                "text":"超链接",
-		                "style": ["bold", "italic"]
-		            },
-		            {
-		                "tag":"at",
-		                "user_id":"@_user_1",
-		                "user_name":"",
-		                "style": []
-		            }
-		        ],
-		        [
-		            {
-		                "tag":"img",
-		                "image_key":"img_47354fbc-a159-40ed-86ab-2ad0f1acb42g"
-		            }
-		        ],
-		        [
-		            {
-		                "tag":"text",
-		                "text":"第二行:",
-		                "style": ["bold", "underline"]
-		            },
-		            {
-		                "tag":"text",
-		                "text":"文本测试",
-		                "style": []
-		            }
-		        ],
-		        [
-		            {
-		                "tag":"img",
-		                "image_key":"img_47354fbc-a159-40ed-86ab-2ad0f1acb42g"
-		            }
-		        ],
-		        [
-		            {
-		                "tag":"media",
-		                "file_key": "file_v2_0dcdd7d9-fib0-4432-a519-41d25aca542j",
-		                "image_key": "img_7ea74629-9191-4176-998c-2e603c9c5e8g"
-		            }
-		        ],
-		        [
-		            {
-		                "tag": "emotion",
-		                "emoji_type": "SMILE"
-		            }
-		        ]
-		    ]
-		}
-	*/
-	var contentMap map[string]interface{}
-	err := json.Unmarshal([]byte(content), &contentMap)
-
-	if err != nil {
-		fmt.Println(err)
-	}
-
-	if contentMap["content"] == nil {
-		return ""
-	}
-	var text string
-	// deal with title
-	if contentMap["title"] != nil && contentMap["title"] != "" {
-		text += contentMap["title"].(string) + "\n"
-	}
-	// deal with content
-	contentList := contentMap["content"].([]interface{})
-	for _, v := range contentList {
-		for _, v1 := range v.([]interface{}) {
-			if v1.(map[string]interface{})["tag"] == "text" {
-				text += v1.(map[string]interface{})["text"].(string)
-			}
-		}
-		// add new line
-		text += "\n"
-	}
-	return msgFilter(text)
-}
-
-func parseContent(content, msgType string) string {
 	//"{\"text\":\"@_user_1  hahaha\"}",
 	//only get text content hahaha
-	if msgType == "post" {
-		return parsePostContent(content)
-	}
-
 	var contentMap map[string]interface{}
 	err := json.Unmarshal([]byte(content), &contentMap)
 	if err != nil {
@@ -128,7 +29,6 @@ func parseContent(content, msgType string) string {
 	text := contentMap["text"].(string)
 	return msgFilter(text)
 }
-
 func processMessage(msg interface{}) (string, error) {
 	msg = strings.TrimSpace(msg.(string))
 	msgB, err := json.Marshal(msg)

code/handlers/event_audio_action.go

@@ -1,69 +0,0 @@
-package handlers
-
-import (
-	"context"
-	"fmt"
-	"os"
-
-	"start-feishubot/initialization"
-	"start-feishubot/utils/audio"
-
-	larkim "github.com/larksuite/oapi-sdk-go/v3/service/im/v1"
-)
-
-type AudioAction struct { /*语音*/
-}
-
-func (*AudioAction) Execute(a *ActionInfo) bool {
-	check := AzureModeCheck(a)
-	if !check {
-		return true
-	}
-
-	// 只有私聊才解析语音,其他不解析
-	if a.info.handlerType != UserHandler {
-		return true
-	}
-
-	//判断是否是语音
-	if a.info.msgType == "audio" {
-		fileKey := a.info.fileKey
-		//fmt.Printf("fileKey: %s \n", fileKey)
-		msgId := a.info.msgId
-		//fmt.Println("msgId: ", *msgId)
-		req := larkim.NewGetMessageResourceReqBuilder().MessageId(
-			*msgId).FileKey(fileKey).Type("file").Build()
-		resp, err := initialization.GetLarkClient().Im.MessageResource.Get(context.Background(), req)
-		//fmt.Println(resp, err)
-		if err != nil {
-			fmt.Println(err)
-			return true
-		}
-		f := fmt.Sprintf("%s.ogg", fileKey)
-		resp.WriteFile(f)
-		defer os.Remove(f)
-
-		//fmt.Println("f: ", f)
-		output := fmt.Sprintf("%s.mp3", fileKey)
-		// 等待转换完成
-		audio.OggToWavByPath(f, output)
-		defer os.Remove(output)
-		//fmt.Println("output: ", output)
-
-		text, err := a.handler.gpt.AudioToText(output)
-		if err != nil {
-			fmt.Println(err)
-
-			sendMsg(*a.ctx, fmt.Sprintf("🤖️：语音转换失败，请稍后再试～\n错误信息: %v", err), a.info.msgId)
-			return false
-		}
-
-		replyMsg(*a.ctx, fmt.Sprintf("🤖️：%s", text), a.info.msgId)
-		//fmt.Println("text: ", text)
-		a.info.qParsed = text
-		return true
-	}
-
-	return true
-
-}

code/handlers/event_common_action.go

@@ -3,12 +3,10 @@ package handlers
 import (
 	"context"
 	"fmt"
-
+	larkim "github.com/larksuite/oapi-sdk-go/v3/service/im/v1"
 	"start-feishubot/initialization"
 	"start-feishubot/services/openai"
 	"start-feishubot/utils"
-
-	larkim "github.com/larksuite/oapi-sdk-go/v3/service/im/v1"
 )
 
 type MsgInfo struct {
@@ -16,6 +14,7 @@ type MsgInfo struct {
 	msgType     string
 	msgId       *string
 	chatId      *string
+	userId      string
 	qParsed     string
 	fileKey     string
 	imageKey    string
@@ -153,15 +152,3 @@ func (*RoleListAction) Execute(a *ActionInfo) bool {
 	}
 	return true
 }
-
-type AIModeAction struct { /*AI模式*/
-}
-
-func (*AIModeAction) Execute(a *ActionInfo) bool {
-	if _, foundMode := utils.EitherCutPrefix(a.info.qParsed,
-		"/ai_mode", "AI模式"); foundMode {
-		SendAIModeListsCard(*a.ctx, a.info.sessionId, a.info.msgId, openai.AIModeStrs)
-		return false
-	}
-	return true
-}

code/handlers/event_msg_action.go

@@ -1,41 +1,157 @@
 package handlers
 
 import (
+	"encoding/json"
 	"fmt"
-
+	"github.com/k0kubun/pp/v3"
+	"log"
+	"start-feishubot/initialization"
+	"start-feishubot/services/accesscontrol"
+	"start-feishubot/services/chatgpt"
 	"start-feishubot/services/openai"
+	"strings"
+	"time"
 )
 
 type MessageAction struct { /*消息*/
+	chatgpt *chatgpt.ChatGPT
 }
 
-func (*MessageAction) Execute(a *ActionInfo) bool {
+func (m *MessageAction) Execute(a *ActionInfo) bool {
+
+	// Add access control
+	if initialization.GetConfig().AccessControlEnable &&
+		!accesscontrol.CheckAllowAccessThenIncrement(&a.info.userId) {
+
+		msg := fmt.Sprintf("UserId: 【%s】 has accessed max count today! Max access count today %s: 【%d】",
+			a.info.userId, accesscontrol.GetCurrentDateFlag(), initialization.GetConfig().AccessControlMaxCountPerUserPerDay)
+
+		_ = sendMsg(*a.ctx, msg, a.info.chatId)
+		return false
+	}
+
+	//s := "快速响应，用于测试： " + time.Now().String() +
+	//	" accesscontrol.currentDate " + accesscontrol.GetCurrentDateFlag()
+	//_ = sendMsg(*a.ctx, s, a.info.chatId)
+	//log.Println(s)
+	//return false
+
+	cardId, err2 := sendOnProcess(a)
+	if err2 != nil {
+		return false
+	}
+
+	answer := ""
+	chatResponseStream := make(chan string)
+	done := make(chan struct{}) // 添加 done 信号，保证 goroutine 正确退出
+	noContentTimeout := time.AfterFunc(10*time.Second, func() {
+		pp.Println("no content timeout")
+		close(done)
+		err := updateFinalCard(*a.ctx, "请求超时", cardId)
+		if err != nil {
+			return
+		}
+		return
+	})
+	defer noContentTimeout.Stop()
 	msg := a.handler.sessionCache.GetMsg(*a.info.sessionId)
 	msg = append(msg, openai.Messages{
 		Role: "user", Content: a.info.qParsed,
 	})
-	// get ai mode as temperature
+	go func() {
-	aiMode := a.handler.sessionCache.GetAIMode(*a.info.sessionId)
+		defer func() {
-	completions, err := a.handler.gpt.Completions(msg, aiMode)
+			if err := recover(); err != nil {
+				err := updateFinalCard(*a.ctx, "聊天失败", cardId)
 				if err != nil {
-		replyMsg(*a.ctx, fmt.Sprintf(
+					printErrorMessage(a, msg, err)
-			"🤖️：消息机器人摆烂了，请稍后再试～\n错误信息: %v", err), a.info.msgId)
+					return
+				}
+			}
+		}()
+
+		//log.Printf("UserId: %s , Request: %s", a.info.userId, msg)
+
+		if err := m.chatgpt.StreamChat(*a.ctx, msg, chatResponseStream); err != nil {
+			err := updateFinalCard(*a.ctx, "聊天失败", cardId)
+			if err != nil {
+				printErrorMessage(a, msg, err)
+				return
+			}
+			close(done) // 关闭 done 信号
+		}
+
+		close(done) // 关闭 done 信号
+	}()
+	ticker := time.NewTicker(700 * time.Millisecond)
+	defer ticker.Stop() // 注意在函数结束时停止 ticker
+	go func() {
+		for {
+			select {
+			case <-done:
+				return
+			case <-ticker.C:
+				err := updateTextCard(*a.ctx, answer, cardId)
+				if err != nil {
+					printErrorMessage(a, msg, err)
+					return
+				}
+			}
+		}
+	}()
+
+	for {
+		select {
+		case res, ok := <-chatResponseStream:
+			if !ok {
 				return false
 			}
-	msg = append(msg, completions)
+			noContentTimeout.Stop()
+			answer += res
+			//pp.Println("answer", answer)
+		case <-done: // 添加 done 信号的处理
+			err := updateFinalCard(*a.ctx, answer, cardId)
+			if err != nil {
+				printErrorMessage(a, msg, err)
+				return false
+			}
+			ticker.Stop()
+			msg := append(msg, openai.Messages{
+				Role: "assistant", Content: answer,
+			})
 			a.handler.sessionCache.SetMsg(*a.info.sessionId, msg)
+			close(chatResponseStream)
 			//if new topic
-	if len(msg) == 2 {
+			//if len(msg) == 2 {
-		//fmt.Println("new topic", msg[1].Content)
+			//	//fmt.Println("new topic", msg[1].Content)
-		sendNewTopicCard(*a.ctx, a.info.sessionId, a.info.msgId,
+			//	//updateNewTextCard(*a.ctx, a.info.sessionId, a.info.msgId,
-			completions.Content)
+			//	//	completions.Content)
-		return false
+			//}
-	}
+			log.Printf("\n\n\n")
-	err = replyMsg(*a.ctx, completions.Content, a.info.msgId)
+			log.Printf("Success request: UserId: %s , Request: %s , Response: %s", a.info.userId, msg, answer)
+			jsonByteArray, err := json.Marshal(msg)
 			if err != nil {
-		replyMsg(*a.ctx, fmt.Sprintf(
+				log.Printf("Error marshaling JSON request: UserId: %s , Request: %s , Response: %s", a.info.userId, jsonByteArray, answer)
-			"🤖️：消息机器人摆烂了，请稍后再试～\n错误信息: %v", err), a.info.msgId)
+			}
+			jsonStr := strings.ReplaceAll(string(jsonByteArray), "\\n", "")
+			jsonStr = strings.ReplaceAll(jsonStr, "\n", "")
+			log.Printf("\n\n\n")
+			log.Printf("Success request plain jsonStr: UserId: %s , Request: %s , Response: %s",
+				a.info.userId, jsonStr, answer)
 			return false
 		}
-	return true
+	}
+}
+
+func printErrorMessage(a *ActionInfo, msg []openai.Messages, err error) {
+	log.Printf("Failed request: UserId: %s , Request: %s , Err: %s", a.info.userId, msg, err)
+}
+
+func sendOnProcess(a *ActionInfo) (*string, error) {
+	// send 正在处理中
+	cardId, err := sendOnProcessCard(*a.ctx, a.info.sessionId, a.info.msgId)
+	if err != nil {
+		return nil, err
+	}
+	return cardId, nil
+
 }

code/handlers/event_pic_action.go

@@ -1,107 +0,0 @@
-package handlers
-
-import (
-	"context"
-	"fmt"
-	"os"
-
-	"start-feishubot/initialization"
-	"start-feishubot/services"
-	"start-feishubot/services/openai"
-	"start-feishubot/utils"
-
-	larkim "github.com/larksuite/oapi-sdk-go/v3/service/im/v1"
-)
-
-type PicAction struct { /*图片*/
-}
-
-func (*PicAction) Execute(a *ActionInfo) bool {
-	check := AzureModeCheck(a)
-	if !check {
-		return true
-	}
-	// 开启图片创作模式
-	if _, foundPic := utils.EitherTrimEqual(a.info.qParsed,
-		"/picture", "图片创作"); foundPic {
-		a.handler.sessionCache.Clear(*a.info.sessionId)
-		a.handler.sessionCache.SetMode(*a.info.sessionId,
-			services.ModePicCreate)
-		a.handler.sessionCache.SetPicResolution(*a.info.sessionId,
-			services.Resolution256)
-		sendPicCreateInstructionCard(*a.ctx, a.info.sessionId,
-			a.info.msgId)
-		return false
-	}
-
-	mode := a.handler.sessionCache.GetMode(*a.info.sessionId)
-	//fmt.Println("mode: ", mode)
-
-	// 收到一张图片,且不在图片创作模式下, 提醒是否切换到图片创作模式
-	if a.info.msgType == "image" && mode != services.ModePicCreate {
-		sendPicModeCheckCard(*a.ctx, a.info.sessionId, a.info.msgId)
-		return false
-	}
-
-	if a.info.msgType == "image" && mode == services.ModePicCreate {
-		//保存图片
-		imageKey := a.info.imageKey
-		//fmt.Printf("fileKey: %s \n", imageKey)
-		msgId := a.info.msgId
-		//fmt.Println("msgId: ", *msgId)
-		req := larkim.NewGetMessageResourceReqBuilder().MessageId(
-			*msgId).FileKey(imageKey).Type("image").Build()
-		resp, err := initialization.GetLarkClient().Im.MessageResource.Get(context.Background(), req)
-		//fmt.Println(resp, err)
-		if err != nil {
-			//fmt.Println(err)
-			replyMsg(*a.ctx, fmt.Sprintf("🤖️：图片下载失败，请稍后再试～\n 错误信息: %v", err),
-				a.info.msgId)
-			return false
-		}
-
-		f := fmt.Sprintf("%s.png", imageKey)
-		resp.WriteFile(f)
-		defer os.Remove(f)
-		resolution := a.handler.sessionCache.GetPicResolution(*a.
-			info.sessionId)
-
-		openai.ConvertJpegToPNG(f)
-		openai.ConvertToRGBA(f, f)
-
-		//图片校验
-		err = openai.VerifyPngs([]string{f})
-		if err != nil {
-			replyMsg(*a.ctx, fmt.Sprintf("🤖️：无法解析图片，请发送原图并尝试重新操作～"),
-				a.info.msgId)
-			return false
-		}
-		bs64, err := a.handler.gpt.GenerateOneImageVariation(f, resolution)
-		if err != nil {
-			replyMsg(*a.ctx, fmt.Sprintf(
-				"🤖️：图片生成失败，请稍后再试～\n错误信息: %v", err), a.info.msgId)
-			return false
-		}
-		replayImagePlainByBase64(*a.ctx, bs64, a.info.msgId)
-		return false
-
-	}
-
-	// 生成图片
-	if mode == services.ModePicCreate {
-		resolution := a.handler.sessionCache.GetPicResolution(*a.
-			info.sessionId)
-		bs64, err := a.handler.gpt.GenerateOneImage(a.info.qParsed,
-			resolution)
-		if err != nil {
-			replyMsg(*a.ctx, fmt.Sprintf(
-				"🤖️：图片生成失败，请稍后再试～\n错误信息: %v", err), a.info.msgId)
-			return false
-		}
-		replayImageCardByBase64(*a.ctx, bs64, a.info.msgId, a.info.sessionId,
-			a.info.qParsed)
-		return false
-	}
-
-	return true
-}

code/handlers/handler.go

@@ -3,13 +3,14 @@ package handlers
 import (
 	"context"
 	"fmt"
-	"strings"
-
 	"start-feishubot/initialization"
 	"start-feishubot/services"
+	"start-feishubot/services/chatgpt"
 	"start-feishubot/services/openai"
+	"strings"
 
 	larkcard "github.com/larksuite/oapi-sdk-go/v3/card"
+
 	larkim "github.com/larksuite/oapi-sdk-go/v3/service/im/v1"
 )
 
@@ -40,11 +41,12 @@ func judgeMsgType(event *larkim.P2MessageReceiveV1) (string, error) {
 	msgType := event.Event.Message.MessageType
 
 	switch *msgType {
-	case "text", "image", "audio", "post":
+	case "text", "image", "audio":
 		return *msgType, nil
 	default:
 		return "", fmt.Errorf("unknown message type: %v", *msgType)
 	}
+
 }
 
 func (m MessageHandler) msgReceivedHandler(ctx context.Context, event *larkim.P2MessageReceiveV1) error {
@@ -75,8 +77,9 @@ func (m MessageHandler) msgReceivedHandler(ctx context.Context, event *larkim.P2
 		handlerType: handlerType,
 		msgType:     msgType,
 		msgId:       msgId,
+		userId:      *event.Event.Sender.SenderId.UserId,
 		chatId:      chatId,
-		qParsed:     strings.Trim(parseContent(*content, msgType), " "),
+		qParsed:     strings.Trim(parseContent(*content), " "),
 		fileKey:     parseFileKey(*content),
 		imageKey:    parseImageKey(*content),
 		sessionId:   sessionId,
@@ -90,18 +93,16 @@ func (m MessageHandler) msgReceivedHandler(ctx context.Context, event *larkim.P2
 	actions := []Action{
 		&ProcessedUniqueAction{}, //避免重复处理
 		&ProcessMentionAction{},  //判断机器人是否应该被调用
-		&AudioAction{},           //语音处理
 		&EmptyAction{},           //空消息处理
 		&ClearAction{},           //清除消息处理
-		&PicAction{},             //图片处理
-		&AIModeAction{},          //模式切换处理
 		&RoleListAction{},        //角色列表处理
 		&HelpAction{},            //帮助处理
-		&BalanceAction{},         //余额处理
 		&RolePlayAction{},        //角色扮演处理
-		&MessageAction{},         //消息处理
+		&MessageAction{
-
+			chatgpt: chatgpt.NewGpt3(&m.config),
+		}, //消息处理
 	}
+
 	chain(data, actions...)
 	return nil
 }

code/handlers/init.go

@@ -2,7 +2,6 @@ package handlers
 
 import (
 	"context"
-
 	"start-feishubot/initialization"
 	"start-feishubot/services/openai"
 

code/handlers/msg.go

@@ -6,7 +6,6 @@ import (
 	"encoding/base64"
 	"errors"
 	"fmt"
-
 	"start-feishubot/initialization"
 	"start-feishubot/services"
 	"start-feishubot/services/openai"
@@ -27,7 +26,6 @@ var (
 	PicVarMoreKind     = CardKind("pic_var_more")     // 变量图片
 	RoleTagsChooseKind = CardKind("role_tags_choose") // 内置角色所属标签选择
 	RoleChooseKind     = CardKind("role_choose")      // 内置角色选择
-	AIModeChooseKind   = CardKind("ai_mode_choose")   // AI模式选择
 )
 
 var (
@@ -76,14 +74,43 @@ func replyCard(ctx context.Context,
 	return nil
 }
 
-func newSendCard(
+func replyCardWithBackId(ctx context.Context,
-	header *larkcard.MessageCardHeader,
+	msgId *string,
-	elements ...larkcard.MessageCardElement) (string,
+	cardContent string,
-	error) {
+) (*string, error) {
+	client := initialization.GetLarkClient()
+	resp, err := client.Im.Message.Reply(ctx, larkim.NewReplyMessageReqBuilder().
+		MessageId(*msgId).
+		Body(larkim.NewReplyMessageReqBodyBuilder().
+			MsgType(larkim.MsgTypeInteractive).
+			Uuid(uuid.New().String()).
+			Content(cardContent).
+			Build()).
+		Build())
+
+	// 处理错误
+	if err != nil {
+		fmt.Println(err)
+		return nil, err
+	}
+
+	// 服务端错误处理
+	if !resp.Success() {
+		fmt.Println(resp.Code, resp.Msg, resp.RequestId())
+		return nil, errors.New(resp.Msg)
+	}
+
+	//ctx = context.WithValue(ctx, "SendMsgId", *resp.Data.MessageId)
+	//SendMsgId := ctx.Value("SendMsgId")
+	//pp.Println(SendMsgId)
+	return resp.Data.MessageId, nil
+}
+
+func newSendCard(header *larkcard.MessageCardHeader, elements ...larkcard.MessageCardElement) (string, error) {
 	config := larkcard.NewMessageCardConfig().
 		WideScreenMode(false).
 		EnableForward(true).
-		UpdateMulti(false).
+		UpdateMulti(true).
 		Build()
 	var aElementPool []larkcard.MessageCardElement
 	for _, element := range elements {
@@ -99,6 +126,26 @@ func newSendCard(
 		String()
 	return cardContent, err
 }
+func newSendCardWithOutHeader(
+	elements ...larkcard.MessageCardElement) (string, error) {
+	config := larkcard.NewMessageCardConfig().
+		WideScreenMode(false).
+		EnableForward(true).
+		UpdateMulti(true).
+		Build()
+	var aElementPool []larkcard.MessageCardElement
+	for _, element := range elements {
+		aElementPool = append(aElementPool, element)
+	}
+	// 卡片消息体
+	cardContent, err := larkcard.NewMessageCard().
+		Config(config).
+		Elements(
+			aElementPool,
+		).
+		String()
+	return cardContent, err
+}
 
 func newSimpleSendCard(
 	elements ...larkcard.MessageCardElement) (string,
@@ -354,7 +401,6 @@ func withPicResolutionBtn(sessionID *string) larkcard.
 		Build()
 	return actions
 }
-
 func withRoleTagsBtn(sessionID *string, tags ...string) larkcard.
 	MessageCardElement {
 	var menuOptions []MenuOption
@@ -409,32 +455,6 @@ func withRoleBtn(sessionID *string, titles ...string) larkcard.
 	return actions
 }
 
-func withAIModeBtn(sessionID *string, aiModeStrs []string) larkcard.MessageCardElement {
-	var menuOptions []MenuOption
-	for _, label := range aiModeStrs {
-		menuOptions = append(menuOptions, MenuOption{
-			label: label,
-			value: label,
-		})
-	}
-
-	cancelMenu := newMenu("选择模式",
-		map[string]interface{}{
-			"value":     "0",
-			"kind":      AIModeChooseKind,
-			"sessionId": *sessionID,
-			"msgId":     *sessionID,
-		},
-		menuOptions...,
-	)
-
-	actions := larkcard.NewMessageCardAction().
-		Actions([]larkcard.MessageCardActionElement{cancelMenu}).
-		Layout(larkcard.MessageCardActionLayoutFlow.Ptr()).
-		Build()
-	return actions
-}
-
 func replyMsg(ctx context.Context, msg string, msgId *string) error {
 	msg, i := processMessage(msg)
 	if i != nil {
@@ -496,7 +516,6 @@ func uploadImage(base64Str string) (*string, error) {
 	}
 	return resp.Data.ImageKey, nil
 }
-
 func replyImage(ctx context.Context, ImageKey *string,
 	msgId *string) error {
 	//fmt.Println("sendMsg", ImageKey, msgId)
@@ -530,6 +549,7 @@ func replyImage(ctx context.Context, ImageKey *string,
 		return errors.New(resp.Msg)
 	}
 	return nil
+
 }
 
 func replayImageCardByBase64(ctx context.Context, base64Str string,
@@ -548,38 +568,6 @@ func replayImageCardByBase64(ctx context.Context, base64Str string,
 	return nil
 }
 
-func replayImagePlainByBase64(ctx context.Context, base64Str string,
-	msgId *string) error {
-	imageKey, err := uploadImage(base64Str)
-	if err != nil {
-		return err
-	}
-	//example := "img_v2_041b28e3-5680-48c2-9af2-497ace79333g"
-	//imageKey := &example
-	//fmt.Println("imageKey", *imageKey)
-	err = replyImage(ctx, imageKey, msgId)
-	if err != nil {
-		return err
-	}
-	return nil
-}
-
-func replayVariantImageByBase64(ctx context.Context, base64Str string,
-	msgId *string, sessionId *string) error {
-	imageKey, err := uploadImage(base64Str)
-	if err != nil {
-		return err
-	}
-	//example := "img_v2_041b28e3-5680-48c2-9af2-497ace79333g"
-	//imageKey := &example
-	//fmt.Println("imageKey", *imageKey)
-	err = sendVarImageCard(ctx, *imageKey, msgId, sessionId)
-	if err != nil {
-		return err
-	}
-	return nil
-}
-
 func sendMsg(ctx context.Context, msg string, chatId *string) error {
 	//fmt.Println("sendMsg", msg, chatId)
 	msg, i := processMessage(msg)
@@ -616,6 +604,37 @@ func sendMsg(ctx context.Context, msg string, chatId *string) error {
 	return nil
 }
 
+func PatchCard(ctx context.Context, msgId *string,
+	cardContent string) error {
+	//fmt.Println("sendMsg", msg, chatId)
+	client := initialization.GetLarkClient()
+	//content := larkim.NewTextMsgBuilder().
+	//	Text(msg).
+	//	Build()
+
+	//fmt.Println("content", content)
+
+	resp, err := client.Im.Message.Patch(ctx, larkim.NewPatchMessageReqBuilder().
+		MessageId(*msgId).
+		Body(larkim.NewPatchMessageReqBodyBuilder().
+			Content(cardContent).
+			Build()).
+		Build())
+
+	// 处理错误
+	if err != nil {
+		fmt.Println(err)
+		return err
+	}
+
+	// 服务端错误处理
+	if !resp.Success() {
+		fmt.Println(resp.Code, resp.Msg, resp.RequestId())
+		return errors.New(resp.Msg)
+	}
+	return nil
+}
+
 func sendClearCacheCheckCard(ctx context.Context,
 	sessionId *string, msgId *string) {
 	newCard, _ := newSendCard(
@@ -635,39 +654,47 @@ func sendSystemInstructionCard(ctx context.Context,
 	replyCard(ctx, msgId, newCard)
 }
 
-func sendPicCreateInstructionCard(ctx context.Context,
+func sendOnProcessCard(ctx context.Context,
-	sessionId *string, msgId *string) {
+	sessionId *string, msgId *string) (*string, error) {
-	newCard, _ := newSendCard(
+	newCard, _ := newSendCardWithOutHeader(
-		withHeader("🖼️ 已进入图片创作模式", larkcard.TemplateBlue),
+		withNote("正在思考，请稍等..."))
-		withPicResolutionBtn(sessionId),
+	id, err := replyCardWithBackId(ctx, msgId, newCard)
-		withNote("提醒：回复文本或图片，让AI生成相关的图片。"))
+	if err != nil {
-	replyCard(ctx, msgId, newCard)
+		return nil, err
+	}
+	return id, nil
 }
 
-func sendPicModeCheckCard(ctx context.Context,
+func updateTextCard(ctx context.Context, msg string,
-	sessionId *string, msgId *string) {
+	msgId *string) error {
-	newCard, _ := newSendCard(
+	newCard, _ := newSendCardWithOutHeader(
-		withHeader("🖼️ 机器人提醒", larkcard.TemplateBlue),
+		withMainText(msg),
-		withMainMd("收到图片，是否进入图片创作模式？"),
+		withNote("正在生成，请稍等..."))
-		withNote("请注意，这将开始一个全新的对话，您将无法利用之前话题的历史信息"),
+	err := PatchCard(ctx, msgId, newCard)
-		withPicModeDoubleCheckBtn(sessionId))
+	if err != nil {
-	replyCard(ctx, msgId, newCard)
+		return err
+	}
+	return nil
 }
-
+func updateFinalCard(
-func sendNewTopicCard(ctx context.Context,
+	ctx context.Context,
-	sessionId *string, msgId *string, content string) {
+	msg string,
-	newCard, _ := newSendCard(
+	msgId *string,
-		withHeader("👻️ 已开启新的话题", larkcard.TemplateBlue),
+) error {
-		withMainText(content),
+	newCard, _ := newSendCardWithOutHeader(
-		withNote("提醒：点击对话框参与回复，可保持话题连贯"))
+		withMainText(msg))
-	replyCard(ctx, msgId, newCard)
+	err := PatchCard(ctx, msgId, newCard)
+	if err != nil {
+		return err
+	}
+	return nil
 }
 
 func sendHelpCard(ctx context.Context,
 	sessionId *string, msgId *string) {
 	newCard, _ := newSendCard(
 		withHeader("🎒需要帮助吗？", larkcard.TemplateBlue),
-		withMainMd("**我是CZLChat-Feishu，一款基于ChatGPT[模型：gpt-3.5-0613]技术的智能聊天机器人**"),
+		withMainMd("**我是具备打字机效果的聊天机器人！**"),
 		withSplitLine(),
 		withMdAndExtraBtn(
 			"** 🆑 清除话题上下文**\n文本回复 *清除* 或 */clear*",
@@ -677,25 +704,9 @@ func sendHelpCard(ctx context.Context,
 				"chatType":  UserChatType,
 				"sessionId": *sessionId,
 			}, larkcard.MessageCardButtonTypeDanger)),
-		withSplitLine(),
-		withMainMd("🤖 **AI模式选择** \n"+" 文本回复 *AI模式* 或 */ai_mode*"),
-		withSplitLine(),
 		withMainMd("🛖 **内置角色列表** \n"+" 文本回复 *角色列表* 或 */roles*"),
-		withSplitLine(),
 		withMainMd("🥷 **角色扮演模式**\n文本回复*角色扮演* 或 */system*+空格+角色信息"),
 		withSplitLine(),
-		withMainMd("🎤 **AI语音对话**\n私聊模式下直接发送语音"),
-		withSplitLine(),
-		withMainMd("🎨 **图片创作模式**\n回复*图片创作* 或 */picture*"),
-		withSplitLine(),
-		withMainMd("🎰 **Token余额查询**\n回复*余额* 或 */balance*"),
-		withSplitLine(),
-		withMainMd("🔃️ **历史话题回档** 🚧\n"+" 进入话题的回复详情页,文本回复 *恢复* 或 */reload*"),
-		withSplitLine(),
-		withMainMd("📤 **话题内容导出** 🚧\n"+" 文本回复 *导出* 或 */export*"),
-		withSplitLine(),
-		withMainMd("🎰 **连续对话与多话题模式**\n"+" 点击对话框参与回复，可保持话题连贯。同时，单独提问即可开启全新新话题"),
-		withSplitLine(),
 		withMainMd("🎒 **需要更多帮助**\n文本回复 *帮助* 或 */help*"),
 	)
 	replyCard(ctx, msgId, newCard)
@@ -719,24 +730,6 @@ func sendImageCard(ctx context.Context, imageKey string,
 	return nil
 }
 
-func sendVarImageCard(ctx context.Context, imageKey string,
-	msgId *string, sessionId *string) error {
-	newCard, _ := newSimpleSendCard(
-		withImageDiv(imageKey),
-		withSplitLine(),
-		//再来一张
-		withOneBtn(newBtn("再来一张", map[string]interface{}{
-			"value":     imageKey,
-			"kind":      PicVarMoreKind,
-			"chatType":  UserChatType,
-			"msgId":     *msgId,
-			"sessionId": *sessionId,
-		}, larkcard.MessageCardButtonTypePrimary)),
-	)
-	replyCard(ctx, msgId, newCard)
-	return nil
-}
-
 func sendBalanceCard(ctx context.Context, msgId *string,
 	balance openai.BalanceResponse) {
 	newCard, _ := newSendCard(
@@ -769,12 +762,3 @@ func SendRoleListCard(ctx context.Context,
 		withNote("提醒：选择内置场景，快速进入角色扮演模式。"))
 	replyCard(ctx, msgId, newCard)
 }
-
-func SendAIModeListsCard(ctx context.Context,
-	sessionId *string, msgId *string, aiModeStrs []string) {
-	newCard, _ := newSendCard(
-		withHeader("🤖 AI模式选择", larkcard.TemplateIndigo),
-		withAIModeBtn(sessionId, aiModeStrs),
-		withNote("提醒：选择内置模式，让AI更好的理解您的需求。"))
-	replyCard(ctx, msgId, newCard)
-}

code/initialization/config.go

@@ -2,14 +2,19 @@ package initialization
 
 import (
 	"fmt"
+	"github.com/spf13/pflag"
 	"os"
 	"strconv"
 	"strings"
+	"sync"
 
 	"github.com/spf13/viper"
 )
 
 type Config struct {
+	// 表示配置是否已经被初始化了。
+	Initialized                        bool
+	EnableLog                          bool
 	FeishuAppId                        string
 	FeishuAppSecret                    string
 	FeishuAppEncryptKey                string
@@ -28,8 +33,35 @@ type Config struct {
 	AzureDeploymentName                string
 	AzureResourceName                  string
 	AzureOpenaiToken                   string
+	AccessControlEnable                bool
+	AccessControlMaxCountPerUserPerDay int
+	OpenAIHttpClientTimeOut            int
+	OpenaiModel                        string
 }
 
+var (
+	cfg    = pflag.StringP("config", "c", "./config.yaml", "apiserver config file path.")
+	config *Config
+	once   sync.Once
+)
+
+/*
+GetConfig will call LoadConfig once and return a global singleton, you should always use this function to get config
+*/
+func GetConfig() *Config {
+
+	once.Do(func() {
+		config = LoadConfig(*cfg)
+		config.Initialized = true
+	})
+
+	return config
+}
+
+/*
+LoadConfig will load config and should only be called once, you should always use GetConfig to get config rather than
+call this function directly
+*/
 func LoadConfig(cfg string) *Config {
 	viper.SetConfigFile(cfg)
 	viper.ReadInConfig()
@@ -41,6 +73,7 @@ func LoadConfig(cfg string) *Config {
 	//fmt.Println(string(content))
 
 	config := &Config{
+		EnableLog:                          getViperBoolValue("ENABLE_LOG", false),
 		FeishuAppId:                        getViperStringValue("APP_ID", ""),
 		FeishuAppSecret:                    getViperStringValue("APP_SECRET", ""),
 		FeishuAppEncryptKey:                getViperStringValue("APP_ENCRYPT_KEY", ""),
@@ -52,13 +85,17 @@ func LoadConfig(cfg string) *Config {
 		UseHttps:                           getViperBoolValue("USE_HTTPS", false),
 		CertFile:                           getViperStringValue("CERT_FILE", "cert.pem"),
 		KeyFile:                            getViperStringValue("KEY_FILE", "key.pem"),
-		OpenaiApiUrl:               getViperStringValue("API_URL", "https://oapi.czl.net"),
+		OpenaiApiUrl:                       getViperStringValue("API_URL", "https://api.openai.com"),
 		HttpProxy:                          getViperStringValue("HTTP_PROXY", ""),
 		AzureOn:                            getViperBoolValue("AZURE_ON", false),
 		AzureApiVersion:                    getViperStringValue("AZURE_API_VERSION", "2023-03-15-preview"),
 		AzureDeploymentName:                getViperStringValue("AZURE_DEPLOYMENT_NAME", ""),
 		AzureResourceName:                  getViperStringValue("AZURE_RESOURCE_NAME", ""),
 		AzureOpenaiToken:                   getViperStringValue("AZURE_OPENAI_TOKEN", ""),
+		AccessControlEnable:                getViperBoolValue("ACCESS_CONTROL_ENABLE", false),
+		AccessControlMaxCountPerUserPerDay: getViperIntValue("ACCESS_CONTROL_MAX_COUNT_PER_USER_PER_DAY", 0),
+		OpenAIHttpClientTimeOut:            getViperIntValue("OPENAI_HTTP_CLIENT_TIMEOUT", 550),
+		OpenaiModel:                        getViperStringValue("OPENAI_MODEL", "gpt-3.5-turbo"),
 	}
 
 	return config
@@ -72,8 +109,8 @@ func getViperStringValue(key string, defaultValue string) string {
 	return value
 }
 
-//OPENAI_KEY: sk-xxx,sk-xxx,sk-xxx
+// OPENAI_KEY: sk-xxx,sk-xxx,sk-xxx
-//result:[sk-xxx sk-xxx sk-xxx]
+// result:[sk-xxx sk-xxx sk-xxx]
 func getViperStringArray(key string, defaultValue []string) []string {
 	value := viper.GetString(key)
 	if value == "" {
@@ -135,8 +172,7 @@ func (config *Config) GetKeyFile() string {
 func filterFormatKey(keys []string) []string {
 	var result []string
 	for _, key := range keys {
-		if strings.HasPrefix(key, "sk-") || strings.HasPrefix(key,
+		if strings.HasPrefix(key, "sk-") {
-			"fk") {
 			result = append(result, key)
 		}
 	}

code/initialization/roles_load.go

@@ -2,12 +2,11 @@ package initialization
 
 import (
 	"errors"
-	"io/ioutil"
-	"log"
-
 	"github.com/duke-git/lancet/v2/slice"
 	"github.com/duke-git/lancet/v2/validator"
 	"gopkg.in/yaml.v2"
+	"io/ioutil"
+	"log"
 )
 
 type Role struct {

code/main.go

@@ -2,41 +2,54 @@ package main
 
 import (
 	"context"
+	"encoding/json"
+	"fmt"
+	larkcard "github.com/larksuite/oapi-sdk-go/v3/card"
+	larkim "github.com/larksuite/oapi-sdk-go/v3/service/im/v1"
+	"gopkg.in/natefinch/lumberjack.v2"
+	"io"
 	"log"
-
+	"os"
 	"start-feishubot/handlers"
 	"start-feishubot/initialization"
 	"start-feishubot/services/openai"
+	"start-feishubot/utils"
 
 	"github.com/gin-gonic/gin"
-	sdkginext "github.com/larksuite/oapi-sdk-gin"
-	larkcard "github.com/larksuite/oapi-sdk-go/v3/card"
-	"github.com/larksuite/oapi-sdk-go/v3/event/dispatcher"
-	larkim "github.com/larksuite/oapi-sdk-go/v3/service/im/v1"
 	"github.com/spf13/pflag"
-)
 
-var (
+	sdkginext "github.com/larksuite/oapi-sdk-gin"
-	cfg = pflag.StringP("config", "c", "./config.yaml", "apiserver config file path.")
+
+	"github.com/larksuite/oapi-sdk-go/v3/event/dispatcher"
 )
 
 func main() {
 	initialization.InitRoleList()
 	pflag.Parse()
-	config := initialization.LoadConfig(*cfg)
+	globalConfig := initialization.GetConfig()
-	initialization.LoadLarkClient(*config)
+
-	gpt := openai.NewChatGPT(*config)
+	// 打印一下实际读取到的配置
-	handlers.InitHandlers(gpt, *config)
+	globalConfigPrettyString, _ := json.MarshalIndent(globalConfig, "", "    ")
+	log.Println(string(globalConfigPrettyString))
+
+	initialization.LoadLarkClient(*globalConfig)
+	gpt := openai.NewChatGPT(*globalConfig)
+	handlers.InitHandlers(gpt, *globalConfig)
+
+	if globalConfig.EnableLog {
+		logger := enableLog()
+		defer utils.CloseLogger(logger)
+	}
 
 	eventHandler := dispatcher.NewEventDispatcher(
-		config.FeishuAppVerificationToken, config.FeishuAppEncryptKey).
+		globalConfig.FeishuAppVerificationToken, globalConfig.FeishuAppEncryptKey).
 		OnP2MessageReceiveV1(handlers.Handler).
 		OnP2MessageReadV1(func(ctx context.Context, event *larkim.P2MessageReadV1) error {
 			return handlers.ReadHandler(ctx, event)
 		})
 
 	cardHandler := larkcard.NewCardActionHandler(
-		config.FeishuAppVerificationToken, config.FeishuAppEncryptKey,
+		globalConfig.FeishuAppVerificationToken, globalConfig.FeishuAppEncryptKey,
 		handlers.CardHandler())
 
 	r := gin.Default()
@@ -51,7 +64,31 @@ func main() {
 		sdkginext.NewCardActionHandlerFunc(
 			cardHandler))
 
-	if err := initialization.StartServer(*config, r); err != nil {
+	err := initialization.StartServer(*globalConfig, r)
+	if err != nil {
 		log.Fatalf("failed to start server: %v", err)
 	}
+
+}
+
+func enableLog() *lumberjack.Logger {
+	// Set up the logger
+	var logger *lumberjack.Logger
+
+	logger = &lumberjack.Logger{
+		Filename: "logs/app.log",
+		MaxSize:  100,      // megabytes
+		MaxAge:   365 * 10, // days
+	}
+
+	fmt.Printf("logger %T\n", logger)
+
+	// Set up the logger to write to both file and console
+	log.SetOutput(io.MultiWriter(logger, os.Stdout))
+	log.SetFlags(log.Ldate | log.Ltime)
+
+	// Write some log messages
+	log.Println("Starting application...")
+
+	return logger
 }

code/role_list.yaml

@@ -1,7 +1,3 @@
-# 可在此处提交你认为不错的角色预设，注意保持格式一致。
-# PR 时的 tag 暂时集中在 [ "日常办公",  "生活助手" ,"代码专家", "文案撰写"]
-# 更多点子可参考我另一个参与的项目: https://open-gpt.app/
-
 - title: 周报生成
   content: 请帮我把以下的工作内容填充为一篇完整的周报，用 markdown 格式以分点叙述的形式输出：
   example: 重新优化设计稿，和前端再次沟通 UI 细节，确保落地

code/services/accesscontrol/access_control.go

@@ -0,0 +1,66 @@
+package accesscontrol
+
+import (
+	"start-feishubot/initialization"
+	"start-feishubot/utils"
+	"sync"
+)
+
+var accessCountMap = sync.Map{}
+var currentDateFlag = ""
+
+/*
+CheckAllowAccessThenIncrement If user has accessed more than 100 times according to accessCountMap, return false.
+Otherwise, return true and increase the access count by 1
+*/
+func CheckAllowAccessThenIncrement(userId *string) bool {
+
+	// Begin a new day, clear the accessCountMap
+	currentDateAsString := utils.GetCurrentDateAsString()
+	if currentDateFlag != currentDateAsString {
+		accessCountMap = sync.Map{}
+		currentDateFlag = currentDateAsString
+	}
+
+	if CheckAllowAccess(userId) {
+		accessedCount, ok := accessCountMap.Load(*userId)
+		if !ok {
+			accessCountMap.Store(*userId, 1)
+		} else {
+			accessCountMap.Store(*userId, accessedCount.(int)+1)
+		}
+		return true
+	} else {
+		return false
+	}
+}
+
+func CheckAllowAccess(userId *string) bool {
+
+	if initialization.GetConfig().AccessControlMaxCountPerUserPerDay <= 0 {
+		return true
+	}
+
+	accessedCount, ok := accessCountMap.Load(*userId)
+
+	if !ok {
+		accessCountMap.Store(*userId, 0)
+		return true
+	}
+
+	// If the user has accessed more than 100 times, return false
+	if accessedCount.(int) >= initialization.GetConfig().AccessControlMaxCountPerUserPerDay {
+		return false
+	}
+
+	// Otherwise, return true
+	return true
+}
+
+func GetCurrentDateFlag() string {
+	return currentDateFlag
+}
+
+func GetAccessCountMap() *sync.Map {
+	return &accessCountMap
+}

code/services/chatgpt/check.go

@@ -0,0 +1,33 @@
+package chatgpt
+
+import (
+	"errors"
+	"github.com/sashabaranov/go-openai"
+)
+
+const (
+	ChatMessageRoleSystem    = "system"
+	ChatMessageRoleUser      = "user"
+	ChatMessageRoleAssistant = "assistant"
+)
+
+func CheckChatCompletionMessages(messages []openai.ChatCompletionMessage) error {
+	hasSystemMsg := false
+	for _, msg := range messages {
+		if msg.Role != ChatMessageRoleSystem && msg.Role != ChatMessageRoleUser && msg.Role != ChatMessageRoleAssistant {
+			return errors.New("invalid message role")
+		}
+		if msg.Role == ChatMessageRoleSystem {
+			if hasSystemMsg {
+				return errors.New("more than one system message")
+			}
+			hasSystemMsg = true
+		} else {
+			// 对于非 system 角色的消息，Content 不能为空
+			if msg.Content == "" {
+				return errors.New("empty content in non-system message")
+			}
+		}
+	}
+	return nil
+}

code/services/chatgpt/gpt3.go

@@ -0,0 +1,90 @@
+package chatgpt
+
+import (
+	"context"
+	"errors"
+	"fmt"
+	"github.com/sashabaranov/go-openai"
+	"io"
+	"start-feishubot/initialization"
+	customOpenai "start-feishubot/services/openai"
+)
+
+type Messages struct {
+	Role    string `json:"role"`
+	Content string `json:"content"`
+}
+
+type ChatGPT struct {
+	config *initialization.Config
+}
+
+type Gpt3 interface {
+	StreamChat() error
+	StreamChatWithHistory() error
+}
+
+func NewGpt3(config *initialization.Config) *ChatGPT {
+	return &ChatGPT{config: config}
+}
+
+func (c *ChatGPT) StreamChat(ctx context.Context,
+	msg []customOpenai.Messages,
+	responseStream chan string) error {
+	//change msg type from Messages to openai.ChatCompletionMessage
+	chatMsgs := make([]openai.ChatCompletionMessage, len(msg))
+	for i, m := range msg {
+		chatMsgs[i] = openai.ChatCompletionMessage{
+			Role:    m.Role,
+			Content: m.Content,
+		}
+	}
+	return c.StreamChatWithHistory(ctx, chatMsgs, 2000,
+		responseStream)
+}
+
+func (c *ChatGPT) StreamChatWithHistory(ctx context.Context, msg []openai.ChatCompletionMessage, maxTokens int,
+	responseStream chan string,
+) error {
+	config := openai.DefaultConfig(c.config.OpenaiApiKeys[0])
+	config.BaseURL = c.config.OpenaiApiUrl + "/v1"
+
+	proxyClient, parseProxyError := customOpenai.GetProxyClient(c.config.HttpProxy)
+	if parseProxyError != nil {
+		return parseProxyError
+	}
+	config.HTTPClient = proxyClient
+
+	client := openai.NewClientWithConfig(config)
+	//pp.Printf("client: %v", client)
+	req := openai.ChatCompletionRequest{
+		Model:       c.config.OpenaiModel,
+		Messages:    msg,
+		N:           1,
+		Temperature: 0.7,
+		MaxTokens:   maxTokens,
+		TopP:        1,
+		//Moderation:     true,
+		//ModerationStop: true,
+	}
+	stream, err := client.CreateChatCompletionStream(ctx, req)
+	if err != nil {
+		fmt.Errorf("CreateCompletionStream returned error: %v", err)
+	}
+
+	defer stream.Close()
+	for {
+		response, err := stream.Recv()
+		if errors.Is(err, io.EOF) {
+			//fmt.Println("Stream finished")
+			return nil
+		}
+		if err != nil {
+			fmt.Printf("Stream error: %v\n", err)
+			return err
+		}
+		responseStream <- response.Choices[0].Delta.Content
+	}
+	return nil
+
+}

code/services/chatgpt/gpt3_test.go

@@ -0,0 +1,62 @@
+package chatgpt
+
+import (
+	"context"
+	"fmt"
+	"start-feishubot/initialization"
+	"start-feishubot/services/openai"
+	"testing"
+	"time"
+)
+
+func TestChatGPT_streamChat(t *testing.T) {
+	// 初始化配置
+	config := initialization.LoadConfig("../../config.yaml")
+
+	// 准备测试用例
+	testCases := []struct {
+		msg        []openai.Messages
+		wantOutput string
+		wantErr    bool
+	}{
+		{
+			msg: []openai.Messages{
+				{
+					Role:    "system",
+					Content: "从现在起你要化身职场语言大师，你需要用婉转的方式回复老板想你提出的问题，或像领导提出请求。",
+				},
+				{
+					Role:    "user",
+					Content: "领导，我想请假一天",
+				},
+			},
+			wantOutput: "",
+			wantErr:    false,
+		},
+	}
+
+	// 执行测试用例
+	for _, tc := range testCases {
+		// 准备输入和输出
+		responseStream := make(chan string)
+		ctx := context.Background()
+		c := &ChatGPT{config: config}
+
+		// 启动一个协程来模拟流式聊天
+		go func() {
+			err := c.StreamChat(ctx, tc.msg, responseStream)
+			if err != nil {
+				t.Errorf("streamChat() error = %v, wantErr %v", err, tc.wantErr)
+			}
+		}()
+
+		// 等待输出并检查是否符合预期
+		select {
+		case gotOutput := <-responseStream:
+			fmt.Printf("gotOutput: %v\n", gotOutput)
+
+		case <-time.After(5 * time.Second):
+			t.Errorf("streamChat() timeout, expected output not received")
+		}
+	}
+}

code/services/chatgpt/tokenizer.go

@@ -0,0 +1,20 @@
+package chatgpt
+
+import (
+	"github.com/pandodao/tokenizer-go"
+	"github.com/sashabaranov/go-openai"
+	"strings"
+)
+
+func CalcTokenLength(text string) int {
+	text = strings.TrimSpace(text)
+	return tokenizer.MustCalToken(text)
+}
+
+func CalcTokenFromMsgList(msgs []openai.ChatCompletionMessage) int {
+	var total int
+	for _, msg := range msgs {
+		total += CalcTokenLength(msg.Content)
+	}
+	return total
+}

code/services/chatgpt/tokenizer_test.go

@@ -0,0 +1,50 @@
+package chatgpt
+
+import "testing"
+
+func TestCalcTokenLength(t *testing.T) {
+	type args struct {
+		text string
+	}
+	tests := []struct {
+		name string
+		args args
+		want int
+	}{
+		{
+			name: "eng",
+			args: args{
+				text: "hello world",
+			},
+			want: 2,
+		},
+		{
+			name: "cn",
+			args: args{
+				text: "我和我的祖国",
+			},
+			want: 13,
+		},
+		{
+			name: "empty",
+			args: args{
+				text: "",
+			},
+			want: 0,
+		},
+		{
+			name: "empty",
+			args: args{
+				text: " ",
+			},
+			want: 0,
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if got := CalcTokenLength(tt.args.text); got != tt.want {
+				t.Errorf("CalcTokenLength() = %v, want %v", got, tt.want)
+			}
+		})
+	}
+}

code/services/msgCache.go

@@ -1,9 +1,8 @@
 package services
 
 import (
-	"time"
-
 	"github.com/patrickmn/go-cache"
+	"time"
 )
 
 type MsgService struct {

code/services/openai/billing.go

@@ -6,12 +6,23 @@ import (
 	"time"
 )
 
-type BillingSubScrip struct {
+//https://api.openai.com/dashboard/billing/credit_grants
-	HardLimitUsd float64 `json:"hard_limit_usd"`
+type Billing struct {
-	AccessUntil  float64 `json:"access_until"`
+	Object         string  `json:"object"`
-}
+	TotalGranted   float64 `json:"total_granted"`
-type BillingUsage struct {
+	TotalUsed      float64 `json:"total_used"`
-	TotalUsage float64 `json:"total_usage"`
+	TotalAvailable float64 `json:"total_available"`
+	Grants         struct {
+		Object string `json:"object"`
+		Data   []struct {
+			Object      string  `json:"object"`
+			ID          string  `json:"id"`
+			GrantAmount float64 `json:"grant_amount"`
+			UsedAmount  float64 `json:"used_amount"`
+			EffectiveAt float64 `json:"effective_at"`
+			ExpiresAt   float64 `json:"expires_at"`
+		} `json:"data"`
+	} `json:"grants"`
 }
 
 type BalanceResponse struct {
@@ -23,47 +34,29 @@ type BalanceResponse struct {
 }
 
 func (gpt *ChatGPT) GetBalance() (*BalanceResponse, error) {
-	fmt.Println("进入")
+	var data Billing
-	var data1 BillingSubScrip
 	err := gpt.sendRequestWithBodyType(
-		gpt.ApiUrl+"/v1/dashboard/billing/subscription",
+		gpt.ApiUrl+"/dashboard/billing/credit_grants",
 		http.MethodGet,
 		nilBody,
 		nil,
-		&data1,
+		&data,
 	)
-	fmt.Println("出错1", err)
 	if err != nil {
-		return nil, fmt.Errorf("failed to get billing subscription: %v", err)
+		return nil, fmt.Errorf("failed to get billing data: %v", err)
-	}
-	nowdate := time.Now()
-	enddate := nowdate.Format("2006-01-02")
-	startdate := nowdate.AddDate(0, 0, -100).Format("2006-01-02")
-	var data2 BillingUsage
-	err = gpt.sendRequestWithBodyType(
-		gpt.ApiUrl+fmt.Sprintf("/v1/dashboard/billing/usage?start_date=%s&end_date=%s", startdate, enddate),
-		http.MethodGet,
-		nilBody,
-		nil,
-		&data2,
-	)
-	fmt.Println(data2)
-	fmt.Println("出错2", err)
-	if err != nil {
-		return nil, fmt.Errorf("failed to get billing subscription: %v", err)
 	}
 
 	balance := &BalanceResponse{
-		TotalGranted:   data1.HardLimitUsd,
+		TotalGranted:   data.TotalGranted,
-		TotalUsed:      data2.TotalUsage / 100,
+		TotalUsed:      data.TotalUsed,
-		TotalAvailable: data1.HardLimitUsd - data2.TotalUsage/100,
+		TotalAvailable: data.TotalAvailable,
 		ExpiresAt:      time.Now(),
 		EffectiveAt:    time.Now(),
 	}
 
-	if data1.AccessUntil > 0 {
+	if len(data.Grants.Data) > 0 {
-		balance.EffectiveAt = time.Now()
+		balance.EffectiveAt = time.Unix(int64(data.Grants.Data[0].EffectiveAt), 0)
-		balance.ExpiresAt = time.Unix(int64(data1.AccessUntil), 0)
+		balance.ExpiresAt = time.Unix(int64(data.Grants.Data[0].ExpiresAt), 0)
 	}
 
 	return balance, nil

code/services/openai/common.go

@@ -9,11 +9,10 @@ import (
 	"mime/multipart"
 	"net/http"
 	"net/url"
-	"strings"
-	"time"
-
 	"start-feishubot/initialization"
 	"start-feishubot/services/loadbalancer"
+	"strings"
+	"time"
 )
 
 type PlatForm string
@@ -38,6 +37,7 @@ type ChatGPT struct {
 	Lb          *loadbalancer.LoadBalancer
 	ApiKey      []string
 	ApiUrl      string
+	ApiModel    string
 	HttpProxy   string
 	Platform    PlatForm
 	AzureConfig AzureConfig
@@ -48,7 +48,7 @@ const (
 	jsonBody requestBodyType = iota
 	formVoiceDataBody
 	formPictureDataBody
-
+	streamBody
 	nilBody
 )
 
@@ -91,6 +91,7 @@ func (gpt *ChatGPT) doAPIRequestWithRetry(url, method string,
 			return err
 		}
 		requestBodyData = formBody.Bytes()
+
 	case nilBody:
 		requestBodyData = nil
 
@@ -111,6 +112,11 @@ func (gpt *ChatGPT) doAPIRequestWithRetry(url, method string,
 	if bodyType == formVoiceDataBody || bodyType == formPictureDataBody {
 		req.Header.Set("Content-Type", writer.FormDataContentType())
 	}
+	if bodyType == streamBody {
+		req.Header.Set("Accept", "text/event-stream")
+		req.Header.Set("Connection", "keep-alive")
+		req.Header.Set("Cache-Control", "no-cache")
+	}
 	if gpt.Platform == OpenAI {
 		req.Header.Set("Authorization", "Bearer "+api.Key)
 	} else {
@@ -120,10 +126,6 @@ func (gpt *ChatGPT) doAPIRequestWithRetry(url, method string,
 	var response *http.Response
 	var retry int
 	for retry = 0; retry <= maxRetries; retry++ {
-		// set body
-		if retry > 0 {
-			req.Body = ioutil.NopCloser(bytes.NewReader(requestBodyData))
-		}
 		response, err = client.Do(req)
 		//fmt.Println("--------------------")
 		//fmt.Println("req", req.Header)
@@ -135,7 +137,7 @@ func (gpt *ChatGPT) doAPIRequestWithRetry(url, method string,
 			fmt.Println("body", string(body))
 
 			gpt.Lb.SetAvailability(api.Key, false)
-			if retry == maxRetries {
+			if retry == maxRetries || bodyType == streamBody {
 				break
 			}
 			time.Sleep(time.Duration(retry+1) * time.Second)
@@ -169,27 +171,38 @@ func (gpt *ChatGPT) sendRequestWithBodyType(link, method string,
 	bodyType requestBodyType,
 	requestBody interface{}, responseBody interface{}) error {
 	var err error
-	client := &http.Client{Timeout: 110 * time.Second}
+	proxyString := gpt.HttpProxy
-	if gpt.HttpProxy == "" {
+
+	client, parseProxyError := GetProxyClient(proxyString)
+	if parseProxyError != nil {
+		return parseProxyError
+	}
+
 	err = gpt.doAPIRequestWithRetry(link, method, bodyType,
 		requestBody, responseBody, client, 3)
-	} else {
+
-		proxyUrl, err := url.Parse(gpt.HttpProxy)
-		if err != nil {
 	return err
+}
+
+func GetProxyClient(proxyString string) (*http.Client, error) {
+	var client *http.Client
+	timeOutDuration := time.Duration(initialization.GetConfig().OpenAIHttpClientTimeOut) * time.Second
+	if proxyString == "" {
+		client = &http.Client{Timeout: timeOutDuration}
+	} else {
+		proxyUrl, err := url.Parse(proxyString)
+		if err != nil {
+			return nil, err
 		}
 		transport := &http.Transport{
 			Proxy: http.ProxyURL(proxyUrl),
 		}
-		proxyClient := &http.Client{
+		client = &http.Client{
 			Transport: transport,
-			Timeout:   110 * time.Second,
+			Timeout:   timeOutDuration,
 		}
-		err = gpt.doAPIRequestWithRetry(link, method, bodyType,
-			requestBody, responseBody, proxyClient, 3)
 	}
-
+	return client, nil
-	return err
 }
 
 func NewChatGPT(config initialization.Config) *ChatGPT {
@@ -212,6 +225,7 @@ func NewChatGPT(config initialization.Config) *ChatGPT {
 		ApiUrl:    config.OpenaiApiUrl,
 		HttpProxy: config.HttpProxy,
 		Platform:  platform,
+		ApiModel:  config.OpenaiModel,
 		AzureConfig: AzureConfig{
 			BaseURL:        AzureApiUrlV1,
 			ResourceName:   config.AzureResourceName,

code/services/openai/gpt3.go

@@ -2,13 +2,8 @@ package openai
 
 import (
 	"errors"
-	"strings"
-
-	"github.com/pandodao/tokenizer-go"
 )
 
-type AIMode float64
-
 const (
 	Fresh      AIMode = 0.1
 	Warmth     AIMode = 0.4
@@ -49,7 +44,6 @@ type ChatGPTResponseBody struct {
 	Choices []ChatGPTChoiceItem    `json:"choices"`
 	Usage   map[string]interface{} `json:"usage"`
 }
-
 type ChatGPTChoiceItem struct {
 	Message      Messages `json:"message"`
 	Index        int      `json:"index"`
@@ -61,24 +55,20 @@ type ChatGPTRequestBody struct {
 	Model            string     `json:"model"`
 	Messages         []Messages `json:"messages"`
 	MaxTokens        int        `json:"max_tokens"`
-	Temperature      AIMode     `json:"temperature"`
+	Temperature      float32    `json:"temperature"`
 	TopP             int        `json:"top_p"`
 	FrequencyPenalty int        `json:"frequency_penalty"`
 	PresencePenalty  int        `json:"presence_penalty"`
+	Stream           bool       `json:"stream" default:"false"`
 }
 
-func (msg *Messages) CalculateTokenLength() int {
+func (gpt *ChatGPT) Completions(msg []Messages) (resp Messages,
-	text := strings.TrimSpace(msg.Content)
-	return tokenizer.MustCalToken(text)
-}
-
-func (gpt *ChatGPT) Completions(msg []Messages, aiMode AIMode) (resp Messages,
 	err error) {
 	requestBody := ChatGPTRequestBody{
-		Model:            engine,
+		Model:            gpt.ApiModel,
 		Messages:         msg,
 		MaxTokens:        maxTokens,
-		Temperature:      aiMode,
+		Temperature:      temperature,
 		TopP:             1,
 		FrequencyPenalty: 0,
 		PresencePenalty:  0,

code/services/openai/gpt3_test.go

@@ -2,9 +2,8 @@ package openai
 
 import (
 	"fmt"
-	"testing"
-
 	"start-feishubot/initialization"
+	"testing"
 )
 
 func TestCompletions(t *testing.T) {
@@ -17,7 +16,7 @@ func TestCompletions(t *testing.T) {
 
 	gpt := NewChatGPT(*config)
 
-	resp, err := gpt.Completions(msgs, Balance)
+	resp, err := gpt.Completions(msgs)
 	if err != nil {
 		t.Errorf("TestCompletions failed with error: %v", err)
 	}

code/services/sessionCache.go

@@ -1,6 +1,7 @@
 package services
 
 import (
+	"encoding/json"
 	"start-feishubot/services/openai"
 	"time"
 
@@ -20,7 +21,6 @@ type SessionMeta struct {
 	Mode       SessionMode       `json:"mode"`
 	Msg        []openai.Messages `json:"msg,omitempty"`
 	PicSetting PicSetting        `json:"pic_setting,omitempty"`
-	AIMode     openai.AIMode     `json:"ai_mode,omitempty"`
 }
 
 const (
@@ -35,14 +35,10 @@ const (
 )
 
 type SessionServiceCacheInterface interface {
-	Get(sessionId string) *SessionMeta
-	Set(sessionId string, sessionMeta *SessionMeta)
 	GetMsg(sessionId string) []openai.Messages
 	SetMsg(sessionId string, msg []openai.Messages)
 	SetMode(sessionId string, mode SessionMode)
 	GetMode(sessionId string) SessionMode
-	GetAIMode(sessionId string) openai.AIMode
-	SetAIMode(sessionId string, aiMode openai.AIMode)
 	SetPicResolution(sessionId string, resolution Resolution)
 	GetPicResolution(sessionId string) string
 	Clear(sessionId string)
@@ -50,22 +46,6 @@ type SessionServiceCacheInterface interface {
 
 var sessionServices *SessionService
 
-// implement Get interface
-func (s *SessionService) Get(sessionId string) *SessionMeta {
-	sessionContext, ok := s.cache.Get(sessionId)
-	if !ok {
-		return nil
-	}
-	sessionMeta := sessionContext.(*SessionMeta)
-	return sessionMeta
-}
-
-// implement Set interface
-func (s *SessionService) Set(sessionId string, sessionMeta *SessionMeta) {
-	maxCacheTime := time.Hour * 12
-	s.cache.Set(sessionId, sessionMeta, maxCacheTime)
-}
-
 func (s *SessionService) GetMode(sessionId string) SessionMode {
 	// Get the session mode from the cache.
 	sessionContext, ok := s.cache.Get(sessionId)
@@ -89,29 +69,6 @@ func (s *SessionService) SetMode(sessionId string, mode SessionMode) {
 	s.cache.Set(sessionId, sessionMeta, maxCacheTime)
 }
 
-func (s *SessionService) GetAIMode(sessionId string) openai.AIMode {
-	sessionContext, ok := s.cache.Get(sessionId)
-	if !ok {
-		return openai.Balance
-	}
-	sessionMeta := sessionContext.(*SessionMeta)
-	return sessionMeta.AIMode
-}
-
-// SetAIMode set the ai mode for the session.
-func (s *SessionService) SetAIMode(sessionId string, aiMode openai.AIMode) {
-	maxCacheTime := time.Hour * 12
-	sessionContext, ok := s.cache.Get(sessionId)
-	if !ok {
-		sessionMeta := &SessionMeta{AIMode: aiMode}
-		s.cache.Set(sessionId, sessionMeta, maxCacheTime)
-		return
-	}
-	sessionMeta := sessionContext.(*SessionMeta)
-	sessionMeta.AIMode = aiMode
-	s.cache.Set(sessionId, sessionMeta, maxCacheTime)
-}
-
 func (s *SessionService) GetMsg(sessionId string) (msg []openai.Messages) {
 	sessionContext, ok := s.cache.Get(sessionId)
 	if !ok {
@@ -189,7 +146,8 @@ func GetSessionCache() SessionServiceCacheInterface {
 func getStrPoolTotalLength(strPool []openai.Messages) int {
 	var total int
 	for _, v := range strPool {
-		total += v.CalculateTokenLength()
+		bytes, _ := json.Marshal(v)
+		total += len(string(bytes))
 	}
 	return total
 }

code/utils/commonUtils.go

@@ -0,0 +1,12 @@
+package utils
+
+import (
+	"time"
+)
+
+func GetCurrentDateAsString() string {
+	return time.Now().Format("2006-01-02")
+
+	// 本地测试可以用这个。将1天缩短到10秒。
+	//return strconv.Itoa((time.Now().Second() + 100000) / 10)
+}

code/utils/logUtils.go

@@ -0,0 +1,24 @@
+package utils
+
+import (
+	"fmt"
+	"gopkg.in/natefinch/lumberjack.v2"
+	"log"
+	"time"
+)
+
+type MyLogWriter struct {
+}
+
+func (writer MyLogWriter) Write(bytes []byte) (int, error) {
+	return fmt.Print(time.Now().UTC().Format("2006-01-02T15:04:05.999Z") + string(bytes))
+}
+
+func CloseLogger(logger *lumberjack.Logger) {
+	err := logger.Close()
+	if err != nil {
+		log.Println(err)
+	} else {
+		log.Println("logger closed")
+	}
+}

docker-compose.yaml

@@ -1,21 +1,26 @@
 version: '3.3'
 services:
   feishu-chatgpt:
-    container_name: feishu-chatgpt
+    container_name: Feishu-OpenAI-Stream-Chatbot
     build:
       context: .
       dockerfile: Dockerfile
     ports:
       - "9000:9000/tcp"
-    # volumes:
+    volumes:
       # - ./code/config.yaml:/app/config.yaml:ro
+      # 要注意，这里右边的容器内的路径，不是从根目录开始的，要参考 dockerfile 中的 WORKDIR
+      - ./logs:/app/logs
     environment:
+      ################ 以下配置建议和 config.example.yaml 里面的配置综合起来看 ################
+      # 日志配置, 默认不开启, 可以开启后查看日志
+      - ENABLE_LOG=false
       - APP_ID=cli_axxx
       - APP_SECRET=xxx
       - APP_ENCRYPT_KEY=xxx
       - APP_VERIFICATION_TOKEN=xxx
       # 请确保和飞书应用管理平台中的设置一致
-      - BOT_NAME=chatGpt
+      - BOT_NAME=xxx
       # OpenAI API Key 支持负载均衡, 可以填写多个 Key 用逗号分隔
       - OPENAI_KEY=sk-xxx,sk-xxx,sk-xxx
       # 服务器配置
@@ -25,6 +30,13 @@ services:
       - CERT_FILE=cert.pem
       - KEY_FILE=key.pem
       # OpenAI 地址, 一般不需要修改, 除非你有自己的反向代理
-      - API_URL=https://oapi.czl.net
+      - API_URL=https://api.openai.com
       # 代理设置, 例如 - HTTP_PROXY=http://127.0.0.1:7890, 默认代表不使用代理
       - HTTP_PROXY
+      ## 访问控制
+      # 是否启用访问控制。默认不启用。
+      - ACCESS_CONTROL_ENABLE=false
+      # 每个用户每天最多问多少个问题。默认为0. 配置成为小于等于0表示不限制。
+      - ACCESS_CONTROL_MAX_COUNT_PER_USER_PER_DAY=0
+      # 访问OpenAi的 普通 Http请求的超时时间，单位秒，不配置的话默认为 550 秒
+      - OPENAI_HTTP_CLIENT_TIMEOUT

s.yaml

@@ -28,7 +28,7 @@ services:
         name: "feishu-chatgpt"
         description: 'a simple feishubot by serverless devs'
         codeUri: './code'
-        caPort: 9000
+        cAPort: 9000
         customRuntimeConfig:
           command:
             - ./target/main