本章主要结介绍如何部署AIChat引擎本章主要介绍X86架构下如何部署AIChat引擎, 升级AIChat引擎请查看文档：AIChat引擎版本升级。
...
安装时用root用户运行或使用具有sudo权限的用户执行，推荐使用root用户安装，如果是普通用户有sudo权限的，在命令前加上sudo。
1.1.3 关闭selinux
临时关闭selinux，立即生效，不需要重启服务器。
| 代码块 |
|---|
setenforce 0 |
永久关闭selinux，修改完配置后需要重启服务器才能生效
| 代码块 |
|---|
sed -i 's/=enforcing/=disabled/g' /etc/selinux/config |
2.安装
(1) 解压安装包
| 代码块 | ||||
|---|---|---|---|---|
| ||||
cd /<安装包路径> tar -xvf nla-agent-setup.tar.gz |
...
输入创建的AIChat安装目录的绝对路径:
当出现如下字样则说明安装完成并已经启动：
检查aichat启动是否成功
| 代码块 | ||
|---|---|---|
| ||
docker ps # 列出当前正在运行的 Docker 容器
docker logs 容器的名称或 ID # 查看 Docker 容器的日志输出 |
如果操作系统已安装了高版本的docker，可能会碰到网络问题，导致启动失败。此时需要修改docker-compose.yml文件，添加network配置。
| 代码块 | ||||||||
|---|---|---|---|---|---|---|---|---|
| ||||||||
version: "3"
services:
propertycenter:
image: "aienv4j:v9.1"
volumes:
- /data/admin/smartbi-nla/logs:/ailogs
- /data/admin/smartbi-nla/services:/jars
- /data/admin/smartbi-nla/scripts/propertycenter:/scripts
links:
- "redis:airedis"
entrypoint:
- /bin/sh
- /scripts/docker-entrypoint.sh
depends_on:
- redis
restart: always
networks:
smartbinla_v1:
aiweb:
image: "aienv4j:v9.1"
ports:
- "9060:9082"
volumes:
- /data/admin/smartbi-nla/logs:/ailogs
- /data/admin/smartbi-nla/services:/jars
- /data/admin/smartbi-nla/scripts/aiweb:/scripts
links:
- "redis:airedis"
entrypoint:
- /bin/sh
- /scripts/docker-entrypoint.sh
depends_on:
- redis
restart: always
networks:
smartbinla_v1:
admin:
image: "aienv4j:v9.1"
volumes:
- /data/admin/smartbi-nla/logs:/ailogs
- /data/admin/smartbi-nla/services:/jars
- /data/admin/smartbi-nla/scripts/admin:/scripts
links:
- "redis:airedis"
- "propertycenter:propertycenter"
- "mongo:aimongo"
entrypoint:
- /bin/sh
- /scripts/docker-entrypoint.sh
depends_on:
- redis
- propertycenter
- mongo
restart: always
networks:
smartbinla_v1:
aliases:
- main
modulemanager:
image: "aienv4j:v9.1"
ports:
- "9081:9081"
volumes:
- /data/admin/smartbi-nla/logs:/ailogs
- /data/admin/smartbi-nla/services:/jars
- /data/admin/smartbi-nla/scripts/modulemanager:/scripts
links:
- "redis:airedis"
- "propertycenter:propertycenter"
entrypoint:
- /bin/sh
- /scripts/docker-entrypoint.sh
depends_on:
- redis
- propertycenter
restart: always
networks:
smartbinla_v1:
aibus:
image: "aienv4j:v9.1"
volumes:
- /data/admin/smartbi-nla/logs:/ailogs
- /data/admin/smartbi-nla/services:/jars
- /data/admin/smartbi-nla/scripts/aibus:/scripts
links:
- "redis:airedis"
- "propertycenter:propertycenter"
entrypoint:
- /bin/sh
- /scripts/docker-entrypoint.sh
depends_on:
- redis
- propertycenter
restart: always
networks:
smartbinla_v1:
smartbiproxy:
image: "aienv4j:v9.1"
volumes:
- /data/admin/smartbi-nla/logs:/ailogs
- /data/admin/smartbi-nla/services:/jars
- /data/admin/smartbi-nla/scripts/smartbiproxy:/scripts
links:
- "redis:airedis"
- "propertycenter:propertycenter"
entrypoint:
- /bin/sh
- /scripts/docker-entrypoint.sh
depends_on:
- redis
- propertycenter
restart: always
networks:
smartbinla_v1:
nl2sql_llm:
image: "aienv4py-llm:v6"
volumes:
- /data/admin/smartbi-nla/logs:/logs
- /data/admin/smartbi-nla/scripts/nl2sql_llm:/scripts
- /data/admin/smartbi-nla/services:/services
- /data/admin/smartbi-nla/data/nl2sql_data:/data
- /data/admin/smartbi-nla/backup/backup_data:/backup_data
links:
- "redis:airedis"
- "admin:main"
depends_on:
- redis
entrypoint:
- /bin/sh
- /scripts/docker-entrypoint.sh
restart: always
networks:
smartbinla_v1:
jupyter:
image: "aienv4py-agent:v1.1"
volumes:
- /data/admin/smartbi-nla/logs:/logs
- /data/admin/smartbi-nla/scripts/agent:/scripts
- /data/admin/smartbi-nla/services:/services
links:
- "redis:airedis"
- "admin:main"
depends_on:
- redis
entrypoint:
- /bin/sh
- /scripts/docker-entrypoint-runtime.sh
restart: always
networks:
smartbinla_v1:
agent:
image: "aienv4py-agent:v1.1"
volumes:
- /data/admin/smartbi-nla/logs:/logs
- /data/admin/smartbi-nla/scripts/agent:/scripts
- /data/admin/smartbi-nla/services:/services
- /data/admin/smartbi-nla/data/nl2sql_data:/data
- /var/run/docker.sock:/var/run/docker.sock
links:
- "redis:airedis"
- "admin:main"
- "jupyter:jupyter"
depends_on:
- redis
- jupyter
entrypoint:
- /bin/sh
- /scripts/docker-entrypoint.sh
restart: always
networks:
smartbinla_v1:
mongo:
image: "mongo:4"
volumes:
- /data/admin/smartbi-nla/data/mongo/data:/data/db
environment:
MONGO_INITDB_ROOT_USERNAME: root
MONGO_INITDB_ROOT_PASSWORD: smartbi1#nla
restart: always
networks:
smartbinla_v1:
aliases:
- aimongo
redis:
image: "airedis:v5"
volumes:
- /data/admin/smartbi-nla/data/config:/var/lib/mysql
- /data/admin/smartbi-nla/backup/config:/backup
restart: always
networks:
smartbinla_v1:
aliases:
- airedis
networks:
smartbinla_v1:
driver: bridge
ipam:
config:
- subnet: 192.168.1.0/24 |
如果需要停止AIChat，可以执行命令
| 代码块 | ||
|---|---|---|
| ||
cd /data/admin/smartbi-nla # AIChat安装目录 sudo bash stop.sh |
...
打开AIChat管理后台，配置smartbi服务器。地址：http://ip:9060/aiweb/#/manager/config 默认密码：admin/manager
配置smartbi服务器地址。
样例：http://10.10.202.101:18080/smartbi
(4) 配置大语言模型
进入“系统运维--自然语言配置“中设置“大模型配置”，新增大模型。新增大模型。参考：连接大模型
支持OpenAI接口，阿里千问、百度等大模型。
配置在线大模型前，要先获取大模型的api_key。
...
