Fast GPT 允许你使用自己的 openai API KEY 来快速的调用 openai 接口,目前集成了 gpt35 和 embedding. 可构建自己的知识库。
配置环境变量
# proxy(可选)
AXIOS_PROXY_HOST=127.0.0.1
AXIOS_PROXY_PORT=7890
# openai 中转连接(可选)
OPENAI_BASE_URL=https://api.openai.com/v1
OPENAI_BASE_URL_AUTH=可选的安全凭证
# 是否开启队列任务。 1-开启,0-关闭(请求 parentUrl 去执行任务,单机时直接填1)
queueTask=1
parentUrl=https://hostname/api/openapi/startEvents
# 发送邮箱验证码配置。参考 nodeMail 获取参数,自行百度。
[email protected]
MAILE_CODE=xxx
# 阿里短信服务(邮箱和短信至少二选一)
aliAccessKeyId=xxx
aliAccessKeySecret=xxx
aliSignName=xxx
aliTemplateCode=SMS_xxx
# token(随便填,作为登录凭证)
TOKEN_KEY=xxx
# openai key
OPENAIKEY=sk-xxx
# mongo连接地址
MONGODB_URI=mongodb:https://username:[email protected]:27017/test?authSource=admin
# mongo数据库名称
MONGODB_NAME=xxx
# pg 数据库相关内容,和 docker-compose 对上
PG_HOST=0.0.0.0
PG_PORT=8102
PG_USER=xxx
PG_PASSWORD=xxx
PG_DB_NAME=xxx
运行
pnpm dev
- clash 方案 - 仅需一台服务器(需要有 clash)
- nginx 方案 - 需要一台国外服务器
- cloudflare 方案 - 需要有域名(每日免费 10w 次代理请求)
这个不同系统略有区别,百度安装下。验证安装成功后进行下一步。下面给出一个例子:
# 安装docker
curl -L https://get.daocloud.io/docker | sh
sudo systemctl start docker
# 安装 docker-compose
curl -L https://github.com/docker/compose/releases/download/1.23.2/docker-compose-`uname -s`-`uname -m` -o /usr/local/bin/docker-compose
sudo chmod +x /usr/local/bin/docker-compose
# 验证安装
docker -v
docker-compose -v
手动创建或者直接把 deploy 里内容复制过去
/root/fast-gpt/pg/init.sql
set -e
psql -v ON_ERROR_STOP=1 --username "$POSTGRES_USER" --dbname "$POSTGRES_DB" <<-EOSQL
CREATE EXTENSION vector;
-- init table
CREATE TABLE modelData (
id BIGSERIAL PRIMARY KEY,
vector VECTOR(1536),
status VARCHAR(50) NOT NULL,
user_id VARCHAR(50) NOT NULL,
model_id VARCHAR(50) NOT NULL,
q TEXT NOT NULL,
a TEXT NOT NULL
);
-- create index
CREATE INDEX modelData_status_index ON modelData USING HASH (status);
CREATE INDEX modelData_userId_index ON modelData USING HASH (user_id);
CREATE INDEX modelData_modelId_index ON modelData USING HASH (model_id);
EOSQL
/root/fast-gpt/nginx/nginx.conf
user nginx;
worker_processes auto;
worker_rlimit_nofile 51200;
events {
worker_connections 1024;
}
http {
access_log off;
server_names_hash_bucket_size 512;
client_header_buffer_size 32k;
large_client_header_buffers 4 32k;
client_max_body_size 50M;
gzip on;
gzip_min_length 1k;
gzip_buffers 4 8k;
gzip_http_version 1.1;
gzip_comp_level 6;
gzip_vary on;
gzip_types text/plain application/x-javascript text/css application/javascript application/json application/xml;
gzip_disable "MSIE [1-6]\.";
open_file_cache max=1000 inactive=1d;
open_file_cache_valid 30s;
open_file_cache_min_uses 8;
open_file_cache_errors off;
server {
listen 443 ssl;
server_name docgpt.ahapocket.cn;
ssl_certificate /ssl/docgpt.pem;
ssl_certificate_key /ssl/docgpt.key;
ssl_session_timeout 5m;
location / {
proxy_pass https://localhost:3000;
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
}
}
server {
listen 80;
server_name docgpt.ahapocket.cn;
rewrite ^(.*) https://$server_name$1 permanent;
}
}
/root/fast-gpt/docker-compose.yml
version: "3.3"
services:
fast-gpt:
image: c121914yu/fast-gpt:latest
network_mode: host
restart: always
container_name: fast-gpt
environment:
# - AXIOS_PROXY_HOST=127.0.0.1
# - AXIOS_PROXY_PORT=7890
# - OPENAI_BASE_URL=https://api.openai.com/v1
# - OPENAI_BASE_URL_AUTH=可选的安全凭证
- [email protected]
- MAILE_CODE=xxxx
- aliAccessKeyId=xxxx
- aliAccessKeySecret=xxxx
- aliSignName=xxxxx
- aliTemplateCode=SMS_xxxx
- TOKEN_KEY=xxxx
- queueTask=1
- parentUrl=https://hostname/api/openapi/startEvents
- MONGODB_URI=mongodb:https://username:[email protected]:27017/?authSource=admin
- MONGODB_NAME=xxx
- PG_HOST=0.0.0.0
- PG_PORT=8100
- PG_USER=xxx
- PG_PASSWORD=xxx
- PG_DB_NAME=xxx
- OPENAIKEY=sk-xxxxx
nginx:
image: nginx:alpine3.17
container_name: nginx
restart: always
network_mode: host
volumes:
- /root/fast-gpt/nginx/nginx.conf:/etc/nginx/nginx.conf:ro
- /root/fast-gpt/nginx/logs:/var/log/nginx
- /root/fast-gpt/nginx/ssl/docgpt.key:/ssl/docgpt.key
- /root/fast-gpt/nginx/ssl/docgpt.pem:/ssl/docgpt.pem
pg:
image: ankane/pgvector
container_name: pg
restart: always
ports:
- 8100:5432
environment:
- POSTGRES_USER=xxx
- POSTGRES_PASSWORD=xxx
- POSTGRES_DB=xxx
volumes:
- /root/fast-gpt/pg/data:/var/lib/postgresql/data
- /root/fast-gpt/pg/init.sql:/docker-entrypoint-initdb.d/init.sh
- /etc/localtime:/etc/localtime:ro
mongodb:
image: mongo:4.0.1
container_name: mongo
restart: always
ports:
- 27017:27017
environment:
- MONGO_INITDB_ROOT_USERNAME=username
- MONGO_INITDB_ROOT_PASSWORD=password
volumes:
- /root/fast-gpt/mongo/data:/data/db
- /root/fast-gpt/mongo/logs:/var/log/mongodb
- /etc/localtime:/etc/localtime:ro
下面是一个辅助脚本,也可以直接 docker-compose up -d run.sh 运行文件
#!/bin/bash
docker-compose pull
docker-compose up -d
echo "Docker Compose 重新拉取镜像完成!"
# 删除本地旧镜像
images=$(docker images --format "{{.ID}} {{.Repository}}" | grep fast-gpt)
# 将镜像 ID 和名称放入数组中
IFS=$'\n' read -rd '' -a image_array <<<"$images"
# 遍历数组并删除所有旧的镜像
for ((i=1; i<${#image_array[@]}; i++))
do
image=${image_array[$i]}
image_id=${image%% *}
docker rmi $image_id
done
.github里拥有一个 git 提交到 main 分支时自动打包 amd64 和 arm64 镜像的 actions。你仅需要提前在 git 配置好 session。
- 创建账号 session: 头像 -> settings -> 最底部 Developer settings -> Personal access tokens -> tokens(classic) -> 创建新 session,把一些看起来需要的权限勾上。
- 添加 session 到仓库: 仓库 -> settings -> Secrets and variables -> Actions -> 创建secret
- 填写 secret: Name-GH_PAT, Secret-第一步的tokens
因为教程有部分镜像不兼容arm64,所以写个文档指导新手如何快速在mac上面搭建fast-gpt如何在mac上面部署fastgpt