Add docker compose and update readme #288
Conversation
☂️ Python Coverage
Overall Coverage
New FilesNo new covered files... Modified FilesNo covered modified files...
|
moria97
changed the title
moria97
requested review from
wwxxzz,
paradiseHIT,
Ceceliachenen and
aero-xi
| environment: | ||
| DASHSCOPE_API_KEY: ${DASHSCOPE_API_KEY} | ||
| OSS_ACCESS_KEY_ID: ${OSS_ACCESS_KEY_ID} | ||
| OSS_ACCESS_KEY_SECRET: ${OSS_ACCESS_KEY_SECRET} |
There was a problem hiding this comment.
什么时候用环境变量，什么时候用配置还没特别理解，比如oss key设置了，bucket还要去控制台再设置一次，用户感觉是都是oss的配置，还要两个地方配置
我也没特别想好，可以讨论下
There was a problem hiding this comment.
包括用docker compose还是用helm，好像更近一步是不是更建议用helm？可以下一步计划里面
There was a problem hiding this comment.
环境变量这个确实和控制台有重合。目前env不是required，感觉可以在这里写一下设置了环境变量，就不用在控制台输入这些key了。如果没设置，后面也可以在控制台手动输入。
docker compose主要是本地启动，helm确实更适合部署的时候，后面可以加一个helm的配置文件。
| ```bash | ||
| # Support custom host (default 0.0.0.0), port (default 8001), config (default src/pai_rag/config/settings.yaml), skip-download-models (default False) | ||
| # Download [bge-large-zh-v1.5, easyocr] by default, you can skip it by setting --skip-download-models. | ||
| # you can use tool "load_model" to download other models including [bge-large-zh-v1.5, easyocr, SGPT-125M-weightedmean-nli-bitfit, bge-large-zh-v1.5, bge-m3, bge-reranker-base, bge-reranker-large, paraphrase-multilingual-MiniLM-L12-v2, qwen_1.8b, text2vec-large-chinese] |
There was a problem hiding this comment.
如果已经load_model了，或者之前已经下载过了；是否会自动skip
There was a problem hiding this comment.
如果已经存在的模型，就会跳过，只下载本地没有的
| pai_rag ui [--host HOST] [--port PORT] [rag-url RAG_URL] | ||
| ``` | ||
|
|
||
| You can also open http://127.0.0.1:8002/ to configure the RAG service and upload local data. |
There was a problem hiding this comment.
有的地方写了localhost，有的写了127.0.0.1
| @@ -103,6 +103,10 @@ def index_url(self): | |||
| def list_index_url(self): | |||
| return f"{self.endpoint}v1/indexes" | |||
|
|
|||
| @property | |||
| def health_check_url(self): | |||
| return f"{self.endpoint}health" | |||
There was a problem hiding this comment.
用urljoin代码会更健壮些，万一endpoint末尾没有以/结尾
| - ./app_data:/app/localdata | ||
| entrypoint: ["pai_rag", "serve"] | ||
| healthcheck: | ||
| test: ["CMD", "curl", "-f", "http://localhost:8001/health"] |
There was a problem hiding this comment.
所有的api，是不是加个版本号
/api/v1/health
| - upload_data | ||
|
|
||
| ```bash | ||
| curl -X 'POST' http://127.0.0.1:8000/service/upload_data -H 'Content-Type: multipart/form-data' -F 'files=@local_path/PAI.txt' |
There was a problem hiding this comment.
API路径，一般都是
这么来设计的，这个改动看大不大，大的话，优先级可以调整下
|
|
||
| ```bash | ||
| # streaming output | ||
| curl -X 'POST' http://127.0.0.1:8000/service/query -H "Content-Type: application/json" -d '{"question":"PAI是什么？", "stream":true}' |
There was a problem hiding this comment.
英文的示例里面，question用英文的好点，而且示例可以用通用点的，防止用户没用这种数据，出来的结果感觉不好
| 使用DashScope API，需要在命令行引入环境变量 | ||
|
|
||
| ```bash | ||
| export DASHSCOPE_API_KEY="" |
There was a problem hiding this comment.
export DASHSCOPE_API_KEY="xxx"
下面补充下请替换xxx为你自己的DASHSCOPE_API_KEY，DASHSCOPE_API_KEY获取地址为 https://dashscope.console.aliyun.com/apiKey
No description provided.