Merge pull request kubeagi#482 from nkwangleiGIT/main

Add ray build file and update readme

.github/workflows/codespell.yaml

@@ -15,5 +15,5 @@ jobs:
         with:
           ignore_words_file: .github/.codespellignore
           check_filenames: true
-          skip: go.*,**/*.drawio,./deploy/charts/*,./config/crd/*
+          skip: go.*,**/*.drawio,./deploy/charts/*,./config/crd/*,./deploy/llms/*
 

README.md

@@ -45,6 +45,8 @@ If you don't have a kubernetes cluster, you can schedule a [kind cluster](https:
 
 Visit our [online documents](http://kubeagi.k8s.com.cn/docs/intro)
 
+Read [user guide](http://kubeagi.k8s.com.cn/docs/UserGuide/intro)
+
 ## CLI
 
 We provide a Command Line Tool `arctl` to interact with `arcadia`. See [here](./pkg/arctl/README.md) for more details.

apiserver/service/chat.go

@@ -35,6 +35,11 @@ import (
 	"github.com/kubeagi/arcadia/apiserver/pkg/requestid"
 )
 
+const (
+	// Time interval to check if the chat stream should be closed if no more message arrives
+	WaitTimeoutForChatStreaming = 5
+)
+
 // @BasePath /chat
 
 // @Summary chat with application
@@ -101,14 +106,14 @@ func chatHandler() gin.HandlerFunc {
 			// Use a ticker to check if there is no data arrived and close the stream
 			// TODO: check if any better solution for this?
 			hasData := true
-			ticker := time.NewTicker(5 * time.Second)
+			ticker := time.NewTicker(WaitTimeoutForChatStreaming * time.Second)
 			quit := make(chan struct{})
 			defer close(quit)
 			go func() {
 				for {
 					select {
 					case <-ticker.C:
-						// If there is no generated data within 5 seconds, just close it
+						// If there is no generated data within WaitTimeoutForChatStreaming seconds, just close it
 						if !hasData {
 							close(respStream)
 						}

deploy/charts/arcadia/Chart.yaml

@@ -2,8 +2,8 @@ apiVersion: v2
 name: arcadia
 description: A Helm chart(KubeBB Component) for KubeAGI Arcadia
 type: application
-version: 0.1.56
-appVersion: "0.0.1"
+version: 0.2.1
+appVersion: "0.1.0"
 
 keywords:
   - LLMOps

deploy/charts/arcadia/values.yaml

@@ -7,7 +7,7 @@ global:
 # @param imagePullPolcy ImagePullPolicy
 # @param resources Resources to be used
 controller:
-  image: kubeagi/arcadia:v0.1.0-20231219-874e2b1
+  image: kubeagi/arcadia:v0.1.0-20231230-c8518fb
   imagePullPolicy: IfNotPresent
   resources:
     limits:
@@ -20,7 +20,7 @@ controller:
 # @section graphql and bff server
 # related project: https://github.com/kubeagi/arcadia/tree/main/apiserver
 apiserver:
-  image: kubeagi/arcadia:v0.1.0-20231219-874e2b1
+  image: kubeagi/arcadia:v0.1.0-20231230-c8518fb
   enableplayground: false
   port: 8081
   ingress:
@@ -38,7 +38,7 @@ apiserver:
 portal:
   enabled: true
   kubebbEnabled: true
-  image: kubeagi/portal:v0.1.0-20231216-df45bd2
+  image: kubeagi/portal:v0.1.0-20231229-f7b210e
   port: 80
   ingress:
     path: kubeagi-portal-public
@@ -48,7 +48,7 @@ portal:
 fastchat:
   image:
     repository: kubeagi/arcadia-fastchat
-    # Use either v0.1.0 or vllm-v0.1.0(with vllm enabled)
+    # Use either v0.1.0 or vllm-v0.1.0 (with vllm enabled)
     tag: v0.1.0
   ingress:
     enabled: true
@@ -113,7 +113,7 @@ chromadb:
 # Related project: https://github.com/kubeagi/arcadia/tree/main/data-processing
 dataprocess:
   enabled: true
-  image: kubeagi/data-processing:v0.1.0-20231219-874e2b1
+  image: kubeagi/data-processing:v0.1.0-20231230-c8518fb
   port: 28888
   config:
     llm:

deploy/llms/Dockerfile.fastchat-worker-ray

@@ -0,0 +1,8 @@
+ARG BASE_IMAGE_VERSION="vllm-v0.1.0"
+FROM kubeagi/arcadia-fastchat-worker:${BASE_IMAGE_VERSION}
+
+# Allow to use environment variable to set ray & python version to pass the version check
+# for now, ray: 2.9.0, python: 3.9.x
+# this utils.py file is from ray 2.9.0 ray-ml image
+# search 'KubeAGI' in utils.py for what's changed
+COPY utils.py /usr/local/lib/python3.9/dist-packages/ray/_private/utils.py