This is an automated email from the ASF dual-hosted git repository.
alexstocks pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/dubbo-go-samples.git
The following commit(s) were added to refs/heads/main by this push:
new a150316c bug fix and rewirte image encode and decode logic (#807)
a150316c is described below
commit a150316c7837b1247070e050466af94cfcccd576
Author: Alan <[email protected]>
AuthorDate: Mon Mar 10 23:59:53 2025 +0800
bug fix and rewirte image encode and decode logic (#807)
* fix import
* fix readme dir
* rewrite image encode and decode logic to support image files robustness
and fix out of bounds err
* fix typo
* limit upload file types and add clarification in the readme
* limit upload file types and add clarification in the readme
* Revert "limit upload file types and add clarification in the readme"
This reverts commit 8c514464208b43e982258724fabd58d09bf2bcb6.
* limit upload file types and add clarification in the readme
---
llm/README.md | 7 ++++---
llm/README_zh.md | 10 +++++-----
llm/go-client/frontend/handlers/chat.go | 18 +++++++++++++++---
llm/go-client/frontend/static/script.js | 3 +++
llm/go-client/frontend/templates/index.html | 2 +-
llm/go-server/cmd/server.go | 12 +++++++-----
6 files changed, 35 insertions(+), 17 deletions(-)
diff --git a/llm/README.md b/llm/README.md
index ed909b81..a7472907 100644
--- a/llm/README.md
+++ b/llm/README.md
@@ -48,7 +48,7 @@ The server integrates the Ollama model and uses Dubbo-go's
RPC service for invoc
Run the server by executing:
```shell
-$ go run llm/go-server/cmd/server.go
+$ go run go-server/cmd/server.go
```
### **Run the Client**
@@ -58,7 +58,7 @@ The client invokes the server's RPC interface to retrieve the
inference results
Run the cli client by executing:
```shell
-$ go run llm/go-client/cmd/client.go
+$ go run go-client/cmd/client.go
```
Cli client supports multi-turn conversations, command interact, context
management.
@@ -66,7 +66,8 @@ Cli client supports multi-turn conversations, command
interact, context manageme
We also support a frontend using Gin framework for users to interact. If you
want run the frontend client you can executing the following command and open
it in ```localhost:8080``` by default:
```shell
-$ go run llm/go-client/frontend/main.go
+$ go run go-client/frontend/main.go
```
Frontend client supports multi-turn conversations, binary file (image) support
for LLM interactions.
+Currently the supported uploaded image types are limited to png, jpeg and gif,
with plans to support more binary file types in the future.
diff --git a/llm/README_zh.md b/llm/README_zh.md
index 7f81488f..3000c1fc 100644
--- a/llm/README_zh.md
+++ b/llm/README_zh.md
@@ -48,7 +48,7 @@ $ cd llm
在服务端目录下运行：
```shell
-$ go run llm/go-server/cmd/server.go
+$ go run go-server/cmd/server.go
```
### **客户端运行**
@@ -58,7 +58,7 @@ $ go run llm/go-server/cmd/server.go
在客户端目录下运行：
```shell
-$ go run llm/go-client/cmd/client.go
+$ go run go-client/cmd/client.go
```
命令行客户端支持多轮对话、命令交互、上下文管理功能。
@@ -66,8 +66,8 @@ $ go run llm/go-client/cmd/client.go
我们也提供了包含前端页面的基于Gin框架的客户端进行交互，运行以下命令然后访问 ```localhost:8080``` 即可使用:
```shell
-$ go run llm/go-client/frontend/main.go
+$ go run go-client/frontend/main.go
```
-包含前端页面的客户端支持多轮对话，支持进行二进制文件（图片）传输并与大模型进行交互.
-
+包含前端页面的客户端支持多轮对话，支持进行二进制文件（图片）传输并与大模型进行交互。
+目前所支持上传的图片类型被限制为 png，jpeg 和 gif 类型，计划在将来支持更多的二进制文件类型。
\ No newline at end of file
diff --git a/llm/go-client/frontend/handlers/chat.go
b/llm/go-client/frontend/handlers/chat.go
index c65ab5d9..e16197e2 100644
--- a/llm/go-client/frontend/handlers/chat.go
+++ b/llm/go-client/frontend/handlers/chat.go
@@ -22,8 +22,8 @@ import (
"io"
"log"
"net/http"
+ "regexp"
"runtime/debug"
- "strings"
"time"
)
@@ -82,13 +82,25 @@ func (h *ChatHandler) Chat(c *gin.Context) {
return
}
- sepIndex := strings.Index(",", req.Bin)
+ var img string
+ if len(req.Bin) > 0 {
+ re :=
regexp.MustCompile(`^data:image/([a-zA-Z]+);base64,([^"]+)$`)
+ // this regex does not support file types like svg
+ matches := re.FindStringSubmatch(req.Bin)
+
+ if len(matches) != 3 {
+ c.JSON(http.StatusBadRequest, gin.H{"error": "invalid
base64 data format"})
+ return
+ }
+
+ img = matches[2]
+ }
messages := h.ctxManager.GetHistory(ctxID)
messages = append(messages, &chat.ChatMessage{
Role: "human",
Content: req.Message,
- Bin: []byte(req.Bin[sepIndex+1:]),
+ Bin: []byte(img),
})
stream, err := h.svc.Chat(context.Background(), &chat.ChatRequest{
diff --git a/llm/go-client/frontend/static/script.js
b/llm/go-client/frontend/static/script.js
index d09d1517..3b56d2cc 100644
--- a/llm/go-client/frontend/static/script.js
+++ b/llm/go-client/frontend/static/script.js
@@ -23,6 +23,7 @@ const createChatLi = (message, className) => {
const handleChat = () => {
userMessage = chatInput.value.trim(); // Get user entered message and
remove extra whitespace
+ userBin = null
if (fileBlobArr.length > 0) {
userBin = fileBlobArr[0]
}
@@ -204,6 +205,8 @@ function clear() {
document.getElementById("drop").style.display = "none";
addBtn.style.display = "flex";
fileInput.value = "";
+ fileBlobArr = [];
+ fileArr = [];
}
document.getElementById("drop").addEventListener('click', clear);
\ No newline at end of file
diff --git a/llm/go-client/frontend/templates/index.html
b/llm/go-client/frontend/templates/index.html
index 68eefdc5..35aef33a 100644
--- a/llm/go-client/frontend/templates/index.html
+++ b/llm/go-client/frontend/templates/index.html
@@ -41,7 +41,7 @@
</div>
- <input id="input" accept="image/*" type="file" style="display: none" >
+ <input id="input" accept="image/png, image/jpeg, image/gif" type="file"
style="display: none" >
</body>
diff --git a/llm/go-server/cmd/server.go b/llm/go-server/cmd/server.go
index c4e752fb..8f2b88ec 100644
--- a/llm/go-server/cmd/server.go
+++ b/llm/go-server/cmd/server.go
@@ -19,8 +19,10 @@ package main
import (
"context"
+ "encoding/base64"
"fmt"
"log"
+ "net/http"
"runtime/debug"
)
@@ -81,13 +83,13 @@ func (s *ChatServer) Chat(ctx context.Context, req
*chat.ChatRequest, stream cha
}
if msg.Bin != nil && len(msg.Bin) != 0 {
- img := string(msg.Bin)
+ decodeByte, err :=
base64.StdEncoding.DecodeString(string(msg.Bin))
if err != nil {
- log.Println("Decode image error:", err)
- return fmt.Errorf("decode image error: %s", err)
+ log.Println("GenerateContent failed: %v", err)
+ return fmt.Errorf("GenerateContent failed")
}
-
- messageContent.Parts = append(messageContent.Parts,
llms.BinaryPart("image/png", img))
+ imgType := http.DetectContentType(decodeByte)
+ messageContent.Parts = append(messageContent.Parts,
llms.BinaryPart(imgType, decodeByte))
}
messages = append(messages, messageContent)
