Dévin.YF

> I like the prompt adjustment but I'm less sure about the quote stripping -- putting this in prevents callers specifying exact-search style queries, right? Yeah, to simulate this problem,...

> I think it's more appropriate for the caller to simply do a strings.Trim() call head of calling search. Yes, that’s more preferable, I merge latest main branch and updated...

update: found that the old implementation in `for xxx := range handler.egress` will never end... And will cause goroutine leak fixed this problem using `HandleAgentFinish` method ```go go func() {...

I followed this example to use memory, it works for me https://github.com/tmc/langchaingo/blob/main/examples/llm-chain-example/llm_chain.go ```go llm, err := openai.New() memory := memory.NewConversationBuffer() llmChain := chains.NewConversation(llm, memory) out, err := chains.Run(ctx, llmChain, "my...

> Does it works for you with latest version? I didn’t encountered any problem using memory in a chain conversion, This is my test code: ```go llm, _ := ollama.New(ollama.WithModel("phi"))...

Just wondering... Is it a good idea to use go-plugins to manage third-party modules like llms, tools ??

官方已经支持了 OpenAI 的兼容接口: `github.com/openai/openai-go` 推荐使用该方案调用api...具体请查阅阿里云百炼平台的文档 出现这个问题是因为你配置的参数: `ResultFormat: "text"`, 官方文档提到过推荐将这个值设置为`"message"` 所以我没有做过对 ‘text’的适配... 以下摘录自官方文档对这个字段的解释: >result_format string （可选） 默认为"text"（[QwQ](https://help.aliyun.com/zh/model-studio/deep-thinking) 模型、Qwen3 开源模型与 Qwen-Long 模型默认值为 "message"） 返回数据的格式。推荐您优先设置为"message"，可以更方便地进行[多轮对话](https://help.aliyun.com/zh/model-studio/multi-round-conversation)。 平台后续将统一将默认值调整为"message"。 如果想自己配制参数 建议使用如下方式: ```go params := qwen.DefaultParameters() params.SetMaxTokens(1000)...