openai-experimental[patch]: Adds experimental raw response field to OpenAI chat models #6087
Conversation
|
The latest updates on your projects. Learn more about Vercel for Git ↗︎
|
dosubot
bot
added
size:L
![dosubot[bot]](https://avatars.githubusercontent.com/in/324583?s=60&v=4){kind=small}
| @@ -28,6 +28,24 @@ import { AzureChatOpenAI } from "../../azure/chat_models.js"; | |||
| // Save the original value of the 'LANGCHAIN_CALLBACKS_BACKGROUND' environment variable | |||
There was a problem hiding this comment.
Hey team, I've flagged the added beforeAll block in the PR for review as it explicitly sets environment variables using process.env. This change should be reviewed to ensure it aligns with our environment variable management practices.
| @@ -5,6 +5,24 @@ import { AIMessageChunk } from "@langchain/core/messages"; | |||
| import { AzureChatOpenAI } from "../../azure/chat_models.js"; | |||
There was a problem hiding this comment.
Hey team, I've flagged the recent change in the PR for review as it explicitly accesses and sets environment variables using process.env. Please take a look and ensure it aligns with our best practices for handling environment variables. Thanks!
| @@ -1,6 +1,24 @@ | |||
| import { test, expect } from "@jest/globals"; | |||
There was a problem hiding this comment.
Hey there! I noticed that the recent PR includes changes to access and set environment variables using process.env. This comment is to flag the change for maintainers to review the handling of environment variables in the code. Great work on the PR!
| @@ -15,6 +15,24 @@ import { AzureOpenAI } from "../../azure/llms.js"; | |||
| // Save the original value of the 'LANGCHAIN_CALLBACKS_BACKGROUND' environment variable | |||
There was a problem hiding this comment.
Hey team, just a heads up that I've flagged the recent changes in the llms.int.test.ts file for review. The added beforeAll block sets environment variables based on corresponding TEST_ prefixed variables, so it's important to ensure everything is handled correctly. Let me know if you have any questions!
…jacob/experimental_raw_response
Passing
__includeRawResponse: trueon instantiation will return the raw response from OpenAI inadditional_kwargs.It's not (and probably won't ever be) exactly 1:1 with OpenAI - we add an extra metadata chunk at the end rather than yielding when OpenAI emits it. The final result of chunk concatenation also doesn't make sense, since it doesn't make sense to merge multiple OpenAI streaming responses.