Test settings #124
Test settings #124
Conversation
| "apiKey": { | ||
| "type": "string", | ||
| "description": "Anthropic API key" | ||
| }, | ||
| "anthropicApiUrl": { |
There was a problem hiding this comment.
Should probably be kept, the attribute doesn't have the same name https://github.com/langchain-ai/langchainjs/blob/c52cc0d53473e8507382c575cba5d46e107783ff/libs/langchain-anthropic/src/chat_models.ts#L701
| @@ -2,10 +2,6 @@ | |||
| "$schema": "http://json-schema.org/draft-07/schema#", | |||
| "type": "object", | |||
| "properties": { | |||
| "concurrency": { | |||
There was a problem hiding this comment.
Not sure about this one.
| @@ -38,10 +38,6 @@ | |||
| "description": "Model name to use (e.g., gemini-pro, gemini-2.0-flash, etc.)", | |||
| "default": "gemini-pro" | |||
| }, | |||
| "baseURL": { | |||
There was a problem hiding this comment.
It doesn't set an attribute of the object, but may be useful https://github.com/langchain-ai/langchainjs/blob/c52cc0d53473e8507382c575cba5d46e107783ff/libs/langchain-google-genai/src/chat_models.ts#L707
| @@ -93,12 +93,6 @@ | |||
| "type": ["string", "number"], | |||
| "default": "5m" | |||
| }, | |||
| "stop": { | |||
There was a problem hiding this comment.
Should be nice to keep it.
Short explanation about stop words (I didn't know what it was used for): https://datasciencedude.com/how-to-easily-get-stop-words-for-your-language-learning-model#heading-what-are-stop-words-and-why-are-they-important
| @@ -113,11 +107,6 @@ | |||
| "description": "The host URL of the Ollama server.", | |||
| "default": "" | |||
| }, | |||
| "headers": { | |||
There was a problem hiding this comment.
| @@ -15,461 +15,6 @@ | |||
| "type": "number", | |||
| "description": "An integer between 0 and 5 specifying the number of most likely tokens to return at each token position, each with an associated log probability. logprobs must be set to true if this parameter is used." | |||
| }, | |||
| "prefixMessages": { | |||
There was a problem hiding this comment.
Add a prefix to the prompt. Since we have a configurable prompt, I guess we should not need it.
This PR should fix #114.
It adds jest test, and uses the fact that "most" of the settings set an attribute in the chat model object.