GigaChat LLM support #917
GigaChat LLM support #917
Conversation
# Conflicts: # Gemfile.lock
There was a problem hiding this comment.
Copilot reviewed 9 out of 12 changed files in this pull request and generated 2 comments.
Files not reviewed (3)
- spec/fixtures/llm/gigachat/chat.json: Language not supported
- spec/fixtures/llm/gigachat/chat_chunk.json: Language not supported
- spec/fixtures/llm/gigachat/chat_with_function_call.json: Language not supported
| ) | ||
| raise ArgumentError.new("text argument is required") if text.empty? | ||
| raise ArgumentError.new("model argument is required") if model.empty? | ||
| raise ArgumentError.new("encoding_format must be either float or base64") if encoding_format && %w[float base64].include?(encoding_format) |
There was a problem hiding this comment.
The condition for raising an ArgumentError for encoding_format appears reversed; it should instead raise an error if encoding_format is provided and is not one of 'float' or 'base64'. For example, consider changing the condition to: if encoding_format && !%w[float base64].include?(encoding_format)
| raise ArgumentError.new("encoding_format must be either float or base64") if encoding_format && %w[float base64].include?(encoding_format) | |
| raise ArgumentError.new("encoding_format must be either float or base64") if encoding_format && !%w[float base64].include?(encoding_format) |
Copilot uses AI. Check for mistakes.
| end | ||
|
|
||
| def default_dimensions | ||
| @defaults[:dimensions] || EMBEDDING_SIZES.fetch(defaults[:embedding_model]) |
There was a problem hiding this comment.
default_dimensions uses EMBEDDING_SIZES.fetch(defaults[:embedding_model]), but EMBEDDING_SIZES is an empty hash which may lead to a KeyError at runtime. Consider revising the logic or providing a proper default embedding size.
| @defaults[:dimensions] || EMBEDDING_SIZES.fetch(defaults[:embedding_model]) | |
| @defaults[:dimensions] || EMBEDDING_SIZES.fetch(@defaults[:embedding_model], nil) |
Copilot uses AI. Check for mistakes.
Implemented GigaChat LLM support. Please take a look.