Use Parsera LLM Specs API instead of capabilities.rb
#132
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Owner
crmne
commented
Apr 22, 2025
crmne
commented
- Removed the OpenAI capabilities module and related methods.
- Simplified model parsing for OpenAI and Gemini providers, focusing on essential attributes.
- Introduced a new utility method for deep key symbolization in hashes.
- Updated OpenRouter model parsing to enhance data extraction and structure.
- Deleted obsolete browser and CLI helper tasks to streamline the codebase.
- Consolidated model updater logic into the Rake task for better maintainability.
- Improved error handling and logging throughout the model update process.
- Removed the OpenAI capabilities module and related methods. - Simplified model parsing for OpenAI and Gemini providers, focusing on essential attributes. - Introduced a new utility method for deep key symbolization in hashes. - Updated OpenRouter model parsing to enhance data extraction and structure. - Deleted obsolete browser and CLI helper tasks to streamline the codebase. - Consolidated model updater logic into the Rake task for better maintainability. - Improved error handling and logging throughout the model update process.
crmne
changed the title
capabilities.rb
jayelkaake
reviewed
Apr 22, 2025
| module Utils | ||
| module_function | ||
|
|
||
| def symbolize_keys_deep(hash) |
There was a problem hiding this comment.
jayelkaake
reviewed
Apr 22, 2025
| end | ||
|
|
||
| def fetch_parsera_models | ||
| response = Faraday.new('https://api.parsera.org/v1/llm-specs', request: { timeout: 60 }) |
There was a problem hiding this comment.
I don't like the idea of a core dependency on a commercial 3rd party's API, though very much appreciate the value that Parsera is doing by standardizing LLM capabilities into a JSON.
I know the code isn't called during any main runtime functions and also appreciate the support from Parsera (seems like great company).
Maybe just move this code to the rake task? Don't feel strongly about this so if you don't want to that's fine.
|
Love this - the initiative by Parsera to standarize capabilities of models is fantastic! Although, I'm not confident that the LLMs are going to follow the same configs within the same capabilities so I question whether the capabilities should just be implemented in the provider classes (and if they're not implemented it means they aren't supported). For example, structured output is supported by Gemini but it is a vastly different implementation than the other LLMs. Anyways, still prob good to do this since it's just a better implementation of how the gem already works today. |
…emspec description
This PR implements the ability to specify custom dimensions when generating embeddings, as requested in issue #47. ### What's included - Added support for passing a dimensions parameter to the embed method - Implemented dimensions handling in both OpenAI and Gemini providers - Added tests to verify dimension param works correctly - Optimized the Gemini provider's `embed` method to reduce unnecessary API calls when embedding texts, resulting in lower token usage. From now on, it uses `batchEmbedContents` endpoint within one request, for both single and multiple text embeddings. - Modernize Gemini embeddings following DIP principle, as implemented in `openai/embeddings.rb`. - The Gemini embeddings API response does not contain the promptTokenCount attribute, so I have removed it. ### Implementation notes I've decided to only implement the per-request dimension configuration and not the global configuration option that was initially proposed in the issue. This is because each embedding model has its own default dimensions, making a global setting potentially confusing. With this implementation, users can set the embedding dimensions like: ```ruby embedding = RubyLLM.embed( "Ruby is a programmer's best friend", model: "text-embedding-3-small", dimensions: 512 ) ``` ### References - OpenAI API docs: https://platform.openai.com/docs/api-reference/embeddings - Gemini API docs: https://ai.google.dev/api/embeddings Resolves #47 --------- Co-authored-by: Carmine Paolino <[email protected]>
…s_chat Previously, API errors during a chat turn would leave an orphaned, empty assistant message record. This caused issues, notably with Gemini rejecting subsequent requests containing empty messages. Fixes #118
…lution in chat, embedding, and image methods Now `assume_model_exists` can be used in `paint` and `embed` methods too.
…al development support Closes #137
…ibuting guidelines
…M 4 and GLM Z1 models
Now uses POROs and simplifies implementation of providers. Fixes #143
…llama provider - Introduced `RubyLLM::Providers::Ollama::Media` module to manage media content formatting for OpenAI APIs. - Implemented `format_content` method to process text and attachments, including images, PDFs, and audio. - Added `format_image` method to convert image attachments into the required format.
…AI models with detailed descriptions
…#151) This updates the acts_as_message, acts_as_chat and acts_as_tool class methods to use Rails-style foreign keys whenever custom class names are used as options. For example: ``` class FooMessage < ActiveRecord::Base acts_as_message chat_class: 'FooChat', tool_call_class: 'FooToolCall' end ``` will now set the foreign key on the `belongs_to :chat` association to be `foo_chat_id`, instead of `chat_id`, and will set the foreign key on `belongs_to :parent_tool_call` association to `foo_tool_call_id` instead of just `tool_call_id`. This is consistent with Rails' naming conventions for class names and foreign keys. Changes are backwards-compatible with existing code/behavior, and don't require a major or minor version bump. Updated test cases to ensure that the associations are working, but didn't re-record VCR tests, since I don't have an OpenAPI key. Closes #150 Co-authored-by: Carmine Paolino <[email protected]>
## Purpose Introduces some basic configuration for the logging capabilities, allowing users to specify a custom log file and log level. This feature enhances debugging and monitoring capabilities by providing more control over where and how logs are recorded. ## Implementation Details - Added log_file configuration option (default: STDOUT) - Added log_level configuration option (default: INFO, or DEBUG if RUBYLLM_DEBUG env var is set) - Updated logger initialisation to use the configured log file and level - Added documentation for logging configuration - Maintained backward compatibility with existing logging behaviour ## Usage Example ```ruby # Global configuration RubyLLM.configure do |config| config.log_file = '/logs/ruby_llm.log' # Custom log file location config.log_level = :debug # Set log level end ``` ## Testing Manual - Verified logger initialisation with custom log file - Confirmed log level changes based on configuration - Tested environment variable override for debug level - Validated default behaviour (STDOUT logging) ## Documentation Added a new section in configuration.md for logging settings Co-authored-by: Carmine Paolino <[email protected]>
- Removed the OpenAI capabilities module and related methods. - Simplified model parsing for OpenAI and Gemini providers, focusing on essential attributes. - Introduced a new utility method for deep key symbolization in hashes. - Updated OpenRouter model parsing to enhance data extraction and structure. - Deleted obsolete browser and CLI helper tasks to streamline the codebase. - Consolidated model updater logic into the Rake task for better maintainability. - Improved error handling and logging throughout the model update process.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.