refactor(core): unify initialization and testing of explainer and `…

[DYung26]

This pull request refactors initialization and test logic for the PlotExplainer and VisualizationRecommender classes, improves test reliability, and introduces a dedicated response parsing class. The main themes are improved default argument handling, more robust provider/model management, and test suite modernization to match the refactored APIs.

Class Initialization and Provider Management:

• Added default values for all arguments in the VisualizationRecommender.__init__, and consistently store api_keys as a class attribute for use in provider management. [1] [2] [3]
• Updated provider manager initialization to always use the internal self.api_keys attribute, ensuring consistent API key handling. [1] [2]

Test Suite Refactoring:

• Refactored tests to use utility functions (save_plot_to_image, encode_image) directly, instead of methods on the explainer instance, improving modularity and reliability. [1] [2]
• Updated tests for model/provider detection to check for available providers and models using the new API, with more robust assertions and error messages. [1] [2]
• Changed plot generator tests to use the _default_plots attribute and its initializer, reflecting internal API changes and improving clarity. [1] [2] [3]

API and Interface Changes:

• Deprecated or commented out direct access to internal attributes and legacy methods in tests, favoring new interfaces and improved encapsulation. [1] [2] [3] [4]
• Updated test logic to use new methods like recommend_visualizations and its returned objects for prompt building and dataframe description, replacing direct calls to internal methods. [1] [2]

Response Parsing Refactor:

• Introduced the ResponseParser class and updated all relevant tests to use it for recommendation parsing, improving separation of concerns and code clarity. [1] [2] [3]

Mocking and Test Reliability:

• Improved test reliability by updating how mock clients and models are injected into recommender instances, ensuring that tests reflect the new model/provider management logic. [1] [2]

These changes collectively modernize the codebase, improve maintainability, and ensure that tests are robust and aligned with the latest API refactors.