Replace explorations with notebook-centric API and redesign dataset module

[Irozuku]

This pull request introduces significant changes to the backend API and frontend design, primarily focusing on migrating functionality from datasets and explorations to a new notebook-centric model, removing the explorations endpoint, and expanding dataset file access capabilities. The changes also update validation logic and add new endpoints for interacting with datasets via file paths.

Notebook-centric API migration and endpoint updates:

• Removed the entire explorations API (endpoints/explorations.py), and updated the main API router to drop the /exploration endpoints and add /notebook endpoints, reflecting a shift from exploration-based to notebook-based workflows. [1] [2] [3]
• Refactored the converters endpoint to operate on notebooks instead of datasets: updated models, parameters, and logic to use notebook_id, and added a new endpoint to fetch finished converters by notebook. [1] [2] [3] [4] [5]
• Updated explorers endpoint validation and logic to reference notebooks (and their datasets) instead of explorations, ensuring all explorer operations are now notebook-centric. [1] [2] [3] [4]

Enhanced dataset file access and utility endpoints:

• Added multiple new endpoints to the datasets API to:
  • Fetch a sample of rows directly from a dataset file path.
  • Retrieve dataset info and column types by file path.
  • Paginate through dataset rows by file, returning row data and total count. [1] [2] [3]
• Improved error handling and JSON serialization for dataset file access. [1] [2]

These changes collectively modernize the backend data model, streamline API usage around notebooks, and provide more flexible, file-based dataset access for downstream consumers.

[cristian-tamblay]

In general it looks good, couldnt review it completely