improve row number estimation

[severo]

fixes #26

[Copilot]

Pull request overview

This PR improves the algorithm for guessing byte offsets when fetching rows at random positions in a CSV file, specifically addressing issue #26 related to handling serial rows more efficiently.

Key Changes:

• Replaced the isStored() and guessByteOffset() methods in the Estimator class with more comprehensive getStatus(), getFirstMissingRow(), and getLastMissingRowNumber() methods
• Updated the fetch logic in csvDataFrame to use the new methods, avoiding unnecessary fetches when rows are already cached
• Enhanced the algorithm to better estimate byte offsets, particularly for scenarios where average row size had been overestimated

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 9 comments.

File	Description
src/cache.ts	Refactored Estimator class to replace isStored() and guessByteOffset() with new methods (getStatus(), getFirstMissingRow(), getLastMissingRowNumber()) that provide more detailed information about row status and byte offsets
src/dataframe.ts	Updated fetch logic to use new estimator methods, eliminating the previous loop-based approach for determining which rows need to be fetched and improving the initial state detection
test/cache.test.ts	Updated and expanded tests to cover the new estimator methods, including edge cases for complete caches, empty caches, and rows stored at various positions
test/dataframe.test.ts	Updated test expectations to reflect improved byte offset estimation behavior for serial rows

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.