* Parquet Query Planner: plan byte ranges, pre-fetch in parallel.
- parquetPlan() that returns lists of byte ranges to fetch.
- prefetchAsyncBuffer() pre-fetches all byte ranges in parallel.
throws exception if non-pre-fetched slice is requested later.
Do this by passing rowGroupStart and rowGroupEnd for the rows to
fetch within a rowgroup. If a page is outside those bounds, we can
skip the page. Replaces rowLimit.
* build types before publishing to npm
* use prepare instead of prepublishOnly + make it clear that we only build types
doc for prepare vs prepublishOnly is here: https://docs.npmjs.com/cli/v8/using-npm/scripts
* no jsx in this lib
* relative imports from the root, so that it works from types/
* remove unused hyparquet.d.ts + report differences to jsdoc in files
* try to understand if this is the cause of the failing CI check
tsc fails: https://github.com/hyparam/hyparquet/actions/runs/12040954822/job/33571851170?pr=46
* Revert "try to understand if this is the cause of the failing CI check"
This reverts commit 5e2fc8ca179064369de71793ab1cda3facefddc7.
* not sure what happens, but we just need to ensure the types are created correctly
* increment version
* Explicitly export types for use in downstream typescript projects
* Use new typescript jsdoc imports for smaller package
* Combine some files and use @import jsdoc
* use the local typescript
---------
Co-authored-by: Kenny Daniel <platypii@gmail.com>
The type change caused a lot of downstream type errors.
If you pass rowFormat: 'object' then it will return Record<string, any>[]
instead of any[][]. This means the types are not aligned with behavior.
Will figure out how to fix it later, for now don't want break downstream projects.
