Seismic-zfp

What's Changed
* Fix headers for files with 512-multiple tracecounts by da-wad in https://github.com/equinor/seismic-zfp/pull/101
* [Snyk] Security upgrade numpy from 1.21.3 to 1.22.2 by ThomasRyschawyEquinor in https://github.com/equinor/seismic-zfp/pull/98


**Full Changelog**: https://github.com/equinor/seismic-zfp/compare/v0.4.1...v0.4.2

What's Changed
* Restore behaviour of accepting numpy integers as arguments to read functions by da-wad in https://github.com/equinor/seismic-zfp/pull/100


**Full Changelog**: https://github.com/equinor/seismic-zfp/compare/v0.4.0...v0.4.1

Overview
This is a **minor** release rather than a patch release to signify that numpy 2 compatibility can be expected, provided that zfpy>=1.0.1 is installed. Functions deprecated in numpy 2 have been removed, but since this is a pure-Python module some environments using numpy<2.0 and zfpy<1.0.1 can _also_ be expected to work.

What's Changed
* Remove numpy-1.x dependency restriction as zfpy and pyzgy are np-2.x compatible by da-wad in https://github.com/equinor/seismic-zfp/pull/96

New Contributors
* TovTyvold made their first contribution in https://github.com/equinor/seismic-zfp/pull/97

**Full Changelog**: https://github.com/equinor/seismic-zfp/compare/v0.3.6...v0.4.0

- adds **use_higher_samples_precision** option to **StreamConverter**
- Specifies whether to use higher precision for the sample interval and sample time.
Default is `False`. When set to `True`, stores sample interval and sample time as
64-bit floating-point numbers for increased precision. If `False`, they are stored
as 32-bit integers.

- Added new StreamConverter
- similar to NumpyConverter
- compresses a 3D numpy array to an SGZ file as a stream of chunks
- enables sequential compression without loading the entire array into memory
- Updated Numpy requirements to <2.0.0

- Allocate space in header for floating-point sample interval + offset