Previously the laszip backend implementation was using
the laszip executable in commandline mode using Popen
and communicating with its stdout, stderr, stdin.
It worked well (albeit the implementation was a bit hard).
However pylas 0.5 introduced 'chunked' reading & writing
which made the implementation of the laszip backend even harder.
It seems to work well, and simplifies the backend
implementation by a lot.
Unify Header classes & Vlr
The hierarchy of header classes (RawHeader1_1, RawHeader1_2, etc)
is removed and now only one class exists LasHeader, where most
of the fields that reader/writers care about are removed and
field that user care about are kept and put in user-friendly classes.
The vlrs are now also a part of the header
(it simplifies synchronizing, the, header, point_format and extra bytes vlr)
Extra Bytes:
The support for extra bytes should now be more complete.
'Scaled Extra Bytes', that is extra bytes with scales and offsets
are supported and can be created within pylas.
Simplify EVLR
Previously we had a EVLR class which was just a VLR subclass.
Now EVLR are just VLR, kepts in a different list, and written
as EVLRs if there are any.
VLRList subclass list
Remove pylas merge
Cleanups
Add a noxfile to simplify testing multiple version
This PR has most changes to 'prepare' pylas 1.0
Change the laszip backend
Previously the laszip backend implementation was using the laszip executable in commandline mode using Popen and communicating with its stdout, stderr, stdin.
It worked well (albeit the implementation was a bit hard).
However pylas 0.5 introduced 'chunked' reading & writing which made the implementation of the laszip backend even harder.
pylas 1.0 plan to replace the laszip backend implementation with proper laszip python bindings. https://github.com/tmontaigu/laszip-python
It seems to work well, and simplifies the backend implementation by a lot.
Unify Header classes & Vlr
The hierarchy of header classes (RawHeader1_1, RawHeader1_2, etc) is removed and now only one class exists LasHeader, where most of the fields that reader/writers care about are removed and field that user care about are kept and put in user-friendly classes.
The vlrs are now also a part of the header (it simplifies synchronizing, the, header, point_format and extra bytes vlr)
Extra Bytes:
The support for extra bytes should now be more complete. 'Scaled Extra Bytes', that is extra bytes with scales and offsets are supported and can be created within pylas.
Simplify EVLR
Previously we had a EVLR class which was just a VLR subclass. Now EVLR are just VLR, kepts in a different list, and written as EVLRs if there are any.
VLRList subclass list
Remove pylas merge
Cleanups
Add a noxfile to simplify testing multiple version