VisiData is an interactive multitool for tabular data. It combines the
clarity of a spreadsheet, the efficiency of the terminal, and the power
of Python, into a lightweight utility which can handle millions of rows
with ease.
It's very flexible - see https://www.visidata.org/blog/2020/ten/ for
examples of what it can do, https://jsvine.github.io/intro-to-visidata/
for a tutorial, and the vd(1) manual for a good reference guide.
All of the obvious data formats are supported, as are more specialist
types including pcap, HTML tables, SQL databases, geographic data and
more - see https://www.visidata.org/formats/. This package does not
force optional dependencies for all of these - check the list or try
opening the file and see the "module not found" message - pkglocate
is useful to find the relevant py3-module to install.
- take maintainer (thanks Ingo)
- enable fortran
- use configure argument for example dir instead of post-install mv(1)
- API changes require major bump to all shlib
- pull patches from upstream to respect DESTDIR for examples
- fix h5repack_plugin.sh test
OK feinerer@, extra testing for gdal,python3 from landry@
- don't patch away upstream's INCLUDEDIR/LIBDIR mechanism to replace
it with something using CPPFLAGS/LDFLAGS which does exactly the same;
simplifies the patch
- building the targets chosen for the port requires a compiler with
128-bit int support; afaik this requires clang, or GCC 4.6 or newer,
and a 64-bit arch. So I've set the compiler and restricted archs.
(alternatively we could use "ALL_TARGETS=lrs64" on those archs but that
would mean extra mess with arch-specific PLISTs).
- simplify the test target
Doing "import sklearn" resulted in breakage at runtime due to sklearn
0.20.3 not being compatible with python 3.8.
In addition:
- add WANTLIB
- rework the DEPs for the new version
- remove SKLEARN_NO_OPENMP which was dropped in commit 9876f748
Breakage initially reported by James Cook (who provided many pointers on
moving forward) and also tested the update.
