2.0.34: Merge pull request #252 from texadactyl/master

In fil2h5, h52fil, dice, and bl_scrunch, it is possible to mangle the output file name if the input file path has an embedded "fil". In essence, the embedded "fil" will be inadvertently changed to "h5". E.g. "abc.filter.def.fil" --> "abc.h5ter.def.h5" (should be "abc.filter.def.h5"). This was caused by misuse of the Python string replace function.

Solution:

• Implemented utils.change_the_ext(path, old_ext, new_ext) will change only the final old_ext in the path or append the new_path if there is no old_ext present.
• fil2h5, h52fil, dice, and bl_scrunch call the new function.

Since the function is generic, it can also be leveraged by other projects such as turbo_seti (probably needed!).

Enhance h5diag for FBH5 version 2.0

For a valid FBH5 version 1.0 file, h5diag reports like this:

h5diag: Begin
h5diag: VERSION attribute: 1.0
h5diag: header: {'DIMENSION_LABELS': array([b'frequency', b'feed_id', b'time'], dtype=object), 'az_start': 0.0, 'data_type': 1, 'fch1': 8421.386717353016, 'foff': -2.7939677238464355e-06, 'ibeam': 1, 'machine_id': 20, 'nbeams': 1, 'nbits': 32, 'nchans': 1048576, 'nifs': 1, 'rawdatafile': 'guppi_57650_67573_Voyager1_0002.0000.raw', 'source_name': 'Voyager1', 'src_dej': <Angle 12.183 deg>, 'src_raj': <Angle 17.16777333 hourangle>, 'telescope_id': 6, 'tsamp': 18.253611008, 'tstart': 57650.78209490741, 'za_start': 0.0}
h5diag: data shape: (16, 1, 1048576)
h5diag: No errors detected

For a valid FBH5 version 2.0 file (created by rawspec 3.0 or later, producing an .h5 file directly), h5diag reports like this:

h5diag: Begin
h5diag: VERSION attribute: 2.0
h5diag: header: {'DIMENSION_LABELS': array([b'time', b'feed_id', b'frequency'], dtype=object), 'az_start': 0.0, 'data_type': 1, 'fch1': 7438.96484375, 'foff': -2.7939677238464355e-06, 'ibeam': -1, 'machine_id': 20, 'nbeams': 1, 'nbits': 32, 'nchans': 67108864, 'nfpc': 1048576, 'nifs': 1, 'rawdatafile': 'blc13_guppi_57991_49836_DIAG_FRB121102_0010.0000.raw', 'source_name': 'DIAG_FRB121102', 'src_dej': <Angle 33.1479 deg>, 'src_raj': <Angle 5.53295333 hourangle>, 'telescope_id': 6, 'tsamp': 18.253611007999982, 'tstart': 57991.57680555555, 'za_start': 0.0}
h5diag: data shape: (3, 1, 67108864)
Rawspec version: 2.5.0+91@gd699853
Librawspec version: 2.5.0+91@gd699853
cuFFT version: 10.2.1.245
HDF5 version: 1.8.16
Bitshuffle: ENABLED
h5diag: No errors detected

New utility: srcname

Created a tool for patching the header source_name in an HDF5 file: srcname.

New blimpy utility: peek

Merge pull request #245 from texadactyl/master

New blimpy utility: peek

Fix issue #243

Merge pull request #244 from texadactyl/master

Fix issue #243

Improve clarity in file writing messages

Impacted source files:

• io/fil_writer.py
• io/hdf_writer.py
• fil2h5.py

Fix issue #238 and a surpise bug in blimpy/utils.py

• There were miscellaneous bugs in stax.py w.r.t. frequency range specifications and large files.
• blimpy/util.py rebin function sometimes crashed due to unexpected dimension input.

Fix issues with stix and stax

1. Recovered lost executable stix.
2. The entry point name for ```stax`` was corrected to be consistent with the rest of blimpy.

Fixed a lot of mishaps in the regression tests

Merge pull request #236 from texadactyl/master

Fix problems in various regression tests

Centralise regression test requirements

Blimpy testing needs setigen for regression testing.
Setigen needs blimpy to save data in Filterbank or HDF5 files.

So, when the blimpy dependency of setigen is installed and blimpy is not yet installed, testing proceeds with the previous release level from pypi.

Solution:
Take setigen out of requirements.txt and put it and all the other test requirements in requirements_test.txt (new file).

Impacted files:

• .github/workflows/python_tests.yml
• Dockerfile