Frédéric Péters
7bc774ae29
debian: use debhelper compat level 12
2021-12-12 14:38:56 +01:00
Frédéric Péters
ba36854230
debian: declary numpy dependency
2020-07-27 07:58:42 +02:00
Frédéric Péters
a61cfc74af
debian: add initial packaging
2020-07-26 21:22:17 +02:00
0b72da7ae6
Version 2.14
2020-03-26 12:41:21 -04:00
193eb4d49d
Slightly simplify new mmap() failure detection
2020-03-26 12:35:43 -04:00
059915a878
Add test of new mmap() OSError catching
2020-03-26 12:33:59 -04:00
dc578f6582
pyodide: no mmap ( #39 )
...
This lets the library work on platforms without `mmap()`.
2020-03-26 12:17:38 -04:00
f5150e4d67
Version 2.13
2020-02-22 11:41:02 -05:00
3b3fc6eba7
gitignore BSP files
2020-02-22 08:12:46 -05:00
cf914ab2ee
Set ValueError .args attribute, for Python 2.7
...
This should hopefully fix up our CI results.
2020-01-04 16:05:54 -05:00
58b889b052
Simplify and test new out-of-range exception
...
1. Remove “Timestamp” from the exception name since no timestamps are
involved; the values are barycentric Julian dates.
2. Display Julian date values in the error message, since those are the
actual time values that users will have submitted.
3. Remove the “min” and “max” attributes because it turns out that (even
though I had forgotten!) those values are already available as
attributes on the class.
4. Added a test to hopefully keep the new feature working.
5. Added the new feature to the changelog.
2020-01-04 16:00:58 -05:00
7c94589a25
Support “-v” to run tests verbosely
2020-01-04 15:31:31 -05:00
2ad36b41d2
Raise a more specific error when the date is not covered ( #36 )
2020-01-04 15:29:35 -05:00
fa1f6f2b19
For Travis CI: a more recent numpy for Python 3.7
2019-12-16 12:16:57 -05:00
6cabfb53c1
Fix load_tests() contract (for Travis CI)
2019-12-16 12:09:17 -05:00
b7d7652c66
Protect doctests against old numpy versions
2019-12-16 11:43:52 -05:00
3f3a762baf
Lock down NumPy version on Travis CI, to fix #35
...
This should stop jplephem from accidentally using features too modern
for users on several-year-old systems.
2019-12-16 09:17:01 -05:00
8927698a85
Release 2.12 that stops using new NumPy flip()
2019-12-13 16:05:11 -05:00
b1a7bf313b
Proclaim Python 3.7 compatability
2019-12-13 09:31:00 -05:00
3b78a1f9b5
Update changelog again and bump version to 2.11
2019-12-13 09:08:58 -05:00
3ee363f243
Remove an unused value
2019-12-13 08:53:02 -05:00
089713252d
Try also flipping SPK coefficient order
2019-12-13 08:46:34 -05:00
804f177b86
Create a changelog entry for 2.11
2019-12-13 07:45:35 -05:00
185eecba0b
Reshape PCK coefficients for natural iteration
2019-12-12 21:53:42 -05:00
901452ab3c
Remove commented out old code
2019-12-12 21:19:09 -05:00
e8c25e1ae9
Fix PCK rate computation
2019-12-12 21:17:22 -05:00
6f792c5865
Try switching velocity computation too
2019-12-12 21:10:49 -05:00
75c9f62bdf
Switch to accumulating Chebyshev in opposite order
...
This gives us exact agreement with the numbers generated by the
equivalent SPICE routines, making comparison easier. After reworking
velocity the same way, I plan to clean up the debugging chaos, but want
to leave it in place until the transform is complete.
2019-12-12 20:51:55 -05:00
5b11851bc8
Switch PCK internals to seconds since epoch
...
This eliminates some internal conversions and more closely models how
other libraries do this computation. I had also hoped it would improve
precision, but am not yet seeing any closer agreement in Skyfield than
previously.
2019-12-12 11:35:43 -05:00
677d5a2656
Add changelog entry for 2.10
2019-12-11 15:29:01 -05:00
9055aed1c9
Tweak to get docstring passing test again
2019-12-11 14:55:43 -05:00
9ef877132f
Add simple shell script for testing locally
2019-12-11 14:48:51 -05:00
187abc1716
Add doctests to main test suite
...
Also, pivot to de421 since that ephemeris is already committed to the
repository and so is available to tests.
2019-12-11 14:48:05 -05:00
6995fe446c
Add ephemeris, docs for rotation; but, untested?
2019-12-11 14:21:43 -05:00
e6f544d353
Rename “binary PCK” to just “PCK” to match “SPK”
...
Also, rename a few variables internal to the class to better match the
definitions in the JPL documentation.
2019-12-10 20:57:16 -05:00
f4fbaa962f
Fix Travis CI by only doing one cd
2019-09-19 16:21:40 -04:00
aae60da592
Check in ephemerides necessary for CI testing
...
This prevents Travis CI from having to hammer the JPL servers when
testing jplephem against several versions of Python at once. It
will also stop CI from routinely erroring out one or more versions
of Python when the JPL’s servers happen to drop a request.
2019-09-19 16:05:24 -04:00
2cca7f11d9
Add ASCL citation to assist scholars
2019-09-17 09:40:16 -04:00
87a2e9f4c7
Try adding support for Type 9 segments
2019-08-26 20:25:58 -04:00
b66da735e7
Create scaffolding for multiple Segment classes
2019-08-22 16:13:47 -04:00
ba560cb086
Travis CI: abandon Python 3.3 but try 3.7
2019-08-22 16:07:43 -04:00
4bc4e355e8
Fix: setup.py did not correctly require NumPy
2019-08-22 15:43:16 -04:00
48c99ce40c
Add documentation on computing velocity per second
...
This addresses a question raised in #30 .
2019-02-03 07:08:20 -05:00
7dae84d506
Version 2.9: add load_array()
2019-01-03 23:34:10 -05:00
31549ce716
Prevent myself from running with old Python
...
Since I use an `ELLIPSIS` option that appears to not work under 2.7.
2019-01-03 23:31:26 -05:00
e39f7d605e
Add license file and include it in sdists ( #27 )
...
Add license file; include it in sdists
2018-12-29 12:52:51 -05:00
79151ac7a4
Make SPK class close mmaps when exiting context manager ( #25 )
...
* Properly close mmaps; respect with context manager
* Update test to use with context manager
* Remove try/except in SPK.close()
* Restore SPK mmap attributes to None on close()
2018-12-07 11:25:24 -05:00
c1b4c1882d
Start writing support for Binary PCK files
2018-10-01 08:58:54 -04:00
9926bb54cb
Version 2.8: single memory map instead of many
2018-07-22 16:16:43 -04:00
e01a35e1d9
Version 2.7
2018-02-11 15:37:58 -05:00
Frédéric Péters
Brandon Rhodes
jcaesar
Jérôme Deuchnord
Todd
James Davies