Permalink
Commits on Apr 2, 2016
  1. Added basic README.rst.

    committed Apr 2, 2016
  2. First step of re-write/simplification.

    Removed all files. Added back src/, pytpm/. setup.py and setup-dev.py
    both modified significantly. Added Makefile. Removed fake_pyrex:
    no-longer needed. Removed all data/, tests/ and doc. I need to
    restrict myself to testing Python code and not TPM code.
    committed Apr 2, 2016
Commits on Jul 10, 2012
  1. Ceasing development.

    :(
    committed Jul 10, 2012
Commits on Feb 2, 2012
  1. New leap second: 2012 July 1

    Accumulated leap seconds now stands at 35.0.
    committed Feb 2, 2012
Commits on Aug 28, 2011
  1. Fixed missing ReST reference link.

    committed Aug 28, 2011
  2. Version 0.7.

    Removed convert.convert.
    Fully documented examples.
    All important conversions tested against SLALIB and results
    included in documentation. Test scripts are available in
    data/.
    Better documentation.
    committed Aug 28, 2011
  3. Changed version to 0.7 from 0.7dev.

    committed Aug 28, 2011
Commits on Aug 25, 2011
  1. Cleaned up scripts in data/.

    Cleaned up aop.py, map.py, summary_test_slalib.py
      aop.py, map.py :Use Numpy array operations when possible.
      In above and summary_test_slalib.py:
          Add ".." to path only if import fails.
    Cleaned up data/summary_test_slalib.py.
        summary_test_slalib.py: print std, instead of variance.
        data/read_data.get_sla(): return Numpy array.
        data/slalib_convert_hip.py: remove trailing space from format
          string for ecleq, eqecl, eqgal and galeq.
        pytpm/tests/data/slalib_hip_<above>.txt were edited to
          remove the trailing space.
    committed Aug 25, 2011
Commits on Aug 24, 2011
  1. MANIFEST.in: Include files from data/, except *vot.

    Minor mofification of devel-notes.rst.
    committed Aug 23, 2011
Commits on Aug 23, 2011
  1. Removed unnecessary function from summary_test_slalib.

    The relevant information can be produced using data/map.py
    and data/aop.py.
    committed Aug 23, 2011
Commits on Aug 19, 2011
Commits on Aug 18, 2011
  1. Change cpdef to def in src/convert.pyx.

    def functions can have cdef and other statements.
    committed Aug 18, 2011
Commits on Aug 17, 2011
  1. Summary of comparisons: sla_aop -> aop.py, sla_map -> map.py.

    These are files used for generating detailed summary of
    comparisons between TPM and SLALIB. This is needed since in the
    automated tests I merely test if the values are within a range.
    
    The file data/summary_test_slalib.py have comparisons for
    other conversions.
    committed Aug 17, 2011
  2. Tests for observed az, zd, ha, dec and ra with SLALIB sla_aop.

    Data from SLALIB conversion is in
    pytpm/tests/data/slalib_hip_aop.txt. This file was created using
    appropriate section of data/slalib_convert_hip.py Tests are in
    pytpm/tests/test_slalib.py.
    committed Aug 17, 2011
Commits on Aug 13, 2011
  1. convert.cat2v6

    Takes one or more coordinates and returns V6C vectors for each
    of the coordinates.
    committed Aug 13, 2011
  2. v62cat extracts normalized alpha. convert.v62cat.

    The new function accepts a list of V6C objects, and returns a
    list of dictionaries containing catalog entires for each of
    the V6C vectors.
    committed Aug 13, 2011
Commits on Aug 12, 2011
Commits on Aug 8, 2011
  1. Tests for FK5 to apparent (i.e., geocentric) RA-DEC.

    The maximum difference between SLALIB and PyTPM is 0.33" in RA
    and 0.03" in Dec. Am I doing something wrong: transformation between
    mean places are identical.
    committed Aug 8, 2011
  2. Test for FK4-FK5 conversion.

    Since I don't have FK4 proper motion data, I am using the
    FK5-FK4 conversion performed using sla_fk524, for testing
    FK5-FK4, as input.
    
    I use sla_fk425 on this data and save it as slalib_fk524_fk425.txt,
    and then compare the results from PyTPM to this.
    
    Function that prints out a summary of comparison results are in
    the file data/summary_test_slalib.py.
    committed Aug 8, 2011
Commits on Aug 6, 2011
  1. Tests for HIP ECl and GAL conversions, and fixed NDWFS tests.

    Added tests for ECL-EQ, EQ-ECL, GAL-EQ and EQ-GAL conversions.
    
    Found out that for GAL<->EQ the galactic data from Vizier is
    for ep=J2000. So for EQ-GAL I need to apply proper motion before
    comparison with SLALIB. For GAL-EQ, SLALIB needs GAL in B1950.0.
    I can't apply proper motion from J2000 to B1950, since I don't
    have velocities. In essence SLLAIB will be converting ep=J2000
    data, and hence the final result will have the extra J2000-B1950
    proper motion. So I just set epoch=tpm.B1950 for PyTPM
    gal-eq conversion. This will leave the result in eq=J2000 and
    ep=B1950. This result will match the output from SLALIB.
    
    At-least this is what I think is happening!
    
    I applied this to NDWFS tests as well.
    
    For ECL<->EQ I was using the wrong MJD; MJD for 2000.0 is
    51544.5 not 51545.0. This change was applied to NDWFS data, i.e.,
    the slalib_ndwfs_ecleq.txt and slalib_ndwfs_eqecl.txt were
    regenerated by applying appropriate changes in
    data/slalib_convert_ndwfs.py.
    
    Summary results for SLALIB HIP data for ECL<->EQ and GAL<->EQ are
    in data/summary_test_slalib.py. Currently this is only for HIP
    data.
    committed Aug 6, 2011
  2. summary_test_slalib.py: functions to summarize test results.

    In the automated tests, I only test if the difference between SLALIB
    and PyTPM is within some limit. But that can be contaminated by a few
    bad data points. So this module  has functions to get some details of
    the comparison. In here I repeat the code used in the test_slalib.py
    file. But then use Numpy and Scipy for obtaining statistics.
    
    In this commit there is one function, hipfk54() that prints summary
    of comparison between FK5 to FK4 conversion using SLALIB fk524 and
    pytpm (convertv6(v, s1=6, s2=5) + PM from J2000 to B1950).
    committed Aug 6, 2011
  3. Test with SLALIB fk524 using HIPPARCOS data.

    Test FK5 to FK4 conversion, using the Vizier generated FK5 values
    for HIPPARCOS data. See end of the method
    TestSLALIBHIPFK54.test_slalib_hip_fk52 for accuracies.
    
    The formatted HIP data was generated using the script
    data/hip_full.py and stores in pytpm/tests/data/hip_full.txt. The
    code used to perform conversions with slalib is in
    data/slalib_hip_fk54.txt. The original HIP data in VOTable
    format is in data/hipparcos.vot.
    
    Added functions to read data from slalib conversions, and the
    formatted HIP data. These are present in data/read_data.py
    and in the test module pytpm/tests/test_slalib.py.
    
    Added NDWFS to docstring of NDWFS based tests for easy
    identification in test output.
    committed Aug 6, 2011
  4. Renamed data/slalib_conversions.py -> slalib_convert_ndwfs.py.

    Also fixed pep8.
    
    Fixed path to test data dir. in comment, in the above file.
    committed Jul 28, 2011
Commits on Jul 26, 2011
  1. Tests with SLALIB conversions of NDWFS data.

    Consistent within data accuracy (or what I think are reasonable
    given number of digits in the input data).
    
    See data/slalib_conversions.py and data/ndwfs.py for conversions
    and how the catalog data was obtained.
    committed Jul 26, 2011
  2. Renamed test data files, in preparation for "external" tests.

    I am going to add tests that compare PyTPM results with those
    from SLALIB. FOr this I have prepared data from HIPPARCOS
    and from the NDWFS catalog.
    committed Jul 26, 2011
  3. Python 3!

    Changes made:
    
    In pytpm/__init__.py:
    Use explicit relative import. Since this works in Python 2.6 +
    there is no problem with 2.6/2.7 support.
    
      from . import tpm
      from . import convert
    
    Deleted __unicode__ and replaced it with __str__.
    
      __str__ returns .decode("utf-8")
    
    Changed all (I think) str() to .decode("utf-8")
    
    In setup.py:
    
    Explicit inclusion of package data, even though MANIFEST.in exists.
    
      use_2to3 = True
    
    Changed xrange to range: but I think 2to3 does this.
    
    Changed test_suite option to function in pytpm/tests/__init__.py
    that returns a test suite containing all tests in all test
    modules i.e., pytpm/tests/*.py. When I add another test module,
    I need to add this in __init__.py file.
    
    The code actually runs on Python 3, only that instead of strings
    it returns b'...'. But to run the tests using setup.py and to
    pass tests that check strings, I need to use 2to3.
    
    All in all an almost painless process, with only problem being
    figuring out which setup() options to use. Only a careful
    reading of distibute documentation can help in this.
    committed Jul 25, 2011
Commits on Jul 25, 2011
  1. Added convert.proper_motion: takes multiple V6 values.

    This merely calls tpm.proper_motion, but since this gets
    translated into C, the performance must be better than doing this
    from within Python.
    
    Test added.
    committed Jul 25, 2011
  2. Removed convert.convert and tests for it.

    This function is not needed. Full 6D information is needed for
    proper usage.
    committed Jul 25, 2011
Commits on Jul 21, 2011