Repository navigation
io.ascii - #154
io.ascii#154
Conversation
Changes: - Make all imports within package be relative - Migrated from nose to py.tests (required compatibility functions) - Added try/except in test modules to retain back-compatibility with standalone asciitable - Reduced excessive data in test_cds_header_from_readme
|
I definitely like it big-picture (I'm a fan of A couple organizational question/comments, though: What is the Also, I would also strongly encourage you to include the docs whenever you get a chance to that (before merging) - one of the best things about |
For unknown reasons this caused an obscure problem when testing on MacOSX.
|
@eteq : I'm working on getting the docs in now. This requires a wee bit of work since some of the content needs to be different (e.g. no installation instructions etc) from standalone
|
The io.ascii docs are automatically derived from the asciitable docs, with some slight modifications like elimination of download/installation instructions. Currently "asciitable" is globally replaced by "astropy.io.ascii", which is formally correct but gives a heavy look to some examples. Some modifications to io.ascii docstrings were also made to fix doc build errors or warnings. The io.ascii docs were refactored into separate .rst files for each section. The same change was made in asciitable docs.
There was a problem hiding this comment.
I would recommend you make this ``astropy.io.ascii Documentation (and add the necessary `===` below it) to be consistent with the rest of the documentation.
|
Looks pretty good (and tests pass for me), aside from the documentation comment I made inline above. Other than that, are you done making changes? |
|
If anybody has additional comments or wants more time for review then speak by Friday. Otherwise I'll plan to merge sometime this weekend. |
|
As you mentioned in #174, maybe you could add a legacy shim into this pull request? |
|
@astrofrog - since there is pending pull request #174 that will change the behavior of the legacy shim, I've opened issue #175 to track adding this for io.ascii and do it after #174 is merged. |
|
@taldcroft - sounds good. From my point of view, this looks ready to merge. |
|
Looks fine to me too... But @taldcroft, I didn't fully understand your comment above: do you mean you want to implement #175 in this pull request, or should we merge this pull request and then you'll issue another one to fix #175? |
When trying to create an overlay from a non-existent frame a crazy error is raised.
This is a preview pull-request for my rework of asciitable as astropy.io.ascii. The goal is to enable initial review and discussion.
All tests pass on linux x86_64 py3.2 and py2.7, but I've just gotten a slew of emails that OSX10.7 is failing miserably on TR's jenkins server. Will need to investigate. EDIT: this was resolved.
This code base is structured to allow rsync'ing of code and test files between standalone asciitable and astropy.io.ascii. Thus there are some things that might look awkward (and of course suggestions for improvement are welcome!). In addition asciitable supports 2.5 and doesn't require NumPy, so there are still more tests that look ugly.
On the plus side both asciitable (optionally) and astropy.io.ascii (by default) now take advantage of astropy.table.Table.
Docs are ported as well.