the path to a 0.7.0 release about chiantipy HOT 20 CLOSED

I am pretty confidant that the ionization cross-sections and rates agree with those from IDL. The original calculations were done in IDL.

from chiantipy.

Sorry I just now saw this. I've been a bit out of the loop with Chianti lately.

The GitHub repo can already be downloaded as a zip file. If possible, I think it best to first port all of the history of the sourceforge repo to GitHub and then delete the sourceforge page entirely. I think hosting the code multiple places mostly confuses people and versions on one site or the other fall out of date quickly.

Agree about the PyPI release.

Another possibility for distribution is putting ChiantiPy on the openastronomy conda channel. This makes the install really easy for those who use the Anaconda distribution. @Cadair has expressed interest in helping with this I think.

from chiantipy.

I think it would be nice to clean up and update some of the tutorials and examples before doing a new release as well.

from chiantipy.

I am happy to make a conda package for chiantipy, as this would be needed for SunPy to rely on ChiantiPy in the future. I had previously voiced the idea of making a separate package for the database do you think this is worthwhile?

from chiantipy.

putting ChiantiPy on Anaconda is fine with me. Also, we release the database as a gzipped tar file and at that point it is free for anyone to do anything with it. So, putting it also on conda, from the released tarball is also fine.

from chiantipy.

We could save all of the stuff that is under svn on sourceforge but this is not my first priority. What I would like to do is a 0.7.0 release as the major error in the ionization cross-sections and rates has been fixed and verified.
On github, I guess that means creating a tag and making a release. I see that the SunPy releases include various tars etc of the package so we should do the same.
Also put it on PyPI. I would not but it on Sourceforge except a simple statement that ChiantiPy has moved to Github.
I asked that the ChiantiPy ASCL.net page be updated to show github but that has not happened yet. I did that some time ago.
Finally, send out a message on the chianti google group

from chiantipy.

Re: moving the svn history from sourceforge, yes I did not mean to imply that should be done before the release (it may be a bit of a task!)

Ah yes I see what you mean about the tarballs as part of the releases. I had forgotten about that.

I also sent a request to ASCL and noticed that the entry had not been updated. I think they are a bit slow when it comes to updating some of their references.

Re: the versioning, since this release is primarily a bug fix, maybe it should be 0.6.6 instead of 0.7? Then release 0.7 once the documentation and tutorials/examples have been a bit more revised. This isn't really that important except that incrementing the second digit in version numbers is typically indicative of major changes in the code.

from chiantipy.

I have been updating the documentation and I would like to do a 0.7.0 release soon.

One problem with the documentation is that sphinx is expecting to find the method setup but it is now _setup so it does not get documented.

from chiantipy.

This is in reference to the _setup method on the ion object?

Methods with an underscore are considered to not be part of the public API and so are not documented. I guess since the user has the option for this method to not be called in the constructor, then it should be public so both _setup and _setupIonrec should have the underscores removed.

from chiantipy.

I agree. I will do that

from chiantipy.

just removing the under-score broke something, will check it out

from chiantipy.

Where specifically is it failing? Could you paste the Traceback?

from chiantipy.

it does not crash, just doesn't work like it should.
in $HOME/chianti/sutherland
there is a k_4.elvlc, .wgfa and .scups

k4 = ch.ion('k -4',1.e+5, 1.e+9, setup=0)
k4.setup(alternate_dir='/home/ken/chianti/sutherland')
and I get
elvlc file does not exist: /data2/svn/chianti/dbase/trunk/k/k_4/k_4.elvlc

it should be looking in $HOME/chianti/sutherland

the setup method with the alternate directory is really for code/data developers but it is very handy for working with data that is not yet in the CHIANTI repository

from chiantipy.

Ok. Can you create a new issue for this?

from chiantipy.

PyPI wants a README.txt file that is in rst format so it can be properly displayed. I will see if I can dig up an old one and go from there

from chiantipy.

see #97. This is a direct translation from the markdown version

from chiantipy.

just what we needed. I plan to do a release this week, hopefully sooner than later

from chiantipy.

Is astropy really a dependency? I think we need to list IPython(3) and Jupyter(3) as very helpful but not dependencies

from chiantipy.

I just made a 0.7.0 release. I did not do anything about making a tag. I will put it on PyPi soon

from chiantipy.

0.7 has been released. This issue can be closed.

from chiantipy.