[Neuroimaging] Website Update

vanessa sochat vsochat at stanford.edu
Mon Aug 24 03:32:47 CEST 2015


Fantastic idea Ben! I think we possibly have found something that would
work.

Github pages:  http://vsoch.github.io/nipy-jekyll/
Circle:
https://circle-artifacts.com/gh/vsoch/nipy-jekyll/53/artifacts/0/home/ubuntu/nipy-jekyll/_site/home.html
Locally (using the --baseurl="/")

A few notes:

- The above was done by replacing the baseurl with one generated from
circle environmental variables
<https://github.com/vsoch/nipy-jekyll/blob/ben/circle.yml#L15>
- Github pages does not need a Gemfile to work.
- I don't know what circle does with gh-pages branches, but they don't
work, even when the code is identical.

It looks like the last little bug is the paths for the posts - we likely
need another regular expression. I'm closing up shop for today, but it's
great that we made progress! We can fix this last bug, then address the
content, and then hopefully have something polished enough to work. I'm
also thinking it would be a good idea to contact circle about this - it
should be their default to change the baseurl so the artifacts render
properly.

On Sun, Aug 23, 2015 at 5:25 PM, vanessa sochat <vsochat at stanford.edu>
wrote:

> ok, I'm testing this out! I was just giving Ariel's solution a try but it
> broke circle in weird ways. Apologies for delay in response, I was working
> on useless and silly things, and have about an hour for nipy right now!
>
> On Sun, Aug 23, 2015 at 5:02 PM, Ben Cipollini <bcipolli at ucsd.edu> wrote:
>
>> OK, I think this will work:
>>
>> * set baseurl in _config.yml to the repo name (/nipy-jekyll), since
>> github has no flexibility.
>> * locally, launch with --baseurl="/" (this is a semi-standard, from what
>> I see while searching)
>> * for circle.ci, write a small bash script that creates the circle.ci
>> base url from environment variables, and dumps it to a custom config.yml
>> file. Then call that shell script from circle.yml, and launch Jekyll with
>> that config file from circle.yml.
>>
>> That should be pretty simple to set up in all cases, keep things very
>> clean in the templates (just use the baseurl on all urls, as is currently
>> done).
>>
>> I'm heading out for the night, but may have time to try this in a branch
>> later tonight or tomorrow.
>>
>>
>> On Sun, Aug 23, 2015 at 2:46 PM, Ben Cipollini <bcipolli at ucsd.edu> wrote:
>>
>>> If it's just a matter of configuring a variable like the base url of the
>>> website, then I say no--no problem.
>>>
>>> If we have to work around it by manipulating URLs in funky ways inside
>>> templates... then I say, probably.
>>>
>>> I don't find this to be a Jekyll issue as much as how the different
>>> publishing sources expose Jekyll config. If they simply set the baseurl
>>> properly for ya, this would all be so simple! That's exactly what that
>>> config variable is there for... and Vanessa's using it quite properly in
>>> all the template code.
>>>
>>> Ben
>>>
>>>
>>>
>>> On Sun, Aug 23, 2015 at 2:43 PM, Matthew Brett <matthew.brett at gmail.com>
>>> wrote:
>>>
>>>> Hi,
>>>>
>>>> On Sun, Aug 23, 2015 at 10:40 PM, Ben Cipollini <bcipolli at ucsd.edu>
>>>> wrote:
>>>> > Hi Vanessa,
>>>> >
>>>> > Another solution is simply to send Jekyll the proper baseurl in all
>>>> cases.
>>>> > In my eyes, that's the simplest and intended solution. I just can't
>>>> > understand why it's so hard to do easily, or at all.
>>>> >
>>>> > The relative paths thing is another solution, but I feel that's a
>>>> > workaround--Jekyll has the baseurl defined and configurable, it's
>>>> clearly
>>>> > part of the software design.
>>>> >
>>>> > Hope we find an elegant solution soon :)
>>>> > Ben
>>>>
>>>> Forgive my anxiety, but this conversation worries me.  It seems to me,
>>>> that if we are having to do some serious custom configuration to make
>>>> a jekyll site work, it goes rather a long way to defeating the purpose
>>>> of jekyll as compared to - say - sphinx - in that it makes the site
>>>> more difficult and fragile to maintain.   Is that worry reasonable do
>>>> you think?
>>>>
>>>> Cheers,
>>>>
>>>> Matthew
>>>> _______________________________________________
>>>> Neuroimaging mailing list
>>>> Neuroimaging at python.org
>>>> https://mail.python.org/mailman/listinfo/neuroimaging
>>>>
>>>
>>>
>>
>> _______________________________________________
>> Neuroimaging mailing list
>> Neuroimaging at python.org
>> https://mail.python.org/mailman/listinfo/neuroimaging
>>
>>
>
>
> --
> Vanessa Villamia Sochat
> Stanford University
> (603) 321-0676
>



-- 
Vanessa Villamia Sochat
Stanford University
(603) 321-0676
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/neuroimaging/attachments/20150823/9cd1dd4d/attachment-0001.html>


More information about the Neuroimaging mailing list