Feature: GetJson and GetCsv in short codes or other layout files.

[SchumacherFM]

I have finished the integration of getJson and getCsv into hugo as far as I think it fits.

I'm not a hugo (and Go) expert of where to put these functions and correct naming of functions and vars.

I would be glad if you can comment on spotted errors, bugs, bad design, etc. or even help me to improve the code more.

All tests and the new tests are passing. Goimports has reordered some import paths.

I have written a blog post about that new feature: http://cyrillschumacher.com/2014/12/21/dynamic-pages-with-gohugo.io/

There is not yet any official documentation because I want to get the code right. I think once this feature is merged as a hidden feature into hugo and some people tested it in the wild with no given complaints I will write the docs.

Looking forward for a successful merge :-)

[bep]

There are som failing tests on Windows. See the AppVeyor details above.

[SchumacherFM]

Fixed

[bep]

I guess @spf13 will chime in, but my first thoughts:

I love the simplicity of this approach. I know I was the one who wanted a cache for the remote resources, and I still believe that is important.

But I had a hope that there was a third-party Go cache libraries out there to use -- cause I would like that my cache entries was evicted and refreshed at some point (I want fresh content - but not on every livereload).

Other than that, this looks great -- a feature I look forward to use.

[bep]

@ahsanulhaque You are the editor (or something) of the gopheracademy and have seen a lot of blog articles floating by - any good Go caching library out there?

[ghost]

Hey, it really depends on what kind of caching you're trying to achieve. This is a fairly popular library and I've seen it used in many cool ways. I would probably write something minimal and use Redis.

[SchumacherFM]

IMHO it would be easier to add a 2nd integer parameter to getJson/getCsv which specifies the cache validity in seconds.

• < 0 means no caching
• 0 means cache is unlimited time valid (or until you remove the files)
• all other values > 0 are the seconds for being valid.
  The cache func only checks the file creation date and that's it.

[spf13]

I like the idea of caching it.
I want to keep it really REALLY simple.
No external dependencies.

My thoughts are using a tmp/ directory or cache/.

• It's really easy to invalidate (rm -rf cache/)
• No external dependencies
• No complicated libraries/logic
• Lasts longer than a single process (really important)

We should add a setting --ignoreCache or something if we go this route.

We can also in the future easily build in some expiration functionality if we start here. It's easy enough to check the mtime on the file and choose not to use it.

This does mean that we do need to wrap all external calls to check cache first. There's probably a number of places where a cache would be helpful if we sit down and think about it.

[thomasmodeneis]

+1.
That will be great value for me, thank you all.

[bep]

@spf13 I feel the API here is pretty set in stone (simple and clear). There may be stuff to improve on caching and documentation, but there are arguments to pull it in as a "silent feature" for a while ... to play with it and get a feeling.

[derekperkins]

👍 Great idea!

[SchumacherFM]

I should resolve the merge conflict and add the option --ignoreCache OR instead of introducing another CLI flag if --cacheDir has been omitted or --cacheDir=false then no caching. Hmm ...

[SchumacherFM]

I've added the option ignoreCache, fixed a bug when the cache folder didn't exists and also rebased my branch with the current commits from master.

[thomasmodeneis]

up !
sorry to disturb,
any project collaborators up to do this merge? o/
thank you

[thomasmodeneis]

I have it running with the fork github.com/SchumacherFM/hugo.

But the url I want to call is composed with prefix + id + sufix
The problem is that I can not make it work a way to concatenate vars with hugo:

{{ $id := .Params._id }}
{{ $url_pre :=  "http://localhost:3000/db/persons/" }}
{{ $url_post := "/limit/10/skip/0" }}
{{ $finalurl  := $url_pre + $id + $url_post }}
 {{ $gistJ := getJson $finalurl }}
{{range first 5 $gistJ }}
<li><a href="#" target="_blank">{{.name}}</a></li>
{{end}}

Is there a way to concatenate vars to build this composed url ??
Thank you.

[bep]

printf - see a recent discussion in the forum.

[SchumacherFM]

I've just started writing and ... how should I name the new page in the docs?

Dynamic Content

Is my current favourite but not really a nice one. Has someone a better idea = name?

[thomasmodeneis]

Dynamic Pages is how they call it on Middleman.

[spf13]

I like dynamic content

[SchumacherFM]

Documentation has been added!

SchumacherFM@4cd023c

Please report any weird stuff and also if it is ok to include future features (people will get more excited about hugo 😋 )

[bep]

The documentation should be about the current feature-set. Nobody knows what the future brings.

[SchumacherFM]

Moved it to another branch. Anything else ok?

[spf13]

I've merged this in. Excellent contribution @SchumacherFM. Looking forward to more.

[SchumacherFM]

Awesome! More stuff is coming.

[github-actions]

This pull request has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.