Add Dockerfile + instructions on how to preview site using docker rather than installing hugo locally

[alamb]

In order to make changes to the website and have confidence that we won't break things we should make sure we can see the results of the work locally.

I can't / don't want to try and figure out how to get a local hugo install running locally, and prefer to use docker.

I figured these instructions might help others

BTW I am happy to make a JIRA for this PR, but it isn't clear to me if that is desired or not in Parquet

[wgtmac]

+1

Thanks for the improvement! My only concern is that these steps may be out of sync easily (e.g. when the provided URLs are broken).

[wgtmac]

Is it better to use a dockerfile which is much easier to use? I'm just asking but not required to change. These steps are helpful enough.

[alamb]

A docker file is a good idea. I will make one

[alamb]

Thanks for the improvement! My only concern is that these steps may be out of sync easily (e.g. when the provided URLs are broken).

Perhaps we can update the instructions over time if/when they become broken? I am sure there are better ways to make such scripts, but in my opinion this is a step in the right direction

[alamb]

Thanks for the review @wgtmac -- I have implemented your suggestion and created a Dockerfile and updated the instructions to use them.

[vinooganesh]

This is a great suggestion and the timing is right. I spend some time a few weeks ago moving the parquet site's docsy dependency to a hugo module, so now they can be managed separately.

The other thing that we haven't been doing a good job of is maintaining the staging website. I made a bunch of changes to get the staging and production branch in sync, but staging still isn't heavily used.

[alamb]

The other thing that we haven't been doing a good job of is maintaining the staging website. I made a bunch of changes to get the staging and production branch in sync, but staging still isn't heavily used.

I wonder what the usecase for the staging website is? (maybe we should just not use it?)

FWIW for https://arrow.apache.org/ and https://datafusion.apache.org/ we simply publish to the production version of the site.

Sometimes the staging site might be helpful to host pre-release api docs or something, but I didn't see any on this site 🤔

[vinooganesh]

Good question @alamb. Technically the "best practice" from the docsy instructions were to create a staging website so I mostly just followed them when I remade the parquet one. Back then, there was a lot of stuff to work through with hugo builds and migrating from the old jenkins site, so having a place to test was definitely helpful. At this point though, I don't think it's necessary to have the staging site anymore.

[wgtmac]

For the staging site, I had a discussion with @gszadovszky here: #31 (comment). I think we can remove the staging site now and use the docker file for debug purpose.

[alamb]

Thanks @wgtmac and @vinooganesh