search

raquo / hnapp

Hacker News faceted search engine, RSS & JSON feeds
http://hnapp.com
MIT License
60 stars 10 forks source link

Planning to terminate hnapp in late 2024. #11

Open raquo opened 1 year ago

raquo commented 1 year ago

Hey everyone, hnapp in its current incarnation has been up and running for 9 years now.

It has served us well, well me at least, and perhaps other people, since someone always notices when it goes down because it's overfilled its disk again (patent pending distributed server monitoring system, just you wait), but well, I haven't touched this code in forever, and if I want to run it anywhere other than its dusty old server, I need to port it to python 3, update all dependencies, set up its hosting anew, and I just don't have the energy for any of that. And, I realized that if hnapp went away, I wouldn't be all that sad. Perhaps it's better to let it go while it's still nice, before it becomes a chore.

Next year this version of hnapp will turn 10 years old (original is 12 years old already), and that's as good a time as any to pull the plug.

I don't run any analytics, and don't collect any personal data, so I don't know how many people still use hnapp, nor do I have a good way to contact any of them. I guess I'll put a banner on the website advising users of the impeding shutdown at some point.

If anyone wants to help upgrade hnapp to python 3 and make it hostable in a docker container, or something like that, please let me know.

PeterDaveHello commented 2 months ago

Hey, just wondering if there might be a chance or possibility to use LLM (AI) to help keep this project going? By the way, I just discovered this project and it looks really awesome and useful!

raquo commented 2 months ago

Heyo, in short – I don't think so. I'm not sure what problems an LLM would solve here, and also, my goal with any changes would be to reduce the cost, complexity and maintenance burden – whereas an LLM would bring the opposite.

I think one option could be to do away with compiling our own database, and just query the hn.algolia.com API for every request. I'm not sure if their query language is fully compatible with ours, but it could be close enough.

PeterDaveHello commented 1 month ago

Oh, I actually meant upgrading to Python 3. I think there are some existing tools, and with the help of LLM, it might make the upgrade less difficult. As for Algolia's API, I'm not sure if it's fully compatible either, but their API is pretty powerful and a lot of other projects use it, so it should be doable!

PeterDaveHello commented 1 month ago

Some reference might be helpful:

  1. 2to3 https://docs.python.org/3/library/2to3.html

  2. Python-Modernize https://github.com/python-modernize/python-modernize

  3. Python-Future https://github.com/PythonCharmers/python-future

  4. Six https://github.com/benjaminp/six

raquo commented 1 month ago

Ah, yes that makes more sense. Although, whoever does the AI upgrade needs to also review and manually test such AI changes, as if it was their own code.