Supporting v1 of PokeAPI

Question

Supporting v1 of PokeAPI

tdmalone opened this issue 6 years ago · 13 comments

I am also thinking to the v1 users. @phalt once said that a lot of calls are directed to the v1 endpoint. Is it possible to make ditto scan also the v1 endpoints?

Answer 1 · 2018-09-09T20:13:41.000Z

I think we should keep supporting it if it's not too difficult. We will never need to add new data to the v1 API, so if modifying ditto is too hard, we might be able to manually save the JSON and just keep it in the repo.

Answer 2 · 2018-09-09T20:39:39.000Z

It should be simple to tweak ditto to handle any arbitrary version of the API. I can take a look at doing that once I've got a PoC of the updater job.

…

On Sep 9, 2018, at 13:13, Charles Marttinen ***@***.***> wrote: I think we should keep supporting it if it's not too difficult. We will never need to add new data, so if modifying ditto is too hard, we might be able to manually save the JSON and just keep it in the repo. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.

Answer 3 · 2018-09-16T09:50:13.000Z

I've been looking at this and it looks like V1 is quite broken, for example https://pokeapi.co/api/v1/pokemon/32/. Is it really worth keeping v1 alive?

Answer 4 · 2018-09-16T09:58:57.000Z

Also a bunch of the build functions are broken. build_egg_pokes, build_type_pokes, build_move_descriptions, and build_pokedex_descriptions all throw exceptions.

IMO we should just drop v1 after a grace period, with intermittent outages to alert folks.

Answer 5 · 2018-09-16T22:24:33.000Z

I would generally agree with dropping v1 too... I'm not sure on the usage stats of it though.

But if it's broken.....!

Answer 6 · 2018-09-16T22:29:44.000Z

If it's broken, and nobody's complained yet, I say drop it. At this point, only very old projects will still be using it (it was deprecated in December 2014).

We can put in a permanent error message for any v1 API requests.

Answer 7 · 2018-09-17T08:47:06.000Z

Some stats I pulled from the live server:

V1 API hits daily ~1k
V2 API hits daily ~20k

If this was a business, I'd deprecate version 1. Cost of maintaining it isn't worth it.

Answer 8 · 2018-09-17T11:41:57.000Z

i like the idea of @cmmartti to drop the v1 and put a permanent error message for any v1 API requests, we know some very old project use the pokeapi v1 in pokeapi.co, but as it will not be in this domain (probably) I think we can only use v2.

Answer 9 · 2018-09-17T13:14:41.000Z

Just return 301 moved permanently responses. That's the responsible HTTP thing to do.

Answer 10 · 2018-09-21T21:57:59.000Z

With the new docs site now online, the deprecation of v1 has a public notification at https://pokeapi.co/docs/v1.html/

Answer 11 · 2018-09-21T22:03:56.000Z

Is 15 October a good date? I put it in as a placeholder because we never talked about it yet, but the outage meant the site went live earlier than expected.

Answer 12 · 2018-09-21T23:30:32.000Z

I have no objections to that - to be honest, I don’t think the date is going to make much of a difference, as anyone using v1 still probably isn’t checking the site anyway.

Answer 13 · 2018-09-22T04:48:21.000Z

Looks like we're dropping v1 support, closing.