Core: add correct flags to precollected items

[kindasneaki]

What is this fixing or adding?

precollected items just had a default flag

How was this tested?

ran generate and multiserver to test the changes

Why I added this

While the flags aren't being used anywhere at the moment I am working on a web based text client that I would like to use the flags to correctly color the precollected items. I think the Archipelago tracker could eventually use this information as well.

[Berserker66]

With only looking at it, and not testing it, I'm pretty sure this will fail to load old multidata files.

[alwaysintreble]

Tested it, and the flags work correctly on receipt, but yeah it does fail to load old multidata.

[kindasneaki]

Im not sure if this is the greatest way to do it. If not can I get some info on a better way?

[black-sliver]

Generating a seed with this and uploading to archipelago.gg will fail as well. You would have to bump the version to 0.4.5 and set multidata's ["minimum_versions"]["server"] to 0.4.5.

I would instead suggest adding decoded_obj["precollected_flags"] and use zip() to iterate over both items and flags at the same time.

--
also this has conflicts now :(

--
i would like to have this in 0.4.5 fwiw

[kindasneaki]

Generating a seed with this and uploading to archipelago.gg will fail as well. You would have to bump the version to 0.4.5 and set multidata's ["minimum_versions"]["server"] to 0.4.5.

This doesn't fail when you upload to archipelago.gg but the TODO was outdated for sure 😐

I would instead suggest adding decoded_obj["precollected_flags"] and use zip() to iterate over both items and flags at the same time.

I'm not quite sure what you mean with doing this

[black-sliver]

This doesn't fail when you upload to archipelago.gg

I have not tested it, but

if items and isinstance(items[0], tuple):

this code does not exist on current archipelago.gg, so by generating a multidata that has a tuple there, the current WebHost will run the old code path, which does

self.start_inventory[slot] = [NetworkItem(item_code, -2, 0) for item_code in item_codes]

where item_code is now a tuple instead of a number, breaking the game.

The way to solve this imo is by providing two separate entries in the multidata:
one is just the codes (as before, fully back compat), and another is just the flags. To get nice-looking code, you can zip the two lists using https://docs.python.org/3/library/itertools.html#itertools.zip_longest , something like

for slot, items in decoded_obj["precollected_items"].items():
    for item_code, item_flags in zip_longest(items, decoded_obj.get("precollected_flags", {}).get(slot, []), fillvalue=0):
        ...