Normalized name missing from FleaMarket

[the-hideout]

Normalized name missing from FleaMarket

[Razzmatazzz]

This regex isn't being run every time there's an API call; it's being run when item prices are updated and data is uploaded to the cloudflare KVs for the API to use:
https://github.com/the-hideout/tarkov-data-manager/blob/9e5472004784624c75e3ee0b3a83ec1598586623/src/tarkov-data-manager/jobs/update-item-cache.js#L471-L477

[Shebuka]

To add: the FleaMarket name is all CAPS, I think it must be like others with only the first letter capital

[GrantBirki]

I'm not sure if there is a historical reason for why we have normalizedName: String excluded from the FleaMarket item. If there isn't a good reason for why this is missing today, then I agree and we should absolutely add it! :)

As for the type name, I think it is correct being FleaMarket. Looking at the schema, many other types have this same case definition. See the examples below:

[Shebuka]

Apologise for wording it wrongly, it's all caps what is returned in the name field by the API, I see "name": "БАРАХОЛКА",

[Shebuka]

p.s. Question: if the flea market is disabled do we still get the flea market in sellto/buyfrom but with enabled: false?

[Razzmatazzz]

We can add normalizedName to the FleaMarket type. It's missing because normalizedName was a recent addition to the traders and it just wasn't added to the FleaMarket type at the time.

The flea market name being all in caps is due to the lowercase regex not properly picking up Cyrillic uppercase characters. Is барахолка a more appropriate capitalization?

[Shebuka]

The flea market name being all in caps is due to the lowercase regex not properly picking up Cyrillic uppercase characters. Is барахолка a more appropriate capitalization?

yes, but with the first letter capital like for all other traders, Барахолка

[Razzmatazzz]

This is the updated code that's being used to convert the uppercase "FLEA MARKET" to "Flea Market" for all languages:

UPPERCASE_STRING.replace(/(?<!\b)([A-ZА-Я])/g, substr => {
    return substr.toLowerCase();
});

What it's supposed to do is capture any upper case letters that aren't at the beginning of a word and convert them to lower case. I added А-Я to capture Cyrillic upper case characters, but it seems to also be capturing the first upper case letter that's at the beginning of the Russian word.

[Razzmatazzz]

I think I figured it out to just return the match if it's at index 1: Барахолка
Apparently issue stems from unicode characters, and the pattern to get a correct match is (?<!^|\s)([A-ZА-Я])

[Shebuka]

As we load these names for a particular lang can't we simply update the translation to the correct case? (just want to understand how it works, we have Therapist and Терапевт in the correct case, why not the Flea?)

[Shebuka]

I was just testing other languages for Flea Market translations and in cz it's returned as BleŠÍ Trh but google returns Bleší Trh (in the first word the last two letters are not capital) and in fr MarchÉ, so I suppose the regex is not working for any language other than one that use only basic Latin chars

[Razzmatazzz]

As we load these names for a particular lang can't we simply update the translation to the correct case? (just want to understand how it works, we have Therapist and Терапевт in the correct case, why not the Flea?)

That's exactly what we're doing. In the localization files we have, the only string for "Flea Market" we have is in all upper case characters, so it has to be converted. Digging a bit more, I believe the proper regex to match all uppercase unicode characters that aren't at the start of a word is /(?<!^|\s)\p{Lu}+/gu.

[Shebuka]

This regex seems to work perfectly. 👍

But why we can't update the localization files directly and save regexs CPU cycles?