search

any2cards / worldhaven

All of the JSON data files and Asset image files that the WAV utilizes
94 stars 24 forks source link

Add ability card data #4

Closed Zachu closed 2 years ago

Zachu commented 2 years ago

Discussion in https://github.com/any2cards/worldhaven/issues/3

Since the values are read in OCR there might be some mistakes. These should be fixed by adding them to scripts/tesseract/manual-fixes-character-ability-cards.jsonl and merging these too in the data/character-ability-cards.js by using scripts/merge_jsonl_to_json.sh

Zachu commented 2 years ago

I don't know if someone would like to verify the data additions somehow before merging

any2cards commented 2 years ago

Just an fyi, but because this repository is relied on by multiple sources, I am going to have to verify everything before I initiate a merge, as I don’t want to break any pre-existing tools, apps, etc.

Sent from my iPhone

On Apr 5, 2022, at 7:14 AM, Jani Korhonen @.***> wrote:

 I don't know if someone would like to verify the data additions somehow before merging

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.

Zachu commented 2 years ago

Just an fyi, but because this repository is relied on by multiple sources, I am going to have to verify everything before I initiate a merge, as I don’t want to break any pre-existing tools, apps, etc.

Would not expect anything less =) Rather what I meant was the new additions if someone else wants to do some verification whether the OCR has been correct and that sort of things.

There was some problems in the not-so-common entries like the card levels A/B (which I now actually read from the filenames since they happen to be there) and the level P whatever that means :sweat_smile:

Also I decided to take just the number part from card names, some having a prefix of "AA-" for example. This limits the character space the tesseract model needed to learn. If they need to be added I would probably script it rather than try to read it with OCR.

any2cards commented 2 years ago

There is no need to process this pull request, as the requested data was added via another process.