check out my other project! a discord utility bot which redeems giftcodes for all your members
all the content in this repo so far has been obtained by me through in game screenshots or manual data entry
current contents:
- chief gear images (only has up to mythic t2 3-star)
- chief gear costs (only has up to mythic t2 3-star)
- chief charm images
- chief charm costs
- fire crystal amounts (only has up to FC5)
- fire crystal images (up to FC5)
chief gear images are organized by rarity into folders with an identifier. for example:
- Mythic T2 3-Star:
mythic_t2_3 - Epic T1:
epic_t1_0 - Rare:
rare_0
in chief_gear.json, each key is an identifier for a rarity of chief gear. the data stored with that key contains the proper name of the rarity, and how much hardened_alloy, polishing_solution, design_plans, and amber (only for >legendary gear, which is not included in this database for now) is needed to upgrade to that rarity.
it also contains the power and stat bonuses for each rarity of chief gear.
chief charm images are organized by their type, e.g. keenness (lancer stat charm), protection (infantry stat charm), or vision (marksman stat charm)
they are sorted into folders and are named by their level.
each key in chief_charm.json are the level of the charm. they contain the cost of the charm in charm_design(s) and charm_guide(s), their stat percentages, and the amount of power they give you.
i used this as an opportunity to learn how to do image recognition with machine learning. using the ultralytics python module, i trained a yolo11n model to recognize chief gear from a set of screenshots (get the model here and the example code here)
then, i used the trained model to recognize chief gear in a folder of 156 screenshots of chief gear to extract cropped images of each piece.
this one does not have as fun of a story as the obtaining of chief gear images. i just took screenshots of every charm and wrote a python script to remove the backgrounds, and trim it, replacing it with