Fixing a None error in contest ID selection

Refactor database query logic to remove Team_count filtering; improve handling of empty DataFrame cases in lineup generation

Fixing typeerror

Okay lets try this again

Adding an improved contest ID flow

Fix lineup query filtering and syntax; ignore Python cache files

fixing min/max stack sizes variable

reverting

Aiming to fix exposure spread which can duplicate in showdown because of the captain flex designation

Adding removal of 6-mn lineups from showdown pull

mlb_db, not nhl_db

fixing some object orientations in the projections import

sigh

Final change for MLB support (I hope)

fixing database queries call for MLB

Add MLB salary data handling and update file loading logic

I'm so annoyed

had the wrong collection labeled in showdown selection for MLB

more MLB support

Ping

Adding MLB support (for real this time)

implementing MLB support

Try/Except for opp map with MMA

adding optimization choice

Clean up of UI

Okay optimize params added (I think)

Mismatch in length of optimized and original

To add in specific players or specific avoidance for re-optimization

need to make sure we can run the exposure player multiple times in the exposure spread

okay for real this time

Fix condition for checking specific exclusions in session state

Specific Exclusiions never added to session state

rip

Oops, didn't declare specific exclusions before use

More work around locking exclusions

Testing some forced excelusions with exposure management

Making some adjustments to try to fix column swapping on reoptimization

Revert

Add warning for NBA Late Slate database load issue

fix index sort

Continuing to improve late swap

Updaing for Cause: row comes from replace_rows.index and player_rows.index, so it’s a index label (e.g. 10, 11, 12 after a slice), not a positional index 0..n-1. The code was using working_frame.iloc[row], which is positional. If working_frame had a non‑contiguous or offset index (e.g. from a slice or after optimize_lineups), a label like 12 could be valid while iloc[12] was out of bounds for a shorter frame.

Aiming to make sure higher medians are maintained

Instead of using the previous lineup, we use if the lineup has been used.

Fixed out, man.

Removing the MIP and cascading logic for a much simpler drop-one insert.

reverting

Okay, this should work now.Cascading optimal is implemented.

Lets see if this works

Reverting