Made several variants of the feature vector and added N shift prediction capacity.

Added another analysis method to XGBoost test 2 and added notes on it.

Increased database size to about 240 proteins, added all star files, and added XGBoost analysis scripts.

Added pH and temperature features and implemented XGBoost on trial set

Mini test data set constructed

Created pipeline that takes the AlphaFlow PDB plus Star File as input and outputs the feature vectors.

Added scripts for downloading star files and doing initial pH and temperature filter

All files involved in processing 34887

Processed one example structure (34887) from alphaflow structure to feature vectors.

good copy script for rotating the nearest neighbors into our invariant frame after they've been determined

One-hot encoding for atom type and amino acid type in training data

Changed nearest neighbor processing to include atom names, made a way to average the H to neighbor vectors across alphaflow models, and made a script to process that data into an initial format for training

Additional AlphaFlow output for testing

Made a script to split alphaflow output models into individual pdb files with added hydrogens and fixed C terminus

Eliminates homodimers now based on name of deposition - hopefully the dimers were all labelled as such...