patch for install without rdkit

0.3.12

Update mol_preprocessor.py

support for python 3.6

What's Changed

• don't require tfrecord_features in preprocessor by @pstjohn in #22
• changes for python 36 by @pstjohn in #23

Full Changelog: 0.3.10...0.3.11

soft rdkit import

What's Changed

• soft rdkit import by @pstjohn in #21

Full Changelog: 0.3.9...0.3.10

Hook for ray tensorflow import

Minor fix for compatibility with ray and rllib

Quick fix for loading previous version json files

Full Changelog: 0.3.7...0.3.8

First release with zenodo DOI

Shree changes (#17)

* 1. xtb_preprocessing : creates a graph object with xtb features 
2. features :  incoporates the functions for bond and atom features from 
xtb
3. preprocessor: function call to incule a json file if it is passed 
from xtb calculations.

* 1. xtb_preprocessing : creates a graph object with xtb features 2. features :  incoporates the functions for bond and atom features from xtb 3. preprocessor: function call to incule a json file if it is passed from xtb calculations.

* 1. adding in examples for xtb json file
2. minor correction to features and xtb_preprocessing

* 1. updated to Mol Preprossor
2. Example showing the the diffrent features
3. added in tests (dont know how to check them though :/)

* 1. making different classes. 
2. addressing all other comments

* edits to the xtb test, a couple minor edits to the xtb and preprocessor classes

* Updating mol features, adding in bond distance and other features
Updated the xtb inputs as well

* Corrections to if else block

* Corrections to if else block

* Correct test_1 for test

* Correct test_1

* I think I removed >>>>>> Head

* test corrections

* As an array graph_features

* Update test_xtb_preprocessor.py

checks for the numpy array global feature and that the padding and batching operations work. Probably could have more informative features in the batching test, but this works for now

Co-authored-by: Peter St. John <[email protected]>

Padding fixes for connectivity

What's Changed

• A few fixes for rlmolecule and the redox models by @pstjohn in #14

Full Changelog: 0.3.5...0.3.6

Minor bugfix

quick edits for handling graphs without edges

New preprocessor inheritance model

Adds the pymatgen preprocessor, the RBFExpansion layer (Radial Basis Function), and a new inheritance model for preprocessor classes that leverages networkx.

What's Changed

• Issues with conda-forge tf2.6 by @pstjohn in #13

Full Changelog: 0.3.3...0.3.4

Additional support for tfrecords, reverting graph_layers

0.3.3

reverting graph_layers