Add support for Pymatgen StructureGraph with different local_env strategies

.gitignore

@@ -93,3 +93,7 @@ Local
 
 # .DS_Store
 .DS_Store
+
+data/atom_init_rev.json
+
+data/readme.txt

README.md

@@ -48,25 +48,33 @@ Please cite the following work if you want to use CGCNN.
 
 This package requires:
 
-- [PyTorch](http://pytorch.org)
-- [scikit-learn](http://scikit-learn.org/stable/)
-- [pymatgen](http://pymatgen.org)
+- [PyTorch](http://pytorch.org) (tested on v.1.4.0)
+- [PyTorch Scatter](https://github.com/rusty1s/pytorch_scatter)
+- [scikit-learn](http://scikit-learn.org/stable/) (tested on v.0.22.1)
+- [pymatgen](http://pymatgen.org) (tested on v.2020.3.13)
 
-If you are new to Python, the easiest way of installing the prerequisites is via [conda](https://conda.io/docs/index.html). After installing [conda](http://conda.pydata.org/), run the following command to create a new [environment](https://conda.io/docs/user-guide/tasks/manage-environments.html) named `cgcnn` and install all prerequisites:
+If you are new to Python, the easiest way of installing the prerequisites is via [conda](https://conda.io/docs/index.html) and `pip`. After installing [conda](http://conda.pydata.org/), run the following command to create a new [environment](https://conda.io/docs/user-guide/tasks/manage-environments.html) named `cgcnn` and install all prerequisites:
 
 ```bash
-conda upgrade conda
-conda create -n cgcnn python=3 scikit-learn pytorch torchvision pymatgen -c pytorch -c conda-forge
+conda create -n cgcnn python=3
+source activate cgcnn
+conda install pytorch torchvision cudatoolkit=10.1 -c pytorch 
+pip install torch-scatter==latest+cu101 -f https://pytorch-geometric.com/whl/torch-1.4.0.html
+conda install scikit-learn
+pip install pymatgen
 ```
-
-*Note: this code is tested for PyTorch v1.0.0+ and is not compatible with versions below v0.4.0 due to some breaking changes.
-
 This creates a conda environment for running CGCNN. Before using CGCNN, activate the environment by:
 
 ```bash
 source activate cgcnn
 ```
 
+finally once inside the environment install torch-scatter
+
+```bash
+pip install torch-scatter
+```
+
 Then, in directory `cgcnn`, you can test if all the prerequisites are installed properly by running:
 
 ```bash
@@ -189,7 +197,7 @@ To reproduce our paper, you can download the corresponding datasets following th
 
 ## Authors
 
-This software was primarily written by [Tian Xie](http://txie.me) who was advised by [Prof. Jeffrey Grossman](https://dmse.mit.edu/faculty/profile/grossman). 
+This software was primarily written by [Tian Xie](http://txie.me) and [Prof. Jeffrey Grossman](https://dmse.mit.edu/faculty/profile/grossman). This version contains [changes made by Rhys Goodall](https://github.com/txie-93/cgcnn/pull/16) to remove zero-padding. This version also contains several changes made by [Andrew S. Rosen](https://asrosen.com/) to enhance the usability of the CGCNN code, including but not limited to: new command-line arguments, saving the CIF data as .json files so they don't have to be regenerated on-the-fly, and [the option](https://github.com/txie-93/cgcnn/pull/17) to use crystal graphs generated by Pymatgen. If using the `--enable-tanh` flag, you should consider referencing the work of [Noh et al.](https://pubs.acs.org/doi/10.1021/acs.jcim.0c00003), who called this approach CGCNN-H.
 
 ## License