Merge pull request #152 from ATOMScience-org/1.4.0

1.4.0

Author: mauvais2

README.md

@@ -26,6 +26,8 @@ Documentation in readthedocs format is available [here](https://ampl.readthedocs
 - [Getting started](#Getting-started)
   - [Prerequisites](#Prerequisites)
   - [Install](#Install)
+     - [Clone the git repository](#clone-repo)
+     - [Create conda environment](#create-conda-env)
      - [dgl and CUDA (**optional**)](#Install-dgl)
   - [Install with Docker](#Install-docker)
   - [AMPL usage survey](#AMPL-usage-survey)
@@ -62,21 +64,27 @@ AMPL is a Python 3 package that has been developed and run in a specific conda e
 
 <a name="Install"></a>
 ### Install
+<a name="clone-repo"></a>
 #### Clone the git repository
 
 `git clone https://github.com/ATOMScience-org/AMPL.git`  
 &nbsp;  
 
 #### Create conda environment
+<a name="create-conda-env"></a>
+There is a script `create_ampl_env.sh` that will automate the conda env creation and the packages install for you.
+
+```
+$ source create_ampl_env.sh
 ```
-cd conda
 
-conda create -y -n atomsci --file conda_package_list.txt
+It will 
 
-conda activate atomsci
-
-pip install -r pip_requirements.txt
-```
+- Prompt for a new conda environment name
+- Then run
+ - conda create 
+ - conda activate 
+ - and pip install and other fixes
 
 - *Note: Depending on system performance, creating the environment can take some time.*  
 &nbsp;  

VERSION

@@ -1 +1 @@
-1.3.0
+1.4.0

atomsci/ddm/docs/PARAMETERS.md

@@ -515,10 +515,58 @@ The AMPL pipeline contains many parameters and options to fit models and make pr
 
 |||
 |-|-|
-|*Description:*|Type of splitter to use: index, random, scaffold, butina, ave\_min, temporal, fingerprint, or stratified. Used to set the splitting.py subclass. Can be input as a comma separated list for hyperparameter search (e.g. 'scaffold','random')|
+|*Description:*|Type of splitter to use: index, random, scaffold, butina, ave\_min, temporal, fingerprint, multitaskscaffold, or stratified. Used to set the splitting.py subclass. Can be input as a comma separated list for hyperparameter search (e.g. 'scaffold','random')|
 |*Default:*|scaffold|
 |*Type:*|str|
 
+- **mtss\_num\_super\_scaffolds**  
+  
+|||
+|-|-|
+|*Description:*|This specifies the number of genes in a chromosome for the genetic algorithm. Scaffolds bins are often very small and only contain 1 compound. Scaffolds are therefore combined into super scaffolds to the number of genes and also reduce complexity and runtime.|
+|*Default:*|40|
+|*Type:*|int|
+
+- **mtss\_num\_generations**  
+  
+|||
+|-|-|
+|*Description:*|The number of generations the genetic algorithm will run.|
+|*Default:*|20|
+|*Type:*|int|
+
+- **mtss\_num\_pop**  
+  
+|||
+|-|-|
+|*Description:*|Size of population per generation in the genetic algorithm.|
+|*Default:*|100|
+|*Type:*|int|
+
+- **mtss\_train\_test\_dist\_weight**  
+  
+|||
+|-|-|
+|*Description:*|How much weight to give the tanimoto distance between training and test partitions.|
+|*Default:*|1.0|
+|*Type:*|float|
+
+- **mtss\_train\_valid\_dist\_weight**  
+  
+|||
+|-|-|
+|*Description:*|How much weight to give the tanimoto distance between training and valid partitions.|
+|*Default:*|1.0|
+|*Type:*|float|
+
+- **mtss\_split\_fraction\_weight**  
+  
+|||
+|-|-|
+|*Description:*|How much weight to give adherence to requested subset franctions.|
+|*Default:*|1.0|
+|*Type:*|float|
+  
 ---
 
 <a name="Transformers"></a>