Psi 4 integration

.coveragerc

@@ -0,0 +1,33 @@
+[paths]
+source =
+  moldesign
+  */site-packages/moldesign
+  */dist-packages/moldesign
+
+
+[run]
+source = moldesign
+omit =
+   moldesign/_notebooks/*
+   moldesign/_tests/*
+   moldesign/_static_data/*
+   moldesign/_version.py
+   moldesign/external/*
+   moldesign/widgets.py
+   moldesign/interfaces/nbo_interface.py
+   moldesign/utils/_deadfunctions.py
+   moldesign/utils/docparsers/google.py
+
+[report]
+exclude_lines =
+    pragma: no cover
+    def __repr__
+    def __str__
+    def _repr_markdown_
+    raise AssertionError
+    raise NotImplementedError
+    if PY2
+    if future.utils.PY2
+    def show
+    def unshow
+

.github/ISSUE_TEMPLATE.md

@@ -0,0 +1,16 @@
+<!-- Delete the text below if you're not filing a bug report -->
+## Bug report
+#### Description
+<!-- A quick description of what's going wrong -->
+
+#### Steps to reproduce
+<!-- If possible, show us how to reproduce the bug, INCLUDING any error messages you receive -->
+
+```python
+### code to reproduce the bug goes here ###
+```
+
+#### Environment
+  - Operating system: <!-- e.g., "MacOS 10.12.3" or "Debian 8.8"-->
+  - `python --version` output: <!-- e.g., "2.7.13" or "Python 3.5.2 :: Continuum Analytics, Inc."-->
+  - `pip show moldesign nbmolviz | grep -B1 Version` output:

DEVELOPMENT.md

@@ -1,32 +1,33 @@
 # DEVELOPING MDT
 
 ### Setting up a dev environment
-(WIP)
+(still under construction)
 
-### Install prerequisites (first time only)
-To make these instructions reasonably cross-platform, we'll use pyenv as the python environment manager. 
+### Install prequisites (first time only)
+You need to install docker, and an environment manager for Python 3 (Miniconda 3). Here's one way to do that:
 1. Install docker: [link]
 2. Install pyenv and pyenv-venv: [link]
-3. Install miniconda2 by running: `pyenv install miniconda2-latest`
-4. Switch to miniconda environment by running: `pyenv shell miniconda2-latest`
+3. Install miniconda3 by running: `pyenv install miniconda3-latest`
+4. Switch to miniconda environment by running: `pyenv shell miniconda3-latest`
 
 ### Set up your environment (first time only)
-1. Change directory to the base of the `molecular-design-toolkit` repository
+1. Get MDT: `git clone http://github.com/Autodesk/molecular-design-toolkit`
+1. `cd molecular-design-toolkit`
 1. Create conda environment (optional but recommended) by running: [command to create conda env]
-2. Activate the environment: `pyenv shell [environment name???]`
-1. Install dev dependencies: `pip install -r requirements.txt DockerMakefiles/requirements.txt moldesign/_tests/requirements.txt`
+2. Activate the environment: `pyenv activate [environment name???]`
+1. Install dev dependencies: `pip install -r requirements.txt DockerMakefiles/requirements.txt deployment/requirements.txt`
 2. Set up for local dev mode (this tells MDT to use your local docker containers):
 ```bash
 mkdir ~/.moldesign
 echo "devmode: true" > ~/.moldesign/moldesign.yml
 ```
-8. Link your installation within your environment
+8. Install MDT in "development mode": 
 ```
 pip install -e molecular-design-toolkit
 ```
 
 ### To activate environment (in any new shell)
-1. Run `pyenv shell [environment name???]`
+1. Run `pyenv activate [environment name???]`
 
 ### To rebuild docker images (first time and after changes that affect dockerized code)
 5. Build development versions of all docker images:
@@ -87,9 +88,18 @@ are considered "stable".
 
 ### Releases
 
-1. Decide on the new version number (using [semantic versioning](http://semver.org/)). For our purposes here, we'll pretend it's `0.9.3`.
+1. Decide on the new version number (see below). For our purposes here, we'll pretend it's `0.9.3`.
 1. Tag the relevant commit (the build must be passing) with a release candidate version number, e.g., `0.9.3rc1`.
-1. Codeship will automatically deploy the updated release to PyPI and DockerHub (still WIP 5/17)
+1. Codeship will automatically deploy the updated release to PyPI and DockerHub
+1. Manually test the example notebooks against this pre-release version.
+1. If succesful, tag the relevant commit with the official release version `0.9.3`
+
+### Versioning
+For now, we're using a subset [PEP 440](https://www.python.org/dev/peps/pep-0440/):
+1. Every release should be of the form MAJOR.MINOR.PATCH, e.g. `0.1.2`
+2. Pre-releases should be numbered consecutively, and may be alpha, beta, or "release candidate", e.g. `1.0.1rc3` or `0.5.3a1`
+3. Our deployment infrastructure uses this regular expression to accept version strings:
+`^(0|[1-9]\d*)\.(0|[1-9]\d*)\.(0|[1-9]\d*)((a|rc|b)(0|[1-9]\d*))?$`
 
 ### Maintainers: updating the documentation
 

DockerMakefiles/DockerMake.yml

@@ -114,11 +114,11 @@ deploy_requirements:
 
 python_deploy_base_py2:
   requires:
-     - miniconda
+     - miniconda_py2
      - deploy_requirements
 
 python_deploy_base:
   requires:
-    - miniconda_py2
+    - miniconda
     - deploy_requirements
 

DockerMakefiles/Moldesign.yml

@@ -17,7 +17,8 @@ moldesign_minimal_requirements:
    && apt-get install -y gcc \
    && pip install -r /tmp/mdtreqs.txt
   COPY . /opt/molecular-design-toolkit
-  RUN pip install -e /opt/molecular-design-toolkit
+  WORKDIR /opt/molecular-design-toolkit
+  RUN python setup.py sdist && pip install `ls -t dist/*.tar.gz | head -n1`
 
 moldesign_minimal:
   requires:

DockerMakefiles/NWChem.yml

@@ -29,7 +29,7 @@ nwchem_build:
  build: |
     RUN         apt-get install -y patch curl perl
     WORKDIR      /opt
-    ENV          NWCHEM_MODULES="nwdft driver stepper" \
+    ENV          NWCHEM_MODULES="nwdft driver stepper mm qmmm nwmd esp " \
                    PYTHONVERSION=2.7 \
                    PYTHONHOME="/usr" \
                    USE_PYTHONCONFIG=Y \
@@ -45,25 +45,32 @@ nwchem_build:
                    && cd nwchem \
                    && svn co --depth empty https://svn.pnl.gov/svn/nwchem/trunk/src \
                    && cd src \
-                   && svn update GNUmakefile nwchem.F config \
-                        tools include basis geom inp input  \
+                   && svn update GNUmakefile nwchem.F config mm qmmm nwmd \
+                        tools include basis geom inp input qhop fft cafe \
                         pstat rtdb task symmetry util peigs perfm bq cons blas lapack \
-                        NWints atomscf cphf ddscf driver gradients hessian nwdft optim property stepper symmetry vib \
+                        NWints atomscf cphf ddscf driver gradients hessian nwdft \
+                        optim property stepper symmetry vib space esp \
                    && rm -rf /opt/nwchem/.svn \
                    && cd tools \
                    && svn export --non-interactive --username nwchem --password nwchem \
                           https://svn.pnl.gov/svn/hpctools/branches/ga-5-5  \
                    && cd .. \
-                   #strip native compiler options as Fedora does
                    && sed -i 's|-march=native||' config/makefile.h \
                    && sed -i 's|-mtune=native|-mtune=generic|' config/makefile.h \
                    && sed -i 's|-mfpmath=sse||' config/makefile.h \
-                   && sed -i 's|-msse3||' config/makefile.h  \
-                   && make nwchem_config && make -j3  \
-                   #clean unnecessary source to reduce docker size
-                   && rm -rf tce tools nwdft NWints geom symmetry util nwxc ddscf lapack blas rism \
-                          argos peigs rmdft gradients symmetry property smd lucia dplot propery \
-                          hessian ccsd mp2_grad moints cafe analyz dimqm /opt/nwchem/lib
+                   && sed -i 's|-msse3||' config/makefile.h
+
+    RUN  cd /opt/nwchem/src \
+         && make nwchem_config \
+         && make -j 8
+
+    RUN mkdir nwchem/contrib \
+       && cd nwchem/contrib \
+       && svn co https://svn.pnl.gov/svn/nwchem/trunk/contrib/mov2asc \
+       && cd mov2asc \
+       && FC=gfortran make \
+       && mv mov2asc asc2mov /opt/nwchem/bin/LINUX64
+
     RUN git clone https://github.com/NWChem-Python/nwapi \
        && cd nwapi \
        && make -f makefile.linux \
@@ -76,19 +83,62 @@ nwchem:
   description: Deployable NWChem image
   build_directory: buildfiles/nwchem/
   requires:
-    - python_deploy_base
+    - python_deploy_base_py2
     - nwchem_requirements
   build: |
-    COPY run.sh run.py getresults.py /usr/bin/
-    RUN chmod +x /usr/bin/getresults.py /usr/bin/run.py /usr/bin/run.sh
-    RUN pip install pint
+    RUN pip install pint cclib
     ENV NWCHEMFILE="From https://svn.pnl.gov/svn/nwchem/trunk/src"
-    ENV PYTHONPATH=$PYTHONPATH:/opt
+    ENV PYTHONPATH=$PYTHONPATH:/opt/nwapi
+    COPY run.py runqmmm.py getresults.py /usr/local/bin/
+    RUN chmod +x /usr/local/bin/getresults.py \
+     /usr/local/bin/run.py \
+     /usr/local/bin/runqmmm.py
   copy_from:
     nwchem_build:
       /usr/local/lib/python2.7/dist-packages: /usr/local/lib/python2.7/
       /opt/nwchem: /opt
       /opt/lib: /opt
-  copy_from:
+      /opt/nwapi: /opt
     openblas:
       /opt/lib: /opt
+
+#################################
+# Experimental QM/MM images below
+nwchem_amber_qmmmm:
+  description: Image with Amber and NWChem for QM/MM
+  requires:
+     - nwchem
+  build: |
+    RUN apt-get update \
+     && apt-get install -y --no-install-recommends \
+         gcc \
+         gfortran \
+     && cleanapt
+  copy_from:
+    amber_dev_branch_build:
+      /opt/amber16: /opt
+      /root/.bash_profile: /root
+
+
+amber_dev_branch_build:
+  build_directory: buildfiles/nwchem
+  requires:
+    - buildbase
+    - ambertools_requirements
+  build: |
+    ADD amber-dev.tgz /opt
+    RUN apt-get install -y \
+         flex \
+         bison \
+         csh \
+         gfortran \
+         g++ \
+         make \
+         patch \
+         python-dev
+    WORKDIR /opt/amber16
+    RUN echo N | ./configure -noX11 --with-python /usr/bin/python --with-netcdf /usr/ gnu \
+     && make -j6 install \
+     && rm -rf test AmberTools doc  # reduce image size for copying to deployment image
+    RUN echo "test -f /opt/amber16/amber.sh && source /opt/amber16/amber.sh" >> /root/.bash_profile
+

DockerMakefiles/PythonTools.yml

@@ -14,6 +14,9 @@ notebook:
      && pip install backports.ssl_match_hostname backports.shutil_get_terminal_size
 
 
+psi4:
+  build: conda install -c psi4 psi4
+
 pyquante2:
  requires:
   - python_deploy_base
@@ -27,13 +30,17 @@ chem_python_conda:
  description: All of the external python chemistry libraries in one place
  build: |
    RUN conda install -qy -c openbabel openbabel=2.4.1
-   RUN conda install -qy -c omnia biopython openmm parmed pdbfixer
-
+   RUN conda install -qy -c omnia \
+      biopython=1.68 \
+      openmm=7.1.1 \
+      parmed=2.7.3 \
+      pdbfixer=1.4
 
 chem_python_requirements:
   requires:
     - pyscf
     - chem_python_conda
+    - psi4
 
 chem_python:
  requires:

DockerMakefiles/README.md

@@ -1,21 +1,51 @@
 # Docker images for MDT
 
-The Molecular Design Toolkit uses a lot of functionality from other open source chemistry packages, such as molecular dynamics from OpenMM, Quantum chemistry from OpenMM, etc., etc.
+The Molecular Design Toolkit does NOT implement molecular modeling algorithms - it instead provides an intuitive, general interface for running calculations with other open source chemistry packages, such as molecular dynamics from OpenMM, Quantum chemistry from PySCF, etc., etc.
 
-To make these calculations 1) easy, 2) portable, and 3) reproducible, users don't need to compile any of these by themselves. Instead, they're provided as docker images, which (as of this writing) are freely available from a Dockerhub repository (see https://hub.docker.com/r/autodesk/moldesign/). Users probably won't need to pull the images manually - they'll be automatically pulled whenever they're needed.
+To make these calculations 1) easy, 2) portable, and 3) reproducible, users don't need to compile any other software. Instead, external open source packages are provided as docker images, which MDT automatically downloads from a public [DockerHub repository](https://hub.docker.com/r/autodesk/moldesign/).
 
+These images are built from definitions contained in the DockerMakefiles in this directory (everythign with a `.yml` extension).
 
-# About the images
+MDT relies on two types of dependencies:
 
-`docker-make` is used to build and push the images to the public repositories. The Dockerfiles for all images are defined in the `DockerMake.yml` and its associates YAML sources in this directory.
+1. CLI executables (e.g., NWChem, AmberTools utilities, etc.). CLI executables come in their own lightweight images. These images let you run the executables without compiling or installing them yourself.
+2. Python libraries (OpenMM, PySCF, etc.). To interact with these, we build a rather large docker image named `moldesign_complete` that includes MDT _and_ all interfaced Python libraries. When MDT needs to use a Python library that isn't installed on your machine, it will run the necessary code in a `moldesign_complete` docker container instead.
 
-If you're just *using* MDT, you'll be relying on the prebuilt MDT images hosted on [DockerHub](https://hub.docker.com/r/autodesk/moldesign/). However, if you are developing MDT, you may need to install docker-make by running `pip install dockermake>=0.5`. 
 
+# Development and building images
 
-MDT relies on two types of dependencies:
+**NOTE**: If you're just *using* MDT, you don't need to do this - MDT will automatically download the images hosted on [DockerHub](https://hub.docker.com/r/autodesk/moldesign/). 
+
+##### Install DockerMake:
+We use the `docker-make` command line utility to build and manage MDT's docker images. To install it:
+```bash
+pip install "DockerMake>=0.5.6"
+```
+
+##### Activate MDT "devmode":
+To tell MDT to use your own, locally built docker images, run:
+```bash
+echo "devmode:true" >> ~/.moldesign/moldesign.yml
+```
+
+Use the following commands to build and manage the docker images. They should be run in this directory (`[...]/molecular-design-toolkit/DockerMakefiles`)
+
+##### List docker images:
+```bash
+docker-make --list
+```
 
-1. CLI executables (e.g., ambertools utilities like antechamber, tleap). CLI executables come in their own lightweight images.
-2. Python libraries (openmm, pyscf, etc.). To interact with these, we build a rather large docker image named "moldesign_complete" that includes both moldesign and all of its dependencies. This allows MDT to call functions from libraries that aren't installed on your machine! Instead, those function calls will be evaluated in the moldesign_complete docker image, and the results will be returned.
+##### Build all docker images:
+```bash
+docker-make --tag dev --all
+```
 
-In the future, we plan to isolate the python dependencies into smaller, lightweight installations that don't include MDT. This will require a common data format with translation layers for each dependency.
+##### Build the MDT docker image for python 3:
+```bash
+docker-make --tag dev moldesign_complete
+```
 
+##### Build the MDT docker image for python 2:
+```bash
+docker-make --tag dev moldesign_complete_py2
+```