High level interface and neural network

.gitignore

@@ -3,6 +3,9 @@ __pycache__/
 *.py[cod]
 *$py.class
 
+# pytest
+.pytest_cache/
+
 # C extensions
 *.so
 
@@ -61,6 +64,7 @@ instance/
 .scrapy
 
 # Sphinx documentation
+docs/build/
 docs/_build/
 
 # PyBuilder

.travis.yml

@@ -32,6 +32,9 @@ install:
    if [ ${TRAVIS_PYTHON_VERSION:0:1} = 3 ]; then
        sudo apt-get install python3-numpy
        pip3 install scipy
+       pip3 install ase
+       pip3 install scikit-learn
+       pip3 install tensorflow
        python3 setup.py build
        python3 setup.py install
        pip3 install sphinx
@@ -41,6 +44,9 @@ install:
    elif [ ${TRAVIS_PYTHON_VERSION} = "2.7" ]; then
        sudo apt-get install python-numpy
        pip2 install scipy
+       pip2 install ase
+       pip2 install scikit-learn
+       pip2 install tensorflow
        python2 setup.py build
        python2 setup.py install
        pip2 install sphinx
@@ -53,10 +59,10 @@ install:
 
 
 before_script:
- - cd ${TRAVIS_BUILD_DIR}/tests/
+ - cd ${TRAVIS_BUILD_DIR}/test/
 
 script:
- - nosetests -v
+ - pytest -v
 
 notifications:
   email: false

README.md

@@ -5,18 +5,17 @@ QML is a Python2/3-compatible toolkit for representation learning of properties
 
 #### Current list of contributors:
 * Anders S. Christensen (University of Basel)
+* Lars A. Bratholm (University of Bristol)
+* Silvia Amabilino (University of Bristol)
 * Felix A. Faber (University of Basel)
 * Bing Huang (University of Basel)
-* Lars A. Bratholm (University of Copenhagen)
-* Alexandre Tkatchenko (University of Luxembourg)
-* Klaus-Robert Müller (Technische Universität Berlin/Korea University)
 * O. Anatole von Lilienfeld (University of Basel)
 
 ## 1) Citing QML:
 
 Until the preprint is available from arXiv, please cite this GitHub repository as:
 
-    AS Christensen, LA Bratholm, FA Faber, B Huang, A Tkatchenko, KR Müller, OA von Lilienfeld (2017) "QML: A Python Toolkit for Quantum Machine Learning" https://github.com/qmlcode/qml
+    AS Christensen, LA Bratholm, S Amabilino, FA Faber, B Huang, GR Glowacki, OA von Lilienfeld (2018) "QML: A Python Toolkit for Quantum Machine Learning" https://github.com/qmlcode/qml
 
 
 ## 2) Get help:

README.rst

@@ -10,11 +10,11 @@ Current list of contributors:
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 -  Anders S. Christensen (University of Basel)
+-  Lars A. Bratholm (University of Bristol)
+-  Silvia Amabilino (University of Bristol)
 -  Felix A. Faber (University of Basel)
 -  Bing Huang (University of Basel)
--  Lars A. Bratholm (University of Copenhagen)
--  Alexandre Tkatchenko (University of Luxembourg)
--  Klaus-Robert Müller (Technische Universität Berlin/Korea University)
+-  David R. Glowacki (University of Bristol) 
 -  O. Anatole von Lilienfeld (University of Basel)
 
 1) Citing QML:
@@ -25,7 +25,7 @@ repository as:
 
 ::
 
-    AS Christensen, LA Bratholm, FA Faber, B Huang, A Tkatchenko, KR Müller, OA von Lilienfeld (2017) "QML: A Python Toolkit for Quantum Machine Learning" https://github.com/qmlcode/qml
+    AS Christensen, LA Bratholm, S Amabilino, FA Faber, B Huang, GR Glowacki, OA von Lilienfeld (2018) "QML: A Python Toolkit for Quantum Machine Learning" https://github.com/qmlcode/qml
 
 2) Get help:
 ------------

docs/source/citation.rst

@@ -9,7 +9,7 @@ Until the preprint is available from arXiv, please cite use of QML as:
 
 ::
 
-    AS Christensen, FA Faber, B Huang, LA Bratholm, A Tkatchenko, KR Müller, OA von Lilienfeld (2017) "QML: A Python Toolkit for Quantum Machine Learning" https://github.com/qmlcode/qml
+    AS Christensen, LA Bratholm, S Amabilino, FA Faber, B Huang, DR Glowacki, OA von Lilienfeld (2018) "QML: A Python Toolkit for Quantum Machine Learning" https://github.com/qmlcode/qml
 
 
 

docs/source/conf.py

@@ -16,10 +16,11 @@
 # add these directories to sys.path here. If the directory is relative to the
 # documentation root, use os.path.abspath to make it absolute, like shown here.
 #
+from __future__ import print_function
 import os
 import sys
-print os.getcwd()
-sys.path.insert(0, os.path.abspath('../../build/lib.linux-x86_64-2.7/'))
+print(os.getcwd())
+# sys.path.insert(0, os.path.abspath('../../build/lib.linux-x86_64-2.7/'))
 import sphinx_rtd_theme
 
 # -- General configuration ------------------------------------------------
@@ -47,17 +48,17 @@
 
 # General information about the project.
 project = u'QML'
-copyright = u'2017, Anders S. Christensen'
+copyright = u'2016-2018, Anders S. Christensen'
 author = u'Anders S. Christensen'
 
 # The version info for the project you're documenting, acts as replacement for
 # |version| and |release|, also used in various other places throughout the
 # built documents.
 #
 # The short X.Y version.
-iversion = u'0.4.0'
+iversion = u'0.5.0.0'
 # The full version, including alpha/beta/rc tags.
-release = u'0.4.0'
+release = u'0.5.0.0'
 
 # The language for content autogenerated by Sphinx. Refer to documentation
 # for a list of supported languages.

docs/source/examples.rst

@@ -5,10 +5,12 @@ Generating representations using the ``Compound`` class
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 The following example demonstrates how to generate a representation via
-the ``qml.Compound`` class.
+the ``qml.data.Compound`` class.
 
 .. code:: python
 
+    from qml.data import Compound
+
     # Read in an xyz or cif file.
     water = Compound(xyz="water.xyz")
 
@@ -31,7 +33,7 @@ Generating representations via the ``qml.representations`` module
 .. code:: python
 
     import numpy as np
-    from qml.representations import *
+    from qml.ml.representations import *
 
     # Dummy coordinates for a water molecule
     coordinates = np.array([[1.464, 0.707, 1.056],
@@ -86,7 +88,7 @@ The input for most of the kernels in QML is a numpy array, where the first dimen
 .. code:: python
 
     import numpy as np
-    from qml.kernels import gaussian_kernel
+    from qml.ml.kernels import gaussian_kernel
 
     # Generate a numpy-array of the representation
     X = np.array([c.representation for c in compounds])
@@ -106,7 +108,7 @@ The easiest way to calculate the kernel matrix using an explicit, local represen
 .. code:: python
 
     import numpy as np
-    from qml.kernels import get_local_kernels_gaussian
+    from qml.ml.kernels import get_local_kernels_gaussian
 
     # Assume the QM7 dataset is loaded into a list of Compound()
     for compound in qm7:
@@ -144,7 +146,7 @@ This input (the types of many-body terms) is generate via the ``get_slatm_mbtype
 
 .. code:: python
 
-    from qml.representations import get_slatm_mbtypes
+    from qml.ml.representations import get_slatm_mbtypes
 
     # Assume 'qm7' is a list of Compound() objects.
     mbtypes = get_slatm_mbtypes([mol.nuclear_charges for compound in qm7])
@@ -155,11 +157,11 @@ This input (the types of many-body terms) is generate via the ``get_slatm_mbtype
         # Generate the desired representation for each compound
         compound.generate_slatm(mbtypes, local=True)
 
-The ``local`` keyword in this example specifies that a local representation is produced. Alternatively the SLATM representation can be generate via the ``qml.representations`` module:
+The ``local`` keyword in this example specifies that a local representation is produced. Alternatively the SLATM representation can be generate via the ``qml.ml.representations`` module:
 
 .. code:: python
 
-    from qml.representations import generate_slatm
+    from qml.ml.representations import generate_slatm
 
     # Dummy coordinates
     coordinates = ... 
@@ -178,7 +180,7 @@ Here ``coordinates`` is an Nx3 numpy array, and ``nuclear_charges`` is simply a
 Generating the FCHL representation
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 The FCHL representation does not have an explicit representation in the form of a vector, and the kernel elements must be calculated analytically in a separate kernel function.
-The syntax is analogous to the explicit representations (e.g. Coulomb matrix, BoB, SLATM, etc), but is handled by kernels from the separate ``qml.fchl`` module.
+The syntax is analogous to the explicit representations (e.g. Coulomb matrix, BoB, SLATM, etc), but is handled by kernels from the separate ``qml.ml.representations.fchl`` module.
 
 The code below show three ways to create the input representations for the FHCL kernel functions.
 
@@ -199,12 +201,12 @@ First using the ``Compound`` class:
 The dimensions of the array should be ``(number_molecules, size, 5, size)``, where ``size`` is the
 size keyword used when generating the representations. 
 
-In addition to using the ``Compound`` class to generate the representations, FCHL representations can also be generated via the ``qml.fchl.generate_fchl_representation()`` function, using similar notation to the functions in the ``qml.representations.*`` functions.
+In addition to using the ``Compound`` class to generate the representations, FCHL representations can also be generated via the ``qml.ml.representations.fchl.generate_fchl_representation()`` function, using similar notation to the functions in the ``qml.ml.representations.*`` functions.
 
 
 .. code:: python
 
-    from qml.fchl import generate_representation 
+    from qml.ml.representations.fchl import generate_representation 
 
     # Dummy coordinates for a water molecule
     coordinates = np.array([[1.464, 0.707, 1.056],
@@ -221,7 +223,7 @@ To create the representation for a crystal, the notation is as follows:
 
 .. code:: python
 
-    from qml.fchl import generate_representation 
+    from qml.ml.representations.fchl import generate_representation 
 
     # Dummy fractional coordinates
     fractional_coordinates = np.array(
@@ -266,7 +268,7 @@ The following example demonstrates how to calculate the local FCHL kernel elemen
 
 .. code:: python
 
-    from qml.fchl import get_local_kernels
+    from qml.ml.representations.fchl import get_local_kernels
 
     # You can get kernels for multiple kernel-widths
     sigmas = [2.5, 5.0, 10.0]
@@ -288,7 +290,7 @@ In case ``X1`` and ``X2`` are identical, K will be symmetrical. This is handled
 
 .. code:: python
 
-    from qml.fchl import get_local_symmetric_kernels
+    from qml.ml.representations.fchl import get_local_symmetric_kernels
 
     # You can get kernels for multiple kernel-widths
     sigmas = [2.5, 5.0, 10.0]
@@ -307,8 +309,8 @@ In addition to the local kernel, the FCHL module also provides kernels for atomi
 
 .. code:: python
 
-    from qml.fchl import get_atomic_kernels
-    from qml.fchl import get_atomic_symmetric_kernels
+    from qml.ml.representations.fchl import get_atomic_kernels
+    from qml.ml.representations.fchl import get_atomic_symmetric_kernels
 
 The only difference between the local and atomic kernels is the shape of the input.
 Since the atomic kernel outputs kernels with atomic resolution, the atomic input has the shape ``(number_atoms, 5, size)``.

docs/source/index.rst

@@ -9,22 +9,22 @@ QML: A Python Toolkit for Quantum Machine Learning
 ==================================================
 
 QML is a Python2/3-compatible toolkit for representation learning of
-properties of molecules and solids. QML is not a high-level framework
-where you can do ``model.train()``, but supplies the building blocks to
+properties of molecules and solids. QML supplies the the building blocks to
 carry out efficient and accurate machine learning on chemical compounds.
-As such, the goal is to provide usable and efficient implementations of
-concepts such as representations and kernels. 
+The goal is to provide usable and efficient implementations of
+concepts such as representations and kernels as well as a high level interface
+to make it easy for computational chemists to use these for machine-learning tasks.
 
 
 Current list of contributors:
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 -  Anders S. Christensen (University of Basel)
+-  Lars A. Bratholm (University of Bristol)
+-  Silvia Amabilino (University of Bristol)
 -  Felix A. Faber (University of Basel)
 -  Bing Huang (University of Basel)
--  Lars A. Bratholm (University of Copenhagen)
--  Alexandre Tkatchenko (University of Luxembourg)
--  Klaus-Robert Müller (Technische Universität Berlin/Korea University)
+-  David R. Glowacki (University of Bristol)
 -  O. Anatole von Lilienfeld (University of Basel)
 
 Code development
@@ -48,7 +48,7 @@ Until the preprint is available from arXiv, please cite use of QML as:
 
 ::
 
-    AS Christensen, FA Faber, B Huang, LA Bratholm, A Tkatchenko, KR Müller, OA von Lilienfeld (2017) "QML: A Python Toolkit for Quantum Machine Learning" https://github.com/qmlcode/qml
+    AS Christensen, LA Bratholm, S Amabilino, FA Faber, B Huang, DR Glowacki, OA von Lilienfeld (2018) "QML: A Python Toolkit for Quantum Machine Learning" https://github.com/qmlcode/qml
 
 
 For citation of the individual procedures of QML, please see the "Citing use of QML" section.

docs/source/installation.rst

@@ -5,7 +5,7 @@ Installing prerequisite modules (for most Linux systems):
 
 .. code:: bash
 
-    sudo apt-get install python-pip gfortran libblas-dev liblapack-dev git
+    sudo apt-get install python-pip gfortran libblas-dev liblapack-dev
 
 These should already be installed on most systems. The Intel compilers
 and MKL math-libraries are supported as well (see section 1.3).

docs/source/qml.rst

@@ -8,39 +8,39 @@ Python API documentation
 qml\.representations module
 ---------------------------
 
-.. automodule:: qml.representations
+.. automodule:: qml.ml.representations.representations
     :members:
     :undoc-members:
     :show-inheritance:
 
 qml\.kernels module
 -------------------
 
-.. automodule:: qml.kernels
+.. automodule:: qml.ml.kernels
     :members:
     :undoc-members:
     :show-inheritance:
 
 qml\.distance module
 --------------------
 
-.. automodule:: qml.distance
+.. automodule:: qml.ml.kernels.distance
     :members:
     :undoc-members:
     :show-inheritance:
 
 qml\.math module
 ----------------
 
-.. automodule:: qml.math
+.. automodule:: qml.ml.math
     :members:
     :undoc-members:
     :show-inheritance:
 
 qml\.Compound class
 --------------------
 
-.. autoclass:: qml.Compound
+.. autoclass:: qml.data.compound.Compound
     :members:
     :undoc-members:
     :show-inheritance:
@@ -55,15 +55,15 @@ qml\.Compound class
 qml\.fchl module
 ----------------
 
-.. automodule:: qml.fchl
+.. automodule:: qml.ml.representations.fchl
     :members:
     :show-inheritance:
 
 
 qml\.wrappers module
 --------------------
 
-.. automodule:: qml.wrappers
+.. automodule:: qml.ml.kernels.wrappers
     :members:
     :undoc-members:
     :show-inheritance:
@@ -85,3 +85,11 @@ qml\.wrappers module
 ..    :undoc-members:
 ..    :show-inheritance:
 
+
+qml\.aglaia module
+------------------
+
+.. automodule:: qml.aglaia.aglaia
+   :inherited-members:
+
+