8. Using MDANSE Command Line Interface

The MDANSE analysis tasks can exist and operate independent of a graphical interface. For this reason, the MDANSE command line interface is a valid alternative to the MDANSE GUI.

8.1. Managing jobs using the MDANSE CLI

While the full list of the CLI options is listed below, here are some useful examples that illustrate how the CLI should be used.

mdanse -r job

will display a list of different tasks that can be run in MDANSE.

mdanse --js dacf

will create a template script for the Dipole AutoCorrelation Function. This script can then be edited to change the job parameters.

mdanse --jr my_modified_script.py

will run the job defined in the file ‘my_modified_script.py’.

mdanse --jl

will display the list of currently existing MDANSE jobs.

Congratulations! Using these four lines (and a text editor) you just performed an analysis task on a trajectory, and confirmed that your task has executed correctly.

8.2. Python interpreter information

Windows, has a dedicated command line which can be run using MDANSE_command_shell file, which sets some environment variables so that, in it, MDANSE python is the default python. On other platforms, you have to use a normal terminal and use MDANSE python by calling its full path, which should be (if MDANSE is in default installation location) /Applications/MDANSE.app/Contents/MacOS/python2 on MacOS, and /usr/local/bin/python on Linux systems. Please only use the python2 file on MacOS, the other python file does not have some environmental variables set up.

These pythons (as discussed in previous paragraph) can then be used to run MDANSE python scripts like any other python script:

python script.py – Windows

path/python2 script.py – MacOS

path/python script.py – Linux

The python can also be used to install other packages, run short code (using the -c option), or to activate python REPL.

8.3. Custom scripts

It is possible to edit MDANSE scripts or even write new ones from scratch. To run an analysis using the MDANSE python library, two steps are necessary; first, set up the parameters that the analysis requires (equivalent to filling in the fields in the GUI), then running the analysis (equivalent to clicking the Run button). For both of these, it is necessary to understand how the analysis’ class works. This can be done by reading MDANSE documentation, either by clicking the analysis’ Help button or by clicking the Open MDANSE API button on the toolbar. An example script is below.

################################################################

# Job parameters #

################################################################


parameters = {}
parameters['atom_charges'] = ''
parameters['atom_selection'] = None
parameters['frames'] = (0, 2258, 1)
parameters['output_files'] =
(u'C:\\\\Users\\\\TACHYON\\\\Downloads\\\\output_NaF', (u'netcdf',))
parameters['running_mode'] = ('monoprocessor',)
parameters['trajectory'] =
u'C:\\\\Users\\\\TACHYON\\\\Downloads\\\\NaF.nc'

################################################################

# Setup and run the analysis #

################################################################

# Create an instance of the class

dacf = REGISTRY['job']['dacf']()

# Run the analysis

dacf.run(parameters,status=True)

8.4. MDANSE Scripts

When MDANSE is installed, multiple scripts come with it, installed into the Scripts\ directory on Windows and bin/ directory on Unix systems. They can be run with python, for example

python mdanse_gui

or by themselves (this mode might be unavailable on Windows):

mdanse_gui

The following scripts come with MDANSE, each of which is described in the following subsections, though more information about a script can also be gained by calling the script with -h, e.g.,

python mdanse_gui -h

  • mdanse

  • mdanse_elements_database

  • mdanse_gui

  • mdanse_job

  • mdanse_periodic_table

  • mdanse_plotter

  • mdanse_ud_editor

  • mdanse_units_editor

8.4.1. mdanse

This script is used to interface with the current installation of MDANSE without running the GUI. It has the following options, where the expected arguments should be inputted after the option as space-separated values:

--version

  • Description: displays the version of the installed MDANSE

  • Expected arguments: None

--add-mmtk-def

  • Description: adds the provided definition to the MMTK database

  • Expected arguments: code, typ, filename

  • code: the MMTK code for the molecule to register (i.e., HOH for water)

  • typ: the molecular type; one of amino_acid, molecule, nucleic_acid

  • filename: the path to the file that stores the MMTK definition of the molecule being added

--database or -d

  • Description: displays chemical information about the provided element

  • Expected arguments: ename

  • ename: the name of a registered element

--registry or -r

  • Description: displays the contents of MDANSE classes registry

  • Expected arguments: None or interface

    • None -> information on all classes is displayed

    • interface: the name of a class -> information on only the subclasses of the provided class is displayed

--traj or -t

  • Description: displays the chemical contents of a trajectory

  • Expected arguments: trajName

  • trajName: the name of a trajectory that has been loaded into MDANSE

--jc

  • Description: shows the status of the provided job

  • Expected arguments: filename

    • filename: the name (not path!) of a file representing an MDANSE job

--jl

  • Description: displays the job list

  • Expected arguments: None

--jr

  • Description: runs the provided MDANSE job(s)

  • Expected arguments: filename

    • filename: the path to an MDANSE python script

--js

  • Description: saves a job script for the provided job with default parameters

  • Expected arguments: name

    • name: the name of a job (e.g., ccf for Current Correlation Function)

--jt

  • Description: saves a new job template

  • Expected arguments: classname, shortname

    • classname: a full name for the new job (e.g., TXTConverter)

    • shortname: a short name for the new job (e.g., txtc)