pyfreedts 2.1.1

Python wrapper for the FreeDTS C++ project

pyFreeDTS

pyFreeDTS is wrapper for the FreeDTS C++ project and aims to simplify installing and interfacing with FreeDTS.

Installation

After ensuring your system satisfies the following requirements

• Python 3.8+
• C++ compiler with C++11 support
• OpenMP (optional)

you can install pyFreeDTS using pip

pip install pyfreedts

Usage

pyFreeDTS mirrors the original FreeDTS binaries, but changed the spelling to lower-ase to avoid collision

dts arg1 arg2 ...
cnv arg1 arg2 ...
gen arg1 arg2 ...

Parameter Screening

From version 2.2 onwards, pyFreeDTS simplifies running parameter screens by introducing templated dts files

Kappa = {{kappa:25.0:35.0:5.0}} 0 0        # Range: 25.0 to 35.0, step 5.0
Temperature = {{temp:1.0,1.5,2.0}} 0       # List: 1.0, 1.5, 2.0
Method = {{method:MC,MD}} 0                # String list
Set_Steps = 1 {{steps:1000,5000}}          # Integer list

A DTS screen based on such a template file can be performed using

dts_screen \
    -in input_template.dts \
    --output-dir screen_result \
    --dts-args "-top /absolute/path/to/topol.top -seed 12345 -e 100000" \
    --backend-args "-j 4 --keep-going" \
    --dry-run

where --dts-args are parameters passed to the dts executable, and --backend-args are parameters passed to the execution backend specified via --backend. The default is snakemake and can be installed using pip

pip install snakemake

Through snakemake, DTS screens can be run locally and on all commonly available HPC platforms.

[!IMPORTANT] The topologies listed in topol.top need to be specified using their absolute path.

Running the pipeline will generate numbered directories in --output-dir, each of which contains the result of a DTS run with corresponding params.json. A summary of all parameters and corresponding directories is given by screen_summary.json in the specified output directory.

screen_result/
├── Snakefile
├── sweep_summary.json
├── run_0001/
│   ├── input.dts
│   ├── params.json
│   ├── dts_*
│   └── completed.flag
└── run_NNNN/
    ├── input.dts
    ├── params.json
    ├── dts_*
    └── completed.flag

[!IMPORTANT] Every screen needs to run for at least 1000 steps (-dts-args '-e 1000'), due to an implementation detail in FreeDTS.

For Developers

Clone and install with Poetry:

git clone https://github.com/maurerv/pyfreedts.git
cd pyfreedts
poetry install

Run tests:

poetry run pytest

License

Like FreeDTS, pyFreeDTS is available under a CC-BY-NC-ND 4.0 International license.