CLN: Refactor Generic Simulation classes to create public API

[flongford]

This PR approaches #26

It introduces new classes that lay the foundation for how chemical and simulation concepts are handled within the BDSS.

Includes contributions from #14 , #28

Change log:

Modules

New:

• force_gromacs.chemicals: contains data frame objects describing chemical entities such as molecular fragments and molecules
• force_gromacs.simulation_builders: contains classes that are used to construct BasePipeline objects to run a simulation

Updated:

• force_gromacs.pipeline --> force_gromacs.pipelines: contains BasePipeline objects that stack a sequence of command line programs
• force_gromacs.core: now contains fundamental classes required for calling command line programs using subprocess library

Classes

New:

• BasePipeline
• IFragment
• FileRegistry: generic class to hold formatting instructions for simulation input / output files
• GromacsFileRegistry: subclass of FileRegistry that also holds example formatted file names for a given prefix
• GromacsTopologyData: container class for all topology related files and objects in a simulation (defining type and numerical of each molecular fragment in a simulation and their forcefields)

Renamed:

• Fragment --> GromacsFragment
• BaseGromacsProcess --> BaseProcess
• IBaseProcess --> IProcess
• GromacsFileTreeBuilder --> FileTreeBuilder

[codecov-io]

Codecov Report

:exclamation: No coverage uploaded for pull request base (master@e1fccf1). Click here to learn what that means. The diff coverage is 98.22%.

@@           Coverage Diff            @@
##             master     #27   +/-   ##
========================================
  Coverage          ?   98.7%           
========================================
  Files             ?      65           
  Lines             ?    2170           
  Branches          ?     138           
========================================
  Hits              ?    2142           
  Misses            ?      16           
  Partials          ?      12

Impacted Files	Coverage Δ
...tion_listeners/hpc_writer/tests/test_hpc_writer.py	100% <ø> (ø)
force_gromacs/chemicals/i_fragment.py	100% <ø> (ø)
force_gromacs/core/i_process.py	100% <100%> (ø)
...romacs/commands/tests/test_base_gromacs_command.py	100% <100%> (ø)
...ders/tests/test_base_gromacs_simulation_builder.py	100% <100%> (ø)
...macs/data_sources/fragment/fragment_data_source.py	100% <100%> (ø)
force_gromacs/commands/gromacs_commands.py	100% <100%> (ø)
...es/simulation/tests/test_simulation_data_source.py	100% <100%> (ø)
...omacs/simulation_builders/gromacs_topology_data.py	100% <100%> (ø)
force_gromacs/io/tests/test_file_tree_builder.py	100% <100%> (ø)
... and 31 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update e1fccf1...357c3e7. Read the comment docs.

[flongford]

Out of curiosity, what is the difference between a GromacsPipeline and a BDSS workflow in evaluation mode?

They are similar, in that they chain together command line calls using subprocess. There's a lot of overlap between the BaseGromacsCommand and SubprocessWorkflow classes actually - we could probably end up refactoring the BaseProcess class if it is useful elsewhere.

The GromacsPipeline just has some extra functionalities ability to stack together multiple command line calls, and also output a bash script capable of performing the same operations.