nepomnyi/stin

"stin" stands for "steady influx". This package is based on drift-flux model (DFM) and devoted for steady inlux simulation.

stin v0.2.1

About

   Steady influx modeling package (stin) allows for simulation of steady influx of gas into a vertical steady flow of liquid for a limited range of flow parameters. It is built on drift-flux model.
   This is a python package, so one should have python (with the standard library) installed on the computer. Required python version is 3.7 or newer.    The package has a documentation website. It is worth mentioning, that the documentation website is still under construction and currently has bad readability. Therefore, it is recommended to use this documentation file along with the report for the purpose of the package exploration. As soon as the documentation website is updated, this paragraph will notify users about that. To date, one can find mathematical description of the package along with the software description in the report.

Installation

Installation from source

• Download the package from GitHub. Then there are 2 options.
• 1st option:
  • On your computer, from the command line of your choice, go to the package root repository.
  • In the command line, type in:
    py setup.py install
    (this command will install the package with all necessary dependencies).
  • In this case, the package can only be uninstalled manually. For Windows go to *C:\Users\your_user_name\AppData\Local\Continuum\anaconda3\Lib\site-packages*
    There you may find a folder called stin-0.2.0-py3.7.egg (the folder name may vary dependently on the package version and version of python you use). Delete this folder and the package will be deleted.
    Note that AppData is a hidden folder.
    For OSX and Linux the process of uninstallation was not tested.
• 2nd option:
  • On your computer, from the command line of your choice, being in any directory (it doesn't matter), type in:
    pip install <path to the package's root directory>
    E.g. for windows:
    pip install C:\Users\your_user_name\Documents\stin
    In this example, stin is the root directory of the package: the one which contains setup.py file.
  • In order to uninstall the package, from command line of your choice, being in any directory (it doesn't matter), type in:
    pip uninstall stin

Installation from TestPyPI

• In the command line of your choice (in any directory - it doesn't matter), type in:
  pip install -i https://test.pypi.org/simple/ stin
• To uninstall the package use:
  pip uninstall stin

Getting started

Let's consider usage of the package on the example.

• In command line (being in any directory - it doesn't matter) type in: py -m stin -alpha 0.01:
  • alternatively, it can be run as: py -m stin --initial_gas_fraction 0.01,
  • -m runs the package as a module,
  • -alpha (or --initial_gas_fraction) is a variable that the package takes to run the code - it is influx's initial volume fraction,
  • 0.01 is the value of the variable: can be anything from 0 to 1 not including the extremes:
    • for the zeroth version of the package it is not recommended to specify initial gas fraction values more than 0.4.
• The output is a set of graphs for each flow variable with relation to the length of the vertical pipe,
  • the flow variables are velocities of gas and liquid phases, volume fractions of gas and liquid phases, gas phase density, flow pressure - their distributions along the flow is the output of the software,
  • one may find plots for the current example with explanation in the section 4 of the report.

Testing

• Using installation from source:
  • In command line, being in the root directory of the package, type in: pytest or pytest -vv --cov=./
  • In both cases the result is the same: a user will see two tables
    • the first one will contain names of the unit and integration tests and passing marks to the right of them,
    • the second one will contain test coverage,
    • test coverage for the package is 82%:
      • test coverage doesn't work in TravisCI, that is why continuous integration status badge at the top of this file shows failing - fixing this issue entails rewriting a chunk of code and is considered to be a part of future work,
      • integration tests are not written for the plotting function, that is why test coverage is 82% only; writing tests for the plotting function is above the author's knowledge, hence, is considered to be time consuming and left for the future work.
• Using installation from TestPyPI:
  • the author is not aware of any way to apply either pytest or pytest-cov on an installed package,
  • testing can be done only from within package's root directory - for that purpose the package should be downloaded from GitHub and the above described procedure can be implemented.

Acknowledgment

FRIDGe, that helped me to set TravisCI up.