Define data paths from a standard TiMBA data directory defined in TIMBA_DATA_DIR

[paulrougieux]

The following proposal addresses the following issues:

• Copying the model's output files shouldn't be necessary for a dedicated visualisation app.
• For development purposes it's alright to store data inside the package's structure. But what happens for persons who install from the python package index? The data should be stored in a dedicated data directory, outside the directory structure of the python package.

Current behaviour

Copy the TiMBA output pickle file from TiMBA to TiMBA_Charts:

- from /home/paul/rp/zz_forks/TiMBA/TiMBA/data/output
- to /home/paul/rp/zz_forks/TiMBA_Charts/Toolbox/Input/Scenario_Files

Expressed as a shell command the copy instruction would be like this:

cp /home/paul/rp/zz_forks/TiMBA/TiMBA/data/output /home/paul/rp/zz_forks/TiMBA_Charts/Toolbox/Input/Scenario_Files

Then to import the data into python. The TiMBA_Charts python package is
strangely called "Toolbox". Current behaviour:

  from Toolbox.classes.import_data import import_pkl_data
  print(import_pkl.SCENARIOPATH)

The paths are defined in Toolbox/parameters/paths.py

PACKAGEDIR = Path(__file__).resolve().parent.parent
SCINPUTPATH = PACKAGEDIR / Path("Input") / Path("Scenario_Files")
AIINPUTPATH = PACKAGEDIR / Path("Input") / Path("Additional_Information")

Proposal

Make SCINPUTPATH and AIINPUTPATH sub directories of TIMBA_DATA_DIR and point them to the model's output directory:

SCINPUTPATH = TIMBA_DATA_DIR / Path("output") / Path("Scenario_Files")
AIINPUTPATH = TIMBA_DATA_DIR / Path("output") / Path("Additional_Information")

You can pull my commit from the fork branch timba_data_dir to get the example
paulrougieux@6e4dc4b

Related issues:

• The proposal above depends on adding TiMBA as a requirement of TiMBA_Charts.

• TiMBA_charts could add TiMBA as a dependency to . See issue: