aradi.notes.txt

* General

√  * Input files contain more than 100 characters per line: shrunk

√  * I would not allow for untagged module imports. We should rather allow for
    module renaming on import or importing specific entries from the module:

    usermodules:
    - skpar.dftbutils.taskdict
    - [skpar.dftbutils.taskdict, td]
    - [skpar.dftbutils.taskdict, [get_evol, set_whatever]]
    
    Implemented

√  * Name skpar.dftbutils.taskdict is way too long! Why not just skpar.dftbutils?
     Implemented


√ * examples/MoS2-repfit/skpar.in.yaml

√   * Without tagimport: how do you know where to get 'get_evol' from?: Gone


* skpar/core/database.py

√  * Do you want to export update()?: not used on own currently

√  * What do we need the Query object for?: used in objectives and .core.plot


* skpar/core/evaluate.py

√  * kwargs is never used in Evaluator(): removed

√  * utopia?: desirable point, which may or may not be within achievable space

√  * Are all class attributes public? Probably don't need to be,
    and will be refactored when style and test-coverage for the module are
    worked on; Currently the code is not very neat, and linter complains about
    too many attributes (Evaluate) and call parameters (__init__).


* skpar/core/test.py

√  * Why is this checked in?: removed

* skpar/core/usertasks.py

√  * ~/.local/share/skpar is a somewhat unfortunate choice. Also users should not
    put their own module into the skpar source folder! They will hopefully
    obtain skpar via pip and should not meddle with its folders themselves.:
    Not used any more

√  * I am not sure, whether trying to look up the file is a good idea at
    all. (What about mypackage.subpackage.module.) Why not just simply try to
    import the module?
    Not done any more

* bin/dftbutils

√ * Calling without argument fails: fixed


* Testing

√  * resulst in 3 failing tests: not if run within the tests directory


----- Additional notes from conversation:

√ * Database -- use or remove: use

* Execute task 
√ - re-think the useage of kwargs[-1]; alternative is explicit: options={}: keep
    as is for the moment
√ - refactor to use popen and ensure continuous output: using subprocess.call()
    is sufficient

* Testing executables -- refactor binaries to only import parsers; then call
  the 'main' with options as though these have been parsed from the command
  line.: work on this on a separate branch (test-coverage)
  
√ remove function definitions inside functions