Research Software Engineering

Criteria and taxonomy for research software engineering (rseng).

Overview

This repository serves a taxonomy and criteria for research software, intended to be used with the research software encyclopedia. The two are maintained separately for development, and because it might be the case that the criteria and taxonomy would want to be used separately from the encyclopedia.

Usage

Usage of the library includes programmatic (within Python or command line) interaction with criteria or taxonomy, and generation of output files.

• Criteria
• Taxonomy
• Generate

Criteria

For usage within Python, you will first want to instantiate a CriteriaSet. If you don't provide a default file, the library default will be used.

from rseng.main.criteria import CriteriaSet
cset = CriteriaSet()
# [CriteriaSet:6]

You can then see questions loaded. Each has a unique id that gives a sense of what is being asked:

cset.criteria                                                                       
{'RSE-research-intention': <rseng.main.criteria.base.Criteria at 0x7f3d2e85d410>,
 'RSE-domain-intention': <rseng.main.criteria.base.Criteria at 0x7f3d2dab8490>,
 'RSE-question-intention': <rseng.main.criteria.base.Criteria at 0x7f3d2dab8910>,
 'RSE-citation': <rseng.main.criteria.base.Criteria at 0x7f3d2db34810>,
 'RSE-usage': <rseng.main.criteria.base.Criteria at 0x7f3d2db340d0>,
 'RSE-absence': <rseng.main.criteria.base.Criteria at 0x7f3d2db34850>}

You can inspect any particular criteria:

cset.criteria['RSE-usage']
<rseng.main.criteria.base.Criteria at 0x7f3d2db340d0>

cset.criteria['RSE-usage'].uid
# 'RSE-usage'

cset.criteria['RSE-usage'].question
# 'Has the software been used by researchers?'

cset.criteria['RSE-usage'].options
# ['yes', 'no']

And further interact with the CriteriaSet, for example export to a tabular file:

print(cset.export()) # You can also define a "filename" and/or "sep" here.
RSE-research-intention	Is the software intended for research?	yes,no
RSE-domain-intention	Is the software intended for a particular domain?	yes,no
RSE-question-intention	Was the software created with intention to solve a research question?	yes,no
RSE-citation	Has the software been cited?	yes,no
RSE-usage	Has the software been used by researchers?	yes,no
RSE-absence	Would taking away the software be a detriment to research?	yes,no

or iterate through the criteria, or get a list of all of them.

> list(cset)
[[Criteria:RSE-research-intention,Is the software intended for research?],
 [Criteria:RSE-domain-intention,Is the software intended for a particular domain?],
 [Criteria:RSE-question-intention,Was the software created with intention to solve a research question?],
 [Criteria:RSE-citation,Has the software been cited?],
 [Criteria:RSE-usage,Has the software been used by researchers?],
 [Criteria:RSE-absence,Would taking away the software be a detriment to research?]]

for criteria in cset:
    print(criteria)

[Criteria:RSE-research-intention,Is the software intended for research?]
[Criteria:RSE-domain-intention,Is the software intended for a particular domain?]
[Criteria:RSE-question-intention,Was the software created with intention to solve a research question?]
[Criteria:RSE-citation,Has the software been cited?]
[Criteria:RSE-usage,Has the software been used by researchers?]
[Criteria:RSE-absence,Would taking away the software be a detriment to research?]

Taxonomy

The taxonomy is interacted with in a similar fashion.

from rseng.main.taxonomy import Taxonomy
tax = Taxonomy()

It will show you the total number of nodes (nested too):

from rseng.main.taxonomy import Taxonomy
tax = Taxonomy()
#  [Taxonomy:24]

Validation happens as the default file is loaded. Akin to criteria, the files are located in rseng/main/taxonomy in yaml format, and are dated. You can quickly print an easily viewable, human understandable version of the tree:

for name in tax.flatten(): 
   ...:     print(name) 
   ...:                                                                                                                                                                                                                      
Software to directly conduct research >> Domain-specific software >> Domain-specific hardware
Software to directly conduct research >> Domain-specific software >> Domain-specific optimized software
Software to directly conduct research >> Domain-specific software >> Domain-specific analysis software
Software to directly conduct research >> General software >> Numerical libraries
Software to directly conduct research >> General software >> Data collection
Software to directly conduct research >> General software >> Visualization
Software to support research >> Explicitly for research >> Workflow managers
Software to support research >> Explicitly for research >> Interactive development environments for research
Software to support research >> Explicitly for research >> Provenance and metadata collection tools
Software to support research >> Used for research but not explicitly for it >> Databases
Software to support research >> Used for research but not explicitly for it >> Application Programming Interfaces
Software to support research >> Used for research but not explicitly for it >> Frameworks
Software to support research >> Incidentally used for research >> Operating systems
Software to support research >> Incidentally used for research >> Personal scheduling and task management
Software to support research >> Incidentally used for research >> Version control
Software to support research >> Incidentally used for research >> Text editors and integrated development environments
Software to support research >> Incidentally used for research >> Communication tools or platforms

Generate

After you install rseng, the rseng executable should be in your path. You can generate output files for the taxonomy or critiera to a folder oath that doesn't exist yet. For example, to generate the markdown files for the static documentation for each of the taxonomy and criteria we do:

# rseng generate <type>   <path>          <version>
$ rseng generate taxonomy docs/_taxonomy
docs/_taxonomy/RSE-taxonomy-domain-hardware.md
docs/_taxonomy/RSE-taxonomy-optimized.md
docs/_taxonomy/RSE-taxonomy-analysis.md
docs/_taxonomy/RSE-taxonomy-numerical libraries.md
docs/_taxonomy/RSE-taxonomy-data-collection.md
docs/_taxonomy/RSE-taxonomy-visualization.md
docs/_taxonomy/RSE-taxonomy-workflow-managers.md
docs/_taxonomy/RSE-taxonomy-ide-research.md
docs/_taxonomy/RSE-taxonomy-provenance-metadata-tools.md
docs/_taxonomy/RSE-taxonomy-databases.md
docs/_taxonomy/RSE-taxonomy-application-programming-interfaces.md
docs/_taxonomy/RSE-taxonomy-frameworks.md
docs/_taxonomy/RSE-taxonomy-operating-systems.md
docs/_taxonomy/RSE-taxonomy-personal-scheduling-task-management.md
docs/_taxonomy/RSE-taxonomy-version-control.md
docs/_taxonomy/RSE-taxonomy-text-editors-ides.md
docs/_taxonomy/RSE-taxonomy-communication-tools.md

The default version generated for each is "latest" but you can add another version as the last argument to change that. Here is generation of the criteria, showing using latest:

# rseng generate <type>   <path>          <version>
$ rseng generate criteria docs/_criteria
docs/_criteria/RSE-research-intention.md
docs/_criteria/RSE-domain-intention.md
docs/_criteria/RSE-question-intention.md
docs/_criteria/RSE-citation.md
docs/_criteria/RSE-usage.md
docs/_criteria/RSE-absence.md

You can also generate a (non flat) version of the taxonomy, specifically a json file that plugs easily into the d3 hierarchy plots.

# rseng generate taxonomy-json <filename>
$ rseng generate taxonomy-json taxonomy.json

License

• Free software: MPL 2.0 License