Add test save #686
Add test save #686
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #686 +/- ##
==========================================
+ Coverage 70.64% 71.00% +0.35%
==========================================
Files 16 16
Lines 3369 3369
==========================================
+ Hits 2380 2392 +12
+ Misses 989 977 -12 ☔ View full report in Codecov by Sentry. |
| from pathlib import Path | ||
|
|
||
| # Save ontology in a different location | ||
| testonto.save(tmpdir / "testonto_saved.ttl") |
There was a problem hiding this comment.
I stedet for å skrive til en temporær mappe som automatisk slettes når testen har kjørt, er det enklere å debugge hvis man skriver til en output/ undermappe under tests/. Hvis du legger inn en .gitignore fil i den mappen som inneholder en linje med "*", vill den mappen bli en del av git-repoet men alle filene som blir generert i den vill bli ignorert av git. Men de vill fortsatt være enkelt tilgjengelige for debugging.
Eneste ulempen med en slik løsning er at output/ ikke er garantert å være tom ved start av testen. Dvs. man må sette overwrite=True. At overwrite fungerer må selvfølgelig testes eksplisitt som du gjør under.
There was a problem hiding this comment.
I prefer to make use of temporary directories that are cleaned up after use and that tests are run in a clean directory every time if that is the intended purpose. Overwrite=False is the default, I don't want to have overwrite=True unless I want to specifically test that. I have added a suggestion for writing to a local directory in debug mode (setting debug=True). Then we meet half way :)
There was a problem hiding this comment.
As you wish... but I really don't see the point of such complexity. Why introducing different test modes? The tests are not part of the distributed production code, but ment to be simple and helpful for ensuring code quality. Easy debugging is part of that.
All extra complexity added to the tests just make the code harder to maintain - and will sooner or later require that we have to add tests for the tests...
| testonto.save(format="rdfxml", dir=tmpdir) | ||
| # 3. | ||
| with open(tmpdir / "testonto_saved.owl") as f: | ||
| owlfile = f.read() | ||
| with open(tmpdir / "testonto.rdfxml") as f: | ||
| owlfile2 = f.read() | ||
| # assert owlfile != owlfile2 # to be uncommented when issue #685 is fixed |
There was a problem hiding this comment.
Hvorfor skal owlfile og owlfile2 være ulike. Har du endret på testonto?
There was a problem hiding this comment.
Because I have saved testonto.rdfxml again. Since overwrite=False is the default it should append to it, thus make the two files different. Stated it even more explicitly in the comment aboe (step2)
There was a problem hiding this comment.
The only thing overwrite=True does is to remove the destination before writing to it. Without overwrite, the behaviour is defined by owlready2 for rdfxml and rdflib for turtle. I haven't checked their documentation, but I don't expect that any of them will duplicate a triple when writing twice to the file.
Removed unecessary print lines
Description
Type of change
Checklist
This checklist can be used as a help for the reviewer.
Comments