Fix broken argument parser test #56
Conversation
| @@ -11,7 +11,7 @@ | |||
|
|
|||
| def test_make_inference_arg_parser(): | |||
| parser = make_inference_arg_parser() | |||
| args = parser.parse_args() | |||
| args = parser.parse_known_args()[0] | |||
There was a problem hiding this comment.
I think you could test this by passing in the command line arguments to parse explicitly. This way, you could test different inputs without needing to run from the command line (and ignore any other conflicting arguments from the pytest invocation). Something like parser.parse_args([]).
Then if you wanted to test specific inputs you'd do something like parser.parse_args(["--flag", "test_value"])
There was a problem hiding this comment.
Changed to match your suggestion.
Codecov ReportAttention:
Additional details and impacted files@@ Coverage Diff @@
## main #56 +/- ##
========================================
+ Coverage 1.23% 4.97% +3.73%
========================================
Files 19 19
Lines 2908 3056 +148
========================================
+ Hits 36 152 +116
- Misses 2872 2904 +32
☔ View full report in Codecov by Sentry. |
Use
parse_known_argsinstead ofparse_argsso the coverage tests do not fail.parse_known_argsreturns a tuple of (known, unknown) args, so we take the first element of that tuple.