Average grades in export #2366
Average grades in export #2366
Conversation
fidoriel
force-pushed
the
average-grades-in-export
branch
3 times, most recently
from
e5117a8 to
d51385d
Compare
fidoriel
force-pushed
the
average-grades-in-export
branch
from
d51385d to
af9697b
Compare
fidoriel
force-pushed
the
average-grades-in-export
branch
from
af9697b to
382680b
Compare
| if semesters: | ||
| evaluations_filter &= Q(course__semester__in=semesters) | ||
| if evaluation_states: | ||
| evaluations_filter &= Q(state__in=evaluation_states) | ||
| if program_ids: | ||
| evaluations_filter &= Q(course__programs__in=program_ids) | ||
| if course_type_ids: | ||
| evaluations_filter &= Q(course__type__in=course_type_ids) |
There was a problem hiding this comment.
The ifs should all use is not None checks to distuingish between empty iterable and None.
There was a problem hiding this comment.
But then empty iterables are not covered, and the code breaks, we want to skip both.
evap/results/exporters.py
Outdated
| @@ -198,6 +203,8 @@ def write_headings_and_evaluation_info( | |||
| else: | |||
| self.write_cell(export_name, "headline") | |||
|
|
|||
| self.write_cell("Average for this Question", "evaluation") | |||
There was a problem hiding this comment.
I think "Average for this Question" would need to be translated here.
The styles used here and in the lines changed below confuse me. Why is this using "evaluation"? Why is the empty cell in the course type column using the "program" style below?
Please double-check that these all make sense, and if we really need some unintuitive value, add some explanation as to why.
evap/results/exporters.py
Outdated
| # One more cell is needed for the question column | ||
| self.write_empty_row_with_styles(["default"] + ["border_left_right"] * len(evaluations_with_results)) | ||
| self.write_empty_row_with_styles(["default"] + ["border_left_right"] * (len(evaluations_with_results) + 1)) |
There was a problem hiding this comment.
The comment does not adequately explain the code anymore. Suggestion:
One column for the question, one column for the average, n columns for the evaluations
evap/results/exporters.py
Outdated
| # Borders only if there is a course grade below. Offset by one column | ||
| self.write_empty_row_with_styles( | ||
| ["default"] + ["border_left_right" if gt1 else "default" for gt1 in count_gt_1] | ||
| ["default", "default"] + ["border_left_right" if gt1 else "default" for gt1 in count_gt_1] |
There was a problem hiding this comment.
Comment doesn't match code anymore
| ) | ||
|
|
||
| self.write_cell(_("Evaluation weight"), "bold") | ||
| weight_percentages = ( | ||
| self.write_cell("") | ||
| weight_percentages = tuple( |
There was a problem hiding this comment.
We need to write an empty cell for the average col
There was a problem hiding this comment.
Tuple was something with typing.
evap/results/exporters.py
Outdated
| avg_values = [] | ||
| count_avg = 0 | ||
| avg_approval = [] | ||
| count_approval = 0 |
There was a problem hiding this comment.
Same performance consideration here: Please don't build lists, just keep track of two numbers instead
evap/results/exporters.py
Outdated
| results.get(questionnaire_id) is None | ||
| ): # ignore all results without the questionaire for average calculation | ||
| continue | ||
| avg, percent_approval = cls._calculate_display_result(questionnaire_id, question, results) |
There was a problem hiding this comment.
@janno42 how much is performance a concern with this exporter? For production, I can see this making the exporter take 7x longer.
There was a problem hiding this comment.
It's not time critical but shouldn't take longer than ~10 seconds (with ~100 evaluations in the semester).
evap/results/exporters.py
Outdated
| # first cell of row is printed above | ||
| self.write_empty_row_with_styles(["border_left_right"] * len(evaluations_with_results)) | ||
|
|
||
| for question in self.filter_text_and_heading_questions(questionnaire.questions.all()): | ||
| self.write_cell(question.text, "italic" if question.is_heading_question else "default") | ||
|
|
||
| question_average, question_approval_count = self._calculate_display_result_average( |
There was a problem hiding this comment.
Suggestion: average_grade and approval_ratio?
evap/results/exporters.py
Outdated
| if question_average is not None: | ||
| if question.is_yes_no_question: | ||
| self.write_cell(f"{question_approval_count:.0%}", self.grade_to_style(question_average)) | ||
| else: | ||
| self.write_cell(question_average, self.grade_to_style(question_average)) | ||
| else: | ||
| self.write_cell("", "border_left_right") |
There was a problem hiding this comment.
Can be unnested:
if approval_ratio is not None:
self.write_cell(...)
elif average_grade is not None:
self.write_cell(...)
else:
self.write_cell(...)
In what situation can we have a question result here that doesn't have an average? We're putting the result here, so there must be at least our result, so there should always be an average, or am I missing something?
evap/results/exporters.py
Outdated
| if ( | ||
| results.get(questionnaire_id) is None | ||
| ): # ignore all results without the questionaire for average calculation | ||
| continue |
There was a problem hiding this comment.
From the code, it is very clear to me that this snippet skips results without a questionnaire. I'd much more appreciate the comment telling me 1. why this can happen in the first place (aren't all results always mapped to some question which needs to be part of some questionaire?) and 2. why we want to skip them
fidoriel
force-pushed
the
average-grades-in-export
branch
from
dced350 to
6163186
Compare
closes #2001