Update number of error occurrences reported + Bug fix in HTML reporter

[sarahsonder]

Motivation and Context

Currently, PythonTA limits the maximum number of the times the same error can shown at 5 errors. However, students have requested for all error occurrences to be shown. Furthermore, there was a bug in the HTML reporter which displayed all error occurrences despite the reporter stating that only a limited number of occurrences were being shown.

Your Changes

Description:

1. Changed the default value of pyta-number-of-messages to 0. This by default displays all instances of the same error.
2. Updated the behaviour of plain_reporter so that it limits the number of error occurrences ONLY when pyta-number-of-messages is not 0.
3. Added limiting behaviour to html_reporter so that its behaviour mirrors that of plain_reporter (previously html_reporter was not limiting the number of errors being shown).
4. Updated the conditions for "First shown" to be displayed only when pyta-number-of-messages is not 0 and the number of error messages exceeds pyta-number-of-messages.
5. Fixed typo in the Configuration page of the PythonTA documentation.

Type of change (select all that apply):

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)
• Documentation update (change that modifies or updates documentation only)

Testing

I tested my code by running both plain_reporter and html_reporter on three cases:

1. Setting pyta-number-of-messages = 0 to check that all error occurrences are being shown
2. Setting pyta-number-of-messages = 2 and running PythonTA on a file that has 5 of the same errors. This was to check that PythonTA was limiting the number of messages shown.
3. Setting pyta-number-of-messages = 6 and running PythonTA on a file that has 5 of the same errors. This was to check that PythonTA would still display all error occurrences.

Questions and Comments

I noticed that the Configuration page in the PythonTA documentation is under construction. Is there anything that I can add to the page?

Checklist

• I have performed a self-review of my own code.
• I have verified that the CI tests have passed.
• I have reviewed the test coverage changes reported on Coveralls.
• I have updated the CHANGELOG.md file.

[coveralls]

Pull Request Test Coverage Report for Build 5449391557

• 33 of 50 (66.0%) changed or added relevant lines in 3 files are covered.
• 19 unchanged lines in 1 file lost coverage.
• Overall coverage increased (+0.5%) to 89.909%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
python_ta/reporters/templates/template.html.jinja	31	48	64.58%

Files with Coverage Reduction	New Missed Lines	%
python_ta/reporters/templates/template.html.jinja	19	67.83%

Totals
Change from base Build 5449285850:	0.5%
Covered Lines:	1782
Relevant Lines:	1982

---

💛 - Coveralls

[david-yz-liu]

@sarahsonder great work! I just left a few minor comments, in addition to my earlier feedback of adding some test cases. 😄

That said, I also realized that I said to use the JSON reporter for testing, but based on the changes you made, I'm not sure whether this reporter is actually using this configuration at all? If not, it would be great if you fixed that in this PR as well.

[sarahsonder]

@david-yz-liu Hi Prof. Liu! I have updated the JSON reporter behaviour to limit the number of error messages and have added tests for my changes. The tests themselves work and the JSON reporter also working when I run PyTA on other files. However, the git tests are not passing due to a type error ('type' object is not subscriptable) caused by the helper function that I wrote in the JSON reporter. Could I please get some advice on how to resolve this error? Thank you!

[sarahsonder]

@david-yz-liu Issue fixed! I needed to use List and Dict instead of list and dict.

[david-yz-liu]

It isn't necessary to define this separately from unique_msg_names. You can instead use a set comprehension to automatically remove duplicates.

(But also see my below comment for additional suggestions for this part.)

[david-yz-liu]

This code is okay but is less efficient than it could be. Instead try this approach:

• loop over just msgs
• use a dict accumulator mapping error msg_id to the count of that message type "seen so far" in the loop. Note that you can initialize the accumulator using all of the unique keys (see above comprehension) with initial corresponding value 0
• then inside the loop use the accumulator to decide whether to append a message to output_lst or not
• after that loop, use a second loop over output_lst that sets the number_of_occurrences for each output message---we don't need to bother with updating this for messages that aren't included in the output!

[david-yz-liu]

@sarahsonder I left a review but accidentally selected "Approve" instead of "Comment". So just to be clear, please address the review comments I left and then re-request a review. Thanks! 😄