Packages/test performance

[abandoned-prototype]

Status

Ready for review

Description of Changes

Changes proposed in this pull request:

• For some reason the mockdata used in the tests not only contains a lot of data, but it was freshly inserted into the database for every test that used it.
  • Now we will load the mockdata into the in-memory database only once (via the fixture with scope session). This makes tests significantly faster.
  • Adding an additional department to the mockdata broke some tests which had to be fixed.
  • Changing anything in the mockdata brakes tests because some of the data is assigned pseudo-randomly (random with fixed seed), so a test might fail because an officer used to not have a link in the profile but does after the change. I fixed the failing tests in a way that this does not happen again.

Notes for Deployment

Changes tests only

Tests and linting

• I have rebased my changes on current develop

• pytests pass in the development environment on my local machine

• pre-commit checks pass

[michplunkett]

Looks good! I just have some formatting suggestions, but they aren't blockers.

[michplunkett]

🔬 I think DEPARTMENT_NAME_XXXX makes more sense as a naming motif.

DEPARTMENT_NAME -> DEPARTMENT_NAME_SPRINGFIELD
DEPARTMENT_TWO_NAME -> DEPARTMENT_NAME_CHICAGO
DEPARTMENT_EMPTY_NAME -> DEPARTMENT_NAME_EMPTY.

[abandoned-prototype]

Changed it slightly, hopefully make it a little more sense. But I want to make it clearer which department is considered the default one.

[michplunkett]

When I see "empty" in the name, I presume it's an empty string.

[abandoned-prototype]

Good point. Changed it to DEPARTMENT_WITHOUT_OFFICERS.

[michplunkett]

🔬 An f-string feels like a more intuitive way to write this, imo.

[abandoned-prototype]

Strongly agree, this is from old code. Changed this one. And maybe we can use https://github.com/ikamensh/flynt or something next to get rid of all of the .format.

[michplunkett]

🔬 The syntax here doesn't feel incredibly intuitive.

[abandoned-prototype]

Tried to make it a little better, by extracting "the math" to a variable and adding a comment.

[michplunkett]

Since these are paired to specific short names, I think it may be valuable to put them in a dictionary.

DEPARTMENT_KEY_SPD = "SPD"
DEPARTMENT_KEY_CPD = "CPD"
DEPARTMENT_KEY_EPD = "EPD"

DEPARTMENTS = {
    DEPARTMENT_KEY_SPD: "Springfield Police Department",
    DEPARTMENT_KEY_CPD: "Empty Police Department",
    DEPARTMENT_KEY_EPD: "Chicago Police Department"
}

You can use the key as the short name (lower or upper case, whateva) and the value as the long name.

[michplunkett]

On second thought, tuples would work as well.

DEPARTMENT_SPD = ("SPD", "Springfield Police Department")
DEPARTMENT_CPD = ("CPD", "Chicago Police Department")
DEPARTMENT_EPD = ("EPD", "Empty Police Department")

[abandoned-prototype]

The department short name does not seem to matter much for tests (they are not used anywhere other than the initial creation) so I decided against giving each their own variable.