Less fp's

[ghost]

• Excludes StackOverflow, Maths, Mathoverflow and Cross Validated from the "post is mostly images" reason (since StackOverflow can have <img> code counted and the other 3 have a lot of MathJax used)
• Adds Cross Validated to the exclusion list for the "mostly punctuation marks in {}" reason (MathJax)

Statistics:

Excluding StackOverflow, Maths, Mathoverflow and Cross Validated from the "post is mostly images" reason, will result in:

• 31 fewer fp's
• 0 fewer tp's

The current accuracy of this reason is 17% (17) New accuracy: 40% (40)

Excluding Cross Validated from the "mostly punctuation marks in {}" reason, will result in:

• 0 fewer tp's (all tp's caught by other reasons)
• 30 fewer fp's

[ghost]

Note: The failures are not because of my code but because of how the tests are set up.

Edit: Fixed now

[user12986714]

Less fp for the mostly-img reason would be great, but I don't think excluding sites is optimal approach. See https://github.com/Charcoal-SE/SmokeDetector/pull/4190

[ghost]

Less fp for the mostly-img reason would be great, but I don't think excluding sites is optimal approach. See #4190

MathJax would still get caught though

[NobodyNada]

Excludes StackOverflow, Maths, Mathoverflow and Cross Validated from the "post is mostly images" reason (since StackOverflow can have code counted and the other 3 have a lot of MathJax used)

Isn't there a stripcodeblocks option that would help on Stack Overflow? As for the math sites...as far as I can tell, MathJax doesn't render as images, it just embeds the text in a <span class="math-container"> (which is rendered by client-side JS).

[ghost]

Excludes StackOverflow, Maths, Mathoverflow and Cross Validated from the "post is mostly images" reason (since StackOverflow can have code counted and the other 3 have a lot of MathJax used) Isn't there a stripcodeblocks option that would help on Stack Overflow? As for the math sites...as far as I can tell, MathJax doesn't render as images, it just embeds the text in a (which is rendered by client-side JS).

I have updated the PR and it has nothing to do with SO now, also the issue is MathJax posted as images and not actual MathJax.