Reduce memory consumption and increase performance for image comparison

[FoxxMD]

• [ ] Determine memory usage for individual image (both raw and loaded into resemble)
• [ ] Determine what impact converting image to smaller size has on cumulative memory usage, comparison speed, and general cpu usage
• [ ] Potentially refactor image comparison object usage (repeat/recent reference objects) based on results from above

[FoxxMD]

Turns out the majority of run time is due to downloading images. Also using "ignore colors" or setting a bounding box on comparison actually slowed comparisons down b/c resemble is doing more work in those instances...

Was able to reduce rule run time in f8fc63991f6b8f732878c722790ee7a3a07d9bff by:

• refactoring image fetch into an async getter (get on-demand)
• use reddit previews where possible so the downloaded image is a smaller resolution (if comparable to reference preview resolutions)
• do image comparisons in parallel using Promise.all instead of for-looping

Unfortunately comparisons in parallel is making memory usage skyrocket. Need to investigate p-map so i can set concurrency to something reasonable (or user-configurable)

[FoxxMD]

Fixed mostly everything by 976f310f512ed20d70dc900b4901e17658ebc18a