Remove Par typeclass

[purrgrammer]

After discussing with @peterneyens how we could get rid of the Par typeclass he suggested using Concurrent#start and Fibers to achieve parallelism whenever we used parTraverse. This is my first approach to implementing it.

• Removes cats-par dependency
• Removes Par implicit requirement
• Implements parallelism in terms of Concurrrent#start and Fiber's
• Support cancellation of parallel fetches

[codecov]

Codecov Report

Merging #166 into master will not change coverage.
The diff coverage is 100%.

@@          Coverage Diff           @@
##           master    #166   +/-   ##
======================================
  Coverage    95.2%   95.2%           
======================================
  Files           5       6    +1     
  Lines         167     167           
  Branches        4       7    +3     
======================================
  Hits          159     159           
  Misses          8       8

Impacted Files	Coverage Δ
shared/src/main/scala/syntax.scala	100% <ø> (ø)	⬆️
shared/src/main/scala/cache.scala	100% <ø> (ø)	⬆️
shared/src/main/scala/datasource.scala	100% <100%> (ø)	⬆️
shared/src/main/scala/fetch.scala	95.39% <100%> (-0.04%)	⬇️
shared/src/main/scala/execution.scala	100% <100%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update fcaac28...f803eab. Read the comment docs.

[purrgrammer]

I'm not sure if we can support running parallel computations and making them cancelable, since the Concurrent#cancelable looks like the following:

import cats.effect.{Async, Fiber, IO, CancelToken}

trait Concurrent[F[_]] extends Async[F] {
  // ...
  def cancelable[A](k: (Either[Throwable, A] => Unit) => CancelToken[F]): F[A]
}

For now I've made the fibers we spawn uncancelable, so in case the fetch computation is cancelled, it will terminate as soon as all concurrent fibers have finished executing.

[purrgrammer]

I have some ideas for supporting cancellation of parallel computations, I'll update the PR when I manage to implement it.

[raulraja]

Looks great so far, let me know when ready for final review. 🍻

[purrgrammer]

@raulraja not sure why the coverage doesn't cover the newly introduced execution.scala file but this PR is ready for review!

[raulraja]

This looks great, if @peterneyens has no additional comments I think this is good to go. kudos and great work!

[raulraja]

Formatting seems off in a few match cases here. Are we enforcing auto formatting with Scalafmt or similar?

[purrgrammer]

Not right now, I can open a separate PR for adding scalafmt and not polluting this one

[purrgrammer]

Cleaned this up following @peterneyens suggestions, let me know if ✔️

[jtjeferreira]

any chance to make a release (RC or milestone) with this PR included? Would like to try this for a small project...