When do we bump package versions and how do we automate the process?

[phutchins]

Currently we use Lerna to manually manage versioning. Lerna has the ability to automatically bump versions based on the Conventional Commits specification. Read more about that here and here.

The format that we are using is SEMVER with the short hash of the git commit on the end for pre-release versions.
For example the progression might be something like this:

0.0.1 - Starting the project
0.0.2-73e21c5.0 - A bug was fixed
0.0.3-339dbs3.0 - A second bug was fixed
0.0.3-fkd930s.1 - The bug actually wasn't fixed so another change was pushed to fix it
0.0.3 - The previous pre-release was graduated to a release (this git tag points to the exact same git hash as 0.0.3-fkd930s.1)
0.1.0-b0d8dse.0 - A feature was added
1.0.0 - A major release is cut signifying that breaking changes can no longer happen without increasing X.0.0

Fore more on semantic versioning, read here

Some of the questions that we should discuss here are...

1. Do we increase versions outside of the main branch? or only for main branch?
2. Can we use hooks to automatically update versions for us?
3. What is the best way to attach the short hash of the git commit to the version? (keep in mind that if you update the version in package.json and push that change, the commit hash changes) (we're currently using tags so we could simply use tags to capture that portion of the version)
4. How and where do we signify changes that affect the version number? I.E. do we use github PR tags? or commit prefixes? something else?