[feature/changelog] Order Change Types Based on Bump Map Rules

[KyleKing]

Description

The CHANGELOG is ordered with the most recent change type first. This makes the diffs hard to read and may put less important change types over more important ones

Example diff from one of my projects: KyleKing/calcipy@a61d0a4#diff-06572a96a58dc510037d5efa622f9bec8519bc1beab13c9f251e97e657a9d4ed

Possible Solution

Could the CHANGELOG be ordered based on the bump rules?

The CHANGELOG could be ordered in descending importance from (MAJOR, MINOR, PATCH, OTHER) where the types within each section would be sorted alphabetically and the OTHER section would capture any change types shown in the CHANGELOG, but not part of the bump rules

Related Issue

I looked through the issues (matching order or changelog) but didn't find any that appear to be asking for the same feature

PR Notes

I would be happy to try to add this functionality, if this feature would be of interest to the maintainers. If not, I may look into separate customization as an alternative to standard functionality)

I created a cz_legacy customization last night, so I have some familiarity with the commitizen code base

[woile]

Hello Kyle! Thanks, that would be an awesome feature, feel free to create a PR with it. Also, you can add your custom rule to our docs in here: https://commitizen-tools.github.io/commitizen/third-party-commitizen/

[KyleKing]

I found a few small questions/issues that seemed too trivial for a whole issue:

• For the issues template, python -c "import platform; print(platform.platform())" might provide more useful debugging information (i.e. macOS-10.15.7-x86_64-i386-64bit vs. Darwin)
• In my fork, the format script made changes to import order. I found that setting commitizen and tests as known first parties for isort kept the correct import order
• (WIP) I ran into issues with Unicode characters in commit messages on Windows. I need to revisit and fully understand why open was defaulting to cp1252 (I thought Python 3 would use Unicode by default?). Initially, specifying the encoding appears to fix the issue, but I'm not sure if that is the best fix (with open(self.file_name, "w") as changelog_file:)