Typer looks neat. I don't think it was well-established when I started on this, there were quite a few offerings but the only one I managed to find was python-fire. It didn't quite do what my team wanted and there were a few behaviours that we didn't quite like so we decided to make this.
Having said that bargeparse was meant to rely solely on stdlib as the environment we were operating in was tightly controlled and any 3rd party packages had to go through a lengthy review process to be whitelisted.
Additionally, since our team (and others in our organisation) were mostly using argparse, the goal was to provide a way to declaratively interact with it using existing features in the language itself.
I used typer as an example because it has a modern twist, yet is based on click which has been the leader for 7 years.
If you want something older and with a similar api (wrapping argparse), using only the stdlib, I can ask the same question with argh, which already worked in python 2.6.
I can make a guess: GP does not know, he started working on the software and pointlessly duplicated existing projects, but with worse quality, because he neglected to survey the landscape.
I have noticed that happening regularly and out of curiousity interviewed those fellow programmers. I found out it's not because of lack of diligence, or because they attribute no value to a survey, but because of a risk-loving psychology trait that expresses itself in staking out ground and claiming it as one's own. (Armchair psychology warning, I am not qualified to make that conclusion; leave the real work confirming this hypothesis to the scientists in the field.)
IMO, this behaviour is harmful and we should work together to socially shun it like we already do for compulsive hoarding of physical objects. We have trouble finding a piece of software that works well because we spend so much effort to repeatedly sift through the mountains of published software that is mostly garbage.
Having said that bargeparse was meant to rely solely on stdlib as the environment we were operating in was tightly controlled and any 3rd party packages had to go through a lengthy review process to be whitelisted.
Additionally, since our team (and others in our organisation) were mostly using argparse, the goal was to provide a way to declaratively interact with it using existing features in the language itself.