Implement minimal coerce and corresponding tests

[deepyaman]

This builds on top of #1766 and serves to:

1. demonstrate how we could generate Ibis data using Hypothesis (by way of Polars)
2. implement basic coerce functionality

[deepyaman]

@datajoely FYI I ended up going down the rabbit hole you mentioned myself:

However, I found the Ibis memtable strategy (leveraging the Hypothesis pandas extra) quite limiting—at least, I couldn't get it working for most types. I found it easier to just wrap the nice Polars strategy. 😁

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 92.78%. Comparing base (4df61da) to head (4525218).
Report is 148 commits behind head on ibis-dev.

Additional details and impacted files

@@             Coverage Diff              @@
##           ibis-dev    #1772      +/-   ##
============================================
- Coverage     94.29%   92.78%   -1.51%     
============================================
  Files            91      130      +39     
  Lines          7024     9458    +2434     
============================================
+ Hits           6623     8776    +2153     
- Misses          401      682     +281     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[deepyaman]

Must have made a mistake when I rebased thus, looking at the diff; let me check...

Oh, I forgot to include #1766, even though this was build on top of that, hence the failures. Since that's merged, I just rebased onto ibis-dev; hopefully it will pass. 🤞

[deepyaman]

CI issue fixed, should be good to go!

[cosmicBboy]

LGTM @deepyaman! Just have a quick question about the test

[cosmicBboy]

why the conversion to polars dtype?

[deepyaman]

https://github.com/deepyaman/pandera/blob/45252183baab6b91416f7373a302ca40979c1e55/tests/ibis/test_ibis_dtypes.py#L30 arguments are passed directly to dataframes strategy from Polars, and then I'm just constructing the ibis.memtable from a Polars dataframe.