🐛 Fix dtype degradation of partial-null bool/int features

[ishitajain9717]

Fixes a ValidationError in Record.from_dataframe(...).save() when a bool/int feature has None on some rows but not others. Internally the save path rebuilt the frame with pd.DataFrame(prepared_rows), which let pandas re-infer and degrade those columns (bool → object, int → float64), failing the dtype check even for correctly-typed nullable input.

The frame is now built column-by-column using each feature's declared dtype (convert_to_pandas_dtype), so dtypes never degrade. Adds tests covering both the DataFrameCurator (accepts nullable boolean/Int64, rejects degraded object/float64) and the from_dataframe save round-trip.

Fixing : pfizer-collab/laminlabs#1009

[codecov]

Codecov Report

❌ Patch coverage is 88.23529% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 86.50%. Comparing base (fce4357) to head (e1c610d).
⚠️ Report is 16 commits behind head on main.

Files with missing lines	Patch %	Lines
lamindb/models/_feature_manager.py	88.23%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3752      +/-   ##
==========================================
- Coverage   91.55%   86.50%   -5.06%     
==========================================
  Files          88       88              
  Lines       15137    15336     +199     
==========================================
- Hits        13859    13266     -593     
- Misses       1278     2070     +792     

Files with missing lines	Coverage Δ
lamindb/models/_feature_manager.py	89.72% <88.23%> (-0.05%)	⬇️

... and 19 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[github-actions]

Deployment URL: https://a411eaaf.lamindb.pages.dev

[falexwolf]

Can you please add a test that makes evident which error message is raised here?

[falexwolf]

Perfect!

[falexwolf]

Looks great!