Add initial pandas 3 support

[larsevj]

No description provided.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 83.45%. Comparing base (31f65ed) to head (518650f).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #844      +/-   ##
==========================================
- Coverage   83.46%   83.45%   -0.01%     
==========================================
  Files          49       49              
  Lines        7293     7291       -2     
==========================================
- Hits         6087     6085       -2     
  Misses       1206     1206              

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

🚀 New features to boost your workflow:

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

[Copilot]

Pull request overview

This pull request adds initial support for pandas 3 by addressing compatibility issues in test assertions and data type handling. The changes primarily focus on making the codebase work with both pandas 2 and pandas 3 by relaxing strict type checking in tests and simplifying data conversion logic.

Key Changes:

• Updated test assertions to use check_index_type=False, check_dtype=False, and check_column_type=False parameters to accommodate pandas 3's different type behaviors
• Refactored compute_date_from_days function to use more explicit logic and avoid deprecated pandas behaviors
• Simplified convert_dframe_date_to_str by removing infer_objects and future.no_silent_downcasting context manager

Reviewed changes

Copilot reviewed 9 out of 9 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
tests/test_ofmvol2csv.py	Added check_index_type=False to test assertions to handle index type differences in pandas 3
tests/test_fmuobs_writers.py	Added check_dtype=False to accommodate dtype variations between pandas versions
tests/test_fmuobs_parsers.py	Added check_dtype=False to test assertions for type compatibility
tests/test_fmuobs.py	Added check_dtype=False to yaml roundtrip test assertions
tests/test_csv2ofmvol.py	Added check_index_type=False to handle index type differences
tests/test_check_swatinit.py	Added check_column_type=False for column type compatibility
src/subscript/fmuobs/writers.py	Removed deprecated infer_objects call and option context manager
src/subscript/fmuobs/parsers.py	Refactored date computation logic to be more explicit and removed numpy import
pyproject.toml	Updated pandas dependency to require version 2 or higher

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The docstring update mentions that 'datetime64 unit depends on pandas' but this is imprecise. Consider clarifying whether this refers to the resolution (e.g., ns vs. us) and documenting which pandas versions use which units, or stating that the specific unit should not be relied upon by consumers of this function.

Suggested change

	(datetime64 unit depends on pandas)
	(typically ``datetime64[ns]``; the exact time unit/resolution is
	determined by pandas and should not be relied upon by callers).

[Copilot]

The refactored logic computes dates for all rows (including those with NaN in DAYS column), which differs from the original behavior that only computed dates for date_needed_rows. While pandas handles NaN gracefully in timedelta operations, this change processes more data than necessary. Consider adding .where(dframe['DAYS'].notna()) to maintain the original selective computation behavior.

Suggested change

	computed_dates = start + pd.to_timedelta(dframe["DAYS"], unit="D")
	computed_dates = (start + pd.to_timedelta(dframe["DAYS"], unit="D")).where(
	dframe["DAYS"].notna()
	)