[BDMS-123]: Add UTC normalization for date_time and measurement_date_time in well inventory importer

[TylerAdamMartinez]

Why

This PR addresses the following problem/context:

• measurement_date_time in the well inventory importer was not normalized to UTC before being written to sample.sample_date and observation.observation_datetime

How

Implementation summary - the following was changed/added/removed:

• Extracted a reusable normalize_datetime_to_utc() helper into a shared utility module
• Updated WellInventoryRow schema validators:
  • date_time is now normalized to UTC
  • measurement_date_time is now normalized to UTC
• Removed logic that rejected timezone-aware datetimes; now both naive and aware datetimes are supported
• Ensured naive datetimes are interpreted as America/Denver before conversion to UTC
• Minor updates to schemas/water_level_csv.py to reuse the shared helper and keep validator behavior consistent (no intentional changes to water-level importer behavior)

Notes

Any special considerations, workarounds, or follow-up work to note?

• Existing well inventory records may contain non-UTC datetimes; backfill or correction may be needed
• Shared helper can now be reused by other importers to enforce consistent datetime handling
• No database schema changes required

[ksmuczynski]

Looks great! The schema-level approach makes sense here. Just a few minor comments/suggestions, all optional.

1. Consider adding a note to the PR description clarifying that the updates to schemas/water_level_csv.py are just to keep the validator behavior consistent and reuse the shared helper. Just want to make it clear we're not intentionally changing the water-level importer's core behavior.
2. Since the PR now accepts timezone-aware datetimes as input, it'd be great to have an explicit WellInventoryRow test covering an aware date_time/measurement_date_time value alongside the existing naive-input coverage. That would help ensure the aware-datetime path stays covered going forward.
3. I noticed the well-inventory feature/spec still describes the datetime inputs as timezone-naive only (tests/features/well-inventory-csv.feature, line 143). Consider aligning the docs with the implementation to avoid any confusion down the line.

[TylerAdamMartinez]

Looks great! The schema-level approach makes sense here. Just a few minor comments/suggestions, all optional.

1. Consider adding a note to the PR description clarifying that the updates to schemas/water_level_csv.py are just to keep the validator behavior consistent and reuse the shared helper. Just want to make it clear we're not intentionally changing the water-level importer's core behavior.
2. Since the PR now accepts timezone-aware datetimes as input, it'd be great to have an explicit WellInventoryRow test covering an aware date_time/measurement_date_time value alongside the existing naive-input coverage. That would help ensure the aware-datetime path stays covered going forward.
3. I noticed the well-inventory feature/spec still describes the datetime inputs as timezone-naive only (tests/features/well-inventory-csv.feature, line 143). Consider aligning the docs with the implementation to avoid any confusion down the line.

• Add PR description node
• Add an explicit WellInventoryRow test covering an aware date_time/measurement_date_time value alongside the existing naive-input coverage
• Update the well-inventory-csv.feature