fix: Support historical feature retrieval with start_date/end_date in RemoteOfflineStore

[aniketpalu]

What this PR does / why we need it:

Problem

• get_historical_features() with start_date/end_date parameters (no entity_df) works in local mode but fails in remote mode with:
  • EntityTimestampInferenceException: Please provide an entity_df with a column named event_timestamp

Root Cause

• Remote client sends mock table {"key": "mock_key"} when entity_df is None
• Server treats mock table as valid entity_df, which lacks event_timestamp column
• Server doesn't forward start_date/end_date to underlying offline store

Changes

• Client (remote.py):
  • Handle None entity_df in _create_retrieval_metadata() and _get_entity_schema()
• Server (offline_server.py):
  • Detect and discard mock tables (single 'key' column)
  • Create minimal entity_df with event_timestamp from start_date/end_date
  • Forward start_date/end_date kwargs to underlying offline store
  • Use pd.to_datetime() and utils.make_tzaware() for proper timestamp types

Which issue(s) this PR fixes:

Misc

[ntkathole]

Looks good! @aniketpalu can you please rebase to resolve conflicts ?

[jyejare]

Handle empty entity_df in offline server.

[jyejare]

We should not need to create dummy entity_df at all. We should handle empty(entity_df=None) entity_df for and in offline_store.get_historical_features which we are returning in this function.

[aniketpalu]

You are right, its not needed to create mock dataframe with timestamps, just had to make sure that empty_df is being passed.