fix: resolve bug on add properties from records related methods

[mcllerena]

No description provided.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 95.66%. Comparing base (c5d9a38) to head (41a380b).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #129      +/-   ##
==========================================
+ Coverage   95.64%   95.66%   +0.02%     
==========================================
  Files           8        8              
  Lines        1721     1729       +8     
==========================================
+ Hits         1646     1654       +8     
  Misses         75       75              

☔ 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.

[pesap]

Index source can drift here. metadata_map keys use row_index from normalized_records, but downstream maps and lookups key off enumerate(params). If a row is skipped before append, metadata can miss on later valid rows.

Suggested change:

Suggested change

	param_tuple = (membership_id, property_id, record["value"])
	row_key = (membership_id, property_id, record["value"], row_index)
	params.append(param_tuple)
	metadata_map[row_key] = {
	insert_index = len(params)
	param_tuple = (membership_id, property_id, record["value"])
	row_key = (membership_id, property_id, record["value"], insert_index)
	params.append(param_tuple)
	metadata_map[row_key] = {

Then use the same insert_index convention in later key lookups too (insert_property_values, apply_scenario_tags, _collect_text_rows).

[pesap]

Great coverage for duplicate value rows. Please add one regression with a skipped row before valid rows to lock the index alignment fix.

Suggested change:

Suggested change

	assert {row[3] for row in rows} == {"row1.csv", "row2.csv"}
	assert {row[3] for row in rows} == {"row1.csv", "row2.csv"}
	def test_add_properties_from_records_duplicate_rows_after_skipped_row_preserve_links(
	db_instance_with_schema: PlexosDB,
	) -> None:
	from plexosdb import ClassEnum, CollectionEnum
	db = db_instance_with_schema
	db.add_object(ClassEnum.Generator, "GEN_MIX")
	records = [
	{
	"name": "MISSING_OBJECT",
	"property": "Max Capacity",
	"value": 0.0,
	"band": 9,
	"datafile_text": "skip.csv",
	},
	{
	"name": "GEN_MIX",
	"property": "Max Capacity",
	"value": 0.0,
	"band": 1,
	"datafile_text": "row1.csv",
	},
	{
	"name": "GEN_MIX",
	"property": "Max Capacity",
	"value": 0.0,
	"band": 2,
	"datafile_text": "row2.csv",
	},
	]
	db.add_properties_from_records(
	records,
	object_class=ClassEnum.Generator,
	parent_class=ClassEnum.System,
	collection=CollectionEnum.Generators,
	scenario="ScenarioMix",
	)
	rows = db.query(
	"""
	SELECT d.data_id, b.band_id
	FROM t_data d
	LEFT JOIN t_band b ON b.data_id = d.data_id
	JOIN t_membership m ON d.membership_id = m.membership_id
	JOIN t_object o ON m.child_object_id = o.object_id
	JOIN t_property p ON d.property_id = p.property_id
	WHERE o.name = ? AND p.name = ?
	ORDER BY d.data_id
	""",
	("GEN_MIX", "Max Capacity"),
	)
	assert len(rows) == 2
	assert {row[1] for row in rows} == {1, 2}

[pesap]

Left two inline comments with concrete suggestions. Main blocker is row index key drift between normalized_records and params, which can drop metadata/text/tag links after skipped rows. Please fix index source and add a mixed valid plus skipped regression test.