Update type_tableview to support list aggregates

[Haigutus]

Optional flag for type_tableview and table_view to graph to support list aggrgates in pivots and in unmeld

[Haigutus]

# rdf_parser.py - updated type_tableview (issue #24)
# support list aggregates for multi-valued RDF props (e.g. Terminal in ACLineSegment, DependentOn etc.)
import pandas as pd
import logging

logger = logging.getLogger(__name__)

def type_tableview(data, type_name, string_to_number=True, type_key="Type", list_aggregates=False):
    """Type tableview for all objects of one CIM class.
    list_aggregates=True collects duplicate (ID, KEY) into lists (pivots safely).
    Returns
    -------
    pd.DataFrame
        Index = object ID, columns = properties (strings/numbers or lists)
    
    Example table (list_aggregates=True):
               name          Terminal.ConnectivityNode
    ID
    _acls1    "Line1"       ["_nodeA", "_nodeB"]
    _acls2    "Line2"       ["_nodeC"]
    """
    # Get IDs of the requested type (fast filter)
    type_id = data[(data["VALUE"] == type_name) & (data["KEY"] == type_key)]
    if type_id.empty:
        logger.warning(f'No data available for {type_name}')
        return None

    # All triplets belonging to these objects
    type_data = pd.merge(type_id[["ID"]], data, on="ID")

    # Key step: aggregate duplicates (multi-value support)
    if list_aggregates:
        type_data = (type_data
            .groupby(["ID", "KEY"], as_index=False)["VALUE"]
            .agg(list)
        )
    else:
        type_data = type_data.drop_duplicates(["ID", "KEY"])

    # Pivot to wide table (now works with list cells)
    table = type_data.pivot(index="ID", columns="KEY", values="VALUE")

    if string_to_number and not list_aggregates:
        for col in table.columns:
            try:
                table[col] = pd.to_numeric(table[col], errors="raise")
            except (ValueError, TypeError):
                pass  # keep object dtype (strings or future lists)

    return table

# attach as pandas extension (existing pattern)
pd.DataFrame.type_tableview = type_tableview

[Haigutus]

# add this helper (in rdf_parser.py or cgmes_tools.py)
def unmeld(table, instance_id=None):
    """Unmeld wide tableview back to long triplets.
    Explodes lists automatically.
    """
    long = (table
        .reset_index()
        .melt(id_vars="ID", var_name="KEY", value_name="VALUE")
        .explode("VALUE")          # handles list aggregates
        .reset_index(drop=True)
    )
    if instance_id is not None:
        long = long.assign(INSTANCE_ID=instance_id)
    return long[["ID", "KEY", "VALUE", "INSTANCE_ID"]].dropna(subset=["VALUE"])

# example round-trip
acls_wide = data.type_tableview("ACLineSegment", list_aggregates=True)
acls_triplets = unmeld(acls_wide, instance_id="my_instance")
data = data.update_triplet_from_triplet(acls_triplets)  # or append