Compute Graph Spread and Portfolio Trades

Problem Statement

You are given a list of trades representing rebalancing actions and a list of correlations representing assets that must be balanced together.

Each trade is a tuple (asset_id, amount), where amount is a signed integer representing the net change in that asset's position. The correlations list contains pairs (asset_a, asset_b), indicating that these two assets belong to the same 'risk group'. Any change in one asset within a risk group must be offset by the aggregate changes of other assets in that same group.

Your task is to identify all connected risk groups and return the net delta for each group. The net delta of a group is the sum of the amount values of all assets within that group. If an asset has no correlations, it forms a group of size one.

Function Signature

def compute_group_deltas(trades: List[Tuple[str, int]], correlations: List[Tuple[str, str]]) -> Dict[str, int]

Constraints

• Number of assets: 1 to 10^5
• Number of correlations: 0 to 10^5
• amount: -10^9 to 10^9

Examples

Example 1:

• trades = [('AAPL', 10), ('MSFT', -5), ('GOOG', 20)]
• correlations = [('AAPL', 'MSFT')]
• Output: {'AAPL_MSFT': 5, 'GOOG': 20}
• Explanation: AAPL and MSFT are linked (10 + -5 = 5). GOOG is independent (20).

Example 2:

• trades = [('A', 1), ('B', 2), ('C', 3)]
• correlations = [('A', 'B'), ('B', 'C')]
• Output: {'A_B_C': 6}
• Explanation: All three are in one component (1 + 2 + 3 = 6).