Merchant Fraud Score (OA)

OA. Given transactions, fraud-scoring rules, and a baseline merchant score, compute each merchant's final fraud score by applying rule contributions over their transactions. Requirements Inputs: transaction = ["merchant1,980,customer1,15", ...] — each row contains merchant_id, amount, customer_id, and a timestamp. rules = ["1500,2,8,12", ...] — each rule contains a threshold and a set of additive score numbers (e.g. +2, +8, +12) keyed off some property of the transaction. merchants = ["merchant1,20", ...] — base score per merchant. Output: ["merchant1,50", ...] — final score per merchant after applying all matching rules across that merchant's transactions. Notes Prompt is long; the rule grammar is the most ambiguous piece — exact rule semantics are not fully recovered from community posts. Estimated 50 lines of code per the source report. Preparation Drill rule-engine problems: parse a list of rules, group transactions, accumulate per-entity totals. Pre-write a small apply_rule(transaction, rule, current_score) skeleton so you can compose rules in Part 2. Practice CSV-by-hand parsing and integer arithmetic; do not assume floats.