Your question is Parse Logs for Error Patterns. Start with the requirements on the right.
Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.
Pinterest's infrastructure team wants to summarize error logs from services supporting PinLater. Each log line uses the format timestamp|service|level|message. Implement a function that parses valid lines, keeps only entries whose level is exactly ERROR, and counts which configured error patterns occur in each entry.
A pattern matches when it appears as a case-insensitive substring of the message. If one message contains the same pattern multiple times, count that pattern only once for the message. One message may match multiple patterns. Ignore malformed lines, lines with a non-ERROR level, and patterns that never match.
Return a list of result objects sorted by descending count, then ascending pattern text. Each object must contain the original pattern, the number of distinct error log lines containing it, and the sorted list of affected services.
logs, a list of strings, and patterns, a list of non-empty strings.pattern, count, and services.ERROR.def analyze_error_logs(logs, patterns):