Logger Handling Multiple Input Types

MediumPython

Problem

Implement a custom logger for an Intercontinental Exchange Holdings ICE Data Services monitoring surface. The logger must filter entries by severity and format messages containing mixed JSON-compatible values consistently, so identical events produce identical output across clients.

Write custom_logger(entries, min_level).

Each entry is a dictionary with level, message, and context. message and context may be strings, integers, floating-point numbers, booleans, null, arrays, or nested objects. Return a list of formatted log lines for entries whose severity is at least min_level. Severity order is debug < info < warn < error.

Use this exact format: [LEVEL] SERIALIZED_MESSAGE | context=SERIALIZED_CONTEXT

Serialization rules match compact JSON: strings use double quotes with necessary escaping, booleans are true or false, null becomes null, arrays preserve order, and object keys are emitted in lexicographic order. Do not mutate the input.

Constraints

1 <= len(entries) <= 10^4
Nesting depth is at most 50
Values contain only JSON-compatible types
Object keys are strings
Input objects contain no cyclic references
min_level is one of debug, info, warn, or error

Function Signature

def custom_logger(entries, min_level):

Write custom_logger(entries, min_level).

Use this exact format: [LEVEL] SERIALIZED_MESSAGE | context=SERIALIZED_CONTEXT