Explain Python code

What this Python sample shows

decorators

@timed above total_by_city is shorthand for total_by_city = timed(total_by_city). Every call now goes through wrapper, which runs the real function, prints how long it took and returns its result.

generators

read_rows uses yield, so it hands back one row at a time instead of building a list. A large file never has to fit in memory at once.

comprehensions

The last line is a dict comprehension: it builds a new dictionary from totals, in city order, with each total rounded to 2 decimal places.

Good to know

• @wraps(func) copies the original name and docstring onto wrapper. Without it, total_by_city.__name__ would be "wrapper", which makes logs and tracebacks harder to read.
• for city, amount in read_rows(path) unpacks each row into exactly two values. A row with more or fewer commas raises ValueError.
• totals.get(city, 0) returns 0 the first time a city appears. collections.defaultdict(float) does the same job without the .get.
• round(total, 2) only changes what is returned. The running totals are plain floats, so use decimal.Decimal when the amounts are money that must add up exactly.

Watch out: A header row such as city,amount crashes the function, because float("amount") raises ValueError. Skip it with next(f) before the loop, or read the file with the csv module.