Combining rows in pandas

I have a DataFrame with an index called city_id of cities in the format [city],[state] (e.g., new york,ny containing integer counts in the columns. The problem is that I have multiple rows for the same city, and I want to collapse the rows sharing a city_id by adding their column values. I looked at groupby() but it wasn't immediately obvious how to apply it to this problem.

Edit:

An example: I'd like to change this:

city_id    val1 val2 val3
houston,tx    1    2    0
houston,tx    0    0    1
houston,tx    2    1    1

into this:

city_id    val1 val2 val3
houston,tx    3    3    2

if there are ~10-20k rows.