Donghyeok built an image character-recognition program. It scans the letters written on a sheet of paper and saves them to a text file. The program is not 100% accurate, so it fails to recognize some characters.
Write a program that computes the recognition rate of this program.
Let $R$ be the number of recognized characters and $A$ the total number of characters. The recognition rate is $R / A$. Newline characters are not counted as characters.
The input consists of several test cases. The first line contains the number of test cases, $N$.
Each test case spans one or more lines, and every unrecognized character is shown as #. A single blank line follows each test case.
Each line is at most 100 characters long, and a single test case has at most 200 lines.
For each test case, compute the recognition rate and print it on its own line in the following format.
Efficiency ratio is X%.
Here $X$ is the recognition rate as a percentage ($R / A \times 100$), shown to one decimal place. Round at the second decimal place, rounding a value that is exactly halfway up. If the rounded result is a whole number, omit the decimal point and print only the integer part.