This is a tool to compute various percentiles from given text files. This tool also incorporate streaming-percentiles, which is intended for extremely high-throughput data systems.
Use 'cmake' and 'make' to build the executable.
mkdir build
cd build
cmake .. && make
To run unit-test after build
ctest
If you build this on Windows, you may need dirent.
The executable takes a config and a path as its inputs.
executable test/config_1.txt test/unit_1
The config file specifies a regex to match files in the given path and another regex for matching the line format in files. Please see examples in test folder.
To utilize streaming-percentiles, where epsil 5E37 on is a float point value between 0 and 1.
executable -e epsilon test/config_1.txt test/unit_1
Config file contains pairs of key-value sets with keywords:
FILE_REGEX # Regular expression for filtering files
LINE_REGEX # Regular expression parsing line format
GROUPID # Specify the target token of the parsed result
PERCENTILE # List intended percentiles to output
Please refer to config files under test folder.
This project is licensed under the MIT License. LICENSE