headers

Classifying function headers

Function headers consist of a function name followed by the parameter types in parentheses. As a crude heuristics, the bash script classify.sh uses regex-based rules to group function headers into categories.

Usage:

classify.sh <headers> <rules-1> <rules-2> ...

<headers> is a text file with one function header per line.

<rules-i> is a text file with one rule per line. Each rule has the form

<inclusion pattern>   <category>   <exclusion pattern (optional)>

A header is assigned to <category> if it matches <inclusion pattern> but not <exclusion pattern>.

classify.sh works in rounds. Initially, all rules in <rules-1> are applied simultaneously to the headers in <headers>. If a rule classifies a header as belonging to <category>, the header is added to the file categories/<category>. One header may end up in several categories, if several rules match. At the end of the round, all headers newly classified in this round are removed from the set of headers. Then <rules-2> is applied to the remaining headers, and so on. After the last round, all classified headers are merged into a single file categories.csv that contains each header with its category, separated by a semicolon.

Required files

headers/classify.sh
headers/rules1.txt
headers/rules2.txt

Try it out

cd ethutils/doc/headers  # go to the directory with the sample data
bash wrapper_classify.sh # apply classify.sh to sample headers

The file headers.csv contains 71673 headers from the contracts on Ethereum's main chain. After running the scripts, categories.csv contains the classified headers with their category.