Viper - a code complexity metrics reporter

Input

all kinds of source code files

Output

For each file, list of functiton information. For a function, analyze to get following data: - scope (namespace, class or function as ancester nodes of this function)

• fqdn (fully qualified domain name: scope + name)
• name (the identifier)
• signature (normalized function signature, keep unchanged when compiled function name is not changed)
• line (line number of identifier)
• start (start line number of function definition)
• end (end line number of function definition)
• ccm (mccabe-style cyclomatic complexity metric, Traditional: each case statement counts +1)
• modified_ccm (mccabe-style cyclomatic complexity metric, Modified: switch statement counts +1)
• statements (number of statements in the function body)

Supported Languages

• C (.c, .h)
• C++ (.cpp, .cxx, .cc, .hpp, .hxx, .hh)
• Python (.py, .pyw)
• SystemVerilog (.sv, .svh)

Output Formats

Viper supports multiple output formats out of the box. Use the --format option:

• default: Human-readable list of functions (default)
• pmccabe: pmccabe-compatible tab-separated format (Modified CCM, Traditional CCM, Statements, First line, # Lines, filename(line): function_name)

Implementation Idea

• use python 3.14.2
• utilize tree_sitter module and tree_sitter_xxx language packs to parse source code to get AST, then analyze AST to get function information
• support inputs of multiple files, enumeration of files in directories.

project and module management

• use tool uv as package/project management tool
• use uv run <script> to run scripts
• put code under src file