Skip to content

Technical Introduction

At a high-level, pymwp reads a C file and produces an analysis result.

pymwp workflow

Fig. 1: pymwp workflow

Under the hood, pymwp is built from various modules. Each module has a specific role. For example, Analysis runs program analysis (i.e., the intermediate steps between parser and output in Fig. 1); the matrix module contains matrix utilities, the parser is the front-end, etc.

This documentation section explains the internal behaviors of these building blocks. Understanding these modules becomes relevant when importing pymwp in scripts.

Scripting Examples

Reusing analysis output

This scenario describes reusing the pymwp analysis result in further computation.

  • An analysis run returns a Result object.
  • Function get_func(...) returns the analysis result of the specified function.
from pymwp import Analysis, Parser
from pprint import pprint

# path to file to analyze
file = 'c_files/basics/if.c'

# parses a C-langugage file using pycparser
ast = Parser.parse(file, use_cpp=True, cpp_path='gcc')

# run analysis, then access result for main function
result = Analysis.run(ast, fin=True, no_save=True).get_func('main')

# do something with the analysis result
# display analysis result and collected data
pprint(result.to_dict())

// contents of if.c
int main(int x, int y){
    if (0) {y = x;}
}

{'bound': {'x': 'x;;', 'y': 'x,y;;'},
'choices': [[[0, 1, 2]]],
'end_time': 1725498245384529000,
'index': 1,
'inf_flows': None,
'infinity': False,
'name': 'foo',
'relation': {'matrix': [[[{'deltas': [], 'scalar': 'm'}],
                         [{'deltas': [], 'scalar': 'm'}]],
                        [[{'deltas': [], 'scalar': 'o'}],
                         [{'deltas': [], 'scalar': 'm'}]]]},
'start_time': 1725498245383927000,
'variables': ['x', 'y']}

Build with analysis modules

Working directly with the internal modules allows customizing the analysis behavior and building related analyses.

from pymwp import Polynomial, Monomial
from pymwp.matrix import identity_matrix, show

matrix = identity_matrix(3)
matrix[0][1] = Polynomial('m')
matrix[1][1] = Polynomial('w')
matrix[2][1] = Polynomial(Monomial('p', (0, 0), (1, 1)))

show(matrix)

[' +m', ' +m', ' +o']
[' +o', ' +w', ' +o']
[' +o', ' +p.delta(0,0).delta(1,1)', ' +m']