Test Generation Workflow

What It Does

TestGenerator analyzes Python source files using AST parsing and generates pytest-structured test case skeletons. For each public function it produces a basic unit test, an edge-case test (when parameters exist), and an error-handling test. For public class methods it generates a unit test that instantiates the class.

CLI

chimera testgen --source ./chimera --output tests/generated

Python API

from chimera.testgen import TestGenerator, TestCase

generator = TestGenerator()

# Analyze a single file
cases = generator.analyze("chimera/core/agent.py")
for case in cases:
    print(f"{case.name} ({case.category}): {case.target_function}")
    print(case.test_code)

Analyze from source string directly:

source = '''
def add(a, b):
    return a + b
'''
cases = generator.analyze_source(source, filepath="math_utils.py")
# Generates: test_add (unit), test_add_edge_empty (edge), test_add_error (error)

Access all accumulated test cases:

all_cases = generator.test_cases
generator.clear()  # reset accumulated cases

Key Classes

TestGenerator

class TestGenerator:
    def __init__(self) -> None
    def analyze(self, filepath: str) -> list[TestCase]
    def analyze_source(self, source: str, filepath: str = "<unknown>") -> list[TestCase]
    def clear(self) -> None

Properties: test_cases (list of all accumulated TestCase instances).

TestCase

Dataclass with fields: name (str), target_function (str), target_file (str), test_code (str), category (str: "unit", "edge", or "error").

CoverageReport

Dataclass with fields: total_statements, total_missing, coverage_percent, file_coverage (dict mapping filepath to percent), uncovered_lines (dict mapping filepath to line numbers).

Properties: covered_statements. Methods: files_below(threshold) -> list[str].

parse_coverage(output: str) -> CoverageReport

Parses pytest-cov / coverage.py text output into a CoverageReport.

Import

from chimera.testgen import TestGenerator, TestCase, CoverageReport, parse_coverage