Python · pytest · Zero dependencies

Know your gaps. Write the missing tests.

code-covered reads your coverage.json, walks the AST to understand context — exception handlers, branches, loops — then generates prioritized, ready-to-use test stubs. Not just what lines are missing. What tests to write.

Get started Read the Handbook

Quick start

pip install code-covered # Run tests with coverage JSON pytest --cov=myapp --cov-report=json # Find what you're missing code-covered coverage.json

Example output

Coverage: 74.5% (35/47 lines) Files analyzed: 1 (1 with gaps) Missing tests: 4 [!!] CRITICAL: 2 [! ] HIGH: 2 Top suggestions: 1. [!!] test_validator_validate_input_handles_exception In validate_input() lines 23-27 — when ValueError is raised 2. [!!] test_validator_parse_data_raises_error In parse_data() lines 45-45 — raise ParseError

Generate stubs

# Write test stubs directly to a file code-covered coverage.json -o tests/test_gaps.py # Filter to critical gaps only code-covered coverage.json --priority critical # JSON output for CI pipelines code-covered coverage.json --format json # Show full test templates code-covered coverage.json -v

More than a coverage report

It understands your code. Not just which lines are missing.

AST-aware context

Walks the source AST to understand what each uncovered block actually does — is it an exception handler? A branch condition? A loop body? The suggestion reflects the context.

Prioritized by risk

Critical first: exception handlers and raise statements that were never triggered. Then branches, then loops. The gaps most likely to break production surface at the top.

Zero dependencies

Pure Python stdlib. No heavy installs, no API calls, no runtime magic. Works anywhere pytest-cov produces a coverage.json. Install once, run anywhere.

Priority levels

Every gap is classified before it's surfaced.

Priority

Triggered by

Example

Critical

Exception handlers, raise statements

except ValueError: never triggered

High

Conditional branches

if x > 0: branch never taken

Medium

Function bodies, loops

Loop body never entered

Low

Other uncovered code

Module-level statements

Get started

Install & run

pip install code-covered

# Generate coverage.json
pytest --cov=myapp --cov-report=json

# Analyze gaps
code-covered coverage.json

# Verbose: show full templates
code-covered coverage.json -v

Generate test stubs

# Write stubs to a file
code-covered coverage.json -o tests/test_missing.py

# Filter by priority
code-covered coverage.json --priority critical
code-covered coverage.json --priority high

# Limit output
code-covered coverage.json --limit 5

CI integration

# JSON output for scripts and pipelines
code-covered coverage.json --format json

# Exit codes
# 0 — success (gaps found or clean)
# 1 — error (file not found, parse error)

# Specify source root if paths are relative
code-covered coverage.json --source-root ./src

Python API

from analyzer import find_coverage_gaps

# Find gaps programmatically
suggestions, warnings = find_coverage_gaps("coverage.json")

for s in suggestions:
    print(f"{s.priority}: {s.test_name}")
    print(f"  Lines: {s.covers_lines}")
    print(f"  Template:\n{s.code_template}")

How it works

Five steps from coverage report to ready-to-paste test stub.

Parse & map

Reads coverage.json from pytest-cov. Parses source files with stdlib ast to map every uncovered line back to the function, class, and block it lives in.

Detect & template

Classifies each gap — exception handler, branch, loop, raise statement — and generates a specific, context-aware test template with Arrange/Act/Assert scaffolding and setup hints.

Prioritize & emit

Ranks by risk and outputs to terminal, test stub files, or JSON for CI. Drop the generated stubs into your test suite and fill in the TODOs.