constantipy 0.1.1

A tool to refactor magic strings and other magic constants into named constants.

Constantipy 🪄

The Magic Removal Tool.

Constantipy is a CLI tool that automatically refactors "magic" literals (strings, numbers, bytes) in your Python codebase into named constants.

It scans your project, identifies repeated literals, names them intelligently (e.g., "SELECT *" → STR_SELECT_SQL), and refactors your code to import them from a central constants.py or define them locally.

---

Features

• Zero Runtime Dependencies: Runs on standard library only.
• Smart Naming Heuristics:
  • Auto-detects SQL (STR_SELECT_SQL), URLs (HTTPS_GOOGLE_COM_URL), and Regex patterns.
  • Splits CamelCase (FieldState → STR_FIELD_STATE).
• Scope Detection:
  • Global: Strings used in multiple files are moved to constants.py.
  • Local: Strings used in only one file are defined at the top of that file (keeps your global constants file clean).
• Safety First:
  • Ignores f-strings (f"Hello {name}") to prevent breaking interpolation.
  • Ignores docstrings.
  • Checks for variable name collisions before generating constants.
  • Ignores structural numbers (0, 1, 2, -1) by default.
• Granular Control:
  • Ignore specific values (--ignore-str desc, --ignore-num 404).
  • Ignore specific types (--no-ints, --no-floats).
  • Ignore specific function calls (like logging.debug).

Installation

pip install constantipy

⚡️ Quick Example

Before Constantipy:

# db_utils.py
import time
import logging

def fetch_users(retries=3):
    # Magic String (SQL)
    query = "SELECT * FROM users WHERE active = 1"

    if retries > 0:
        # Magic String inside logging (can be ignored via config)
        logging.debug("Attempting connection...")
        # Magic Number
        time.sleep(60)

    return "Success"

After running constantipy apply:

# db_utils.py
import time
import logging
from constants import STR_SELECT_SQL, STR_SUCCESS, INT_60

def fetch_users(retries=3):
    # Magic String (SQL)
    query = STR_SELECT_SQL

    if retries > 0:
        # Magic String inside logging (can be ignored via config)
        logging.debug("Attempting connection...")
        # Magic Number
        time.sleep(INT_60)

    return STR_SUCCESS

---

Usage

1. Direct Mode

Scan the codebase and preview or apply changes

Dry Run (Preview changes to stdout)

constantipy

Apply changes to files

constantipy --apply

2. Pipeline Mode

Generate a JSON report, review/edit it manually, and then apply the refactoring.

Generate Report

constantipy report > report.json

Edit report manually (optional)

Validate Report

constantipy validate < report.json

Apply refactoring from report

constantipy refactor --apply < report.json

Configuration

You can fine-tune the extraction process using command-line flags:

Flag	Description	Default
--apply	Apply detected changes to files (direct mode only)	False
--constants-file	File name for storing global constants	constants.py
--exclude	Directories to exclude	None
--extra-constants	Additional Python files to scan for constants	None
--ignore-call	Ignore specific function calls during scanning	None
--min-count	Minimum occurrences for a constant to be considered	2
--min-length	Minimum length for string literals	4
--naming {generic,derived}	Strategy for naming generated constants	derived
--no-local-scope	Force all constants to be global (ignore local scope)	False
--path	Path to scan or refactor (default: current directory)	.

Numeric & Type Controls

Flag	Description
--ignore-num X	Explicitly ignore number X (e.g. --ignore-num 42)
--ignore-str S	Explicitly ignore string S (e.g. --ignore-str "foo")
--include-num X	Explicitly extract number X (overrides default ignore)
--no-bytes	Skip bytes scanning
--no-floats	Skip float scanning
--no-ints	Skip integer scanning
--no-numbers	Disable ALL magic number extraction

Safety Warning ⚠️

Constantipy modifies your source code.

While it contains safety checks (AST parsing, collision detection), automated refactoring always carries risk.

1. Always commit your changes to git before running with --apply.
2. Run your test suite immediately after applying.

License

MIT