🛡️ SafeMathJax

Safe, production-ready mathematical operations for Python, JAX, and Pandas. Never crash on edge cases—get reliable results every time.

---

🚀 Quick Start

pip install SafeMathJax

import safemathjax as sm
import jax.numpy as jnp
import pandas as pd

# Never crashes, always returns a fallback
sm.safe_divide(10, 0)                # nan
sm.safe_log(-1, fallback=0)          # 0
sm.safe_sqrt([-1, 4, 9])             # [nan, 2.0, 3.0]

---

📈 Use Cases

SafeMath is designed for any scenario where mathematical reliability and safety are critical. Here are some ideal use cases:

• Data Science & Analytics:

  • Prevents crashes in data pipelines due to NaNs, infinities, or invalid operations.
  • Seamlessly handles missing, zero, or out-of-domain values in JAX and Pandas workflows.

• Financial & Business Applications:

  • Ensures robust calculations for KPIs, ratios, and financial metrics, even with incomplete or edge-case data.
  • Avoids propagation of errors in chained calculations (e.g., division by zero, log of negative numbers).

• Machine Learning Pipelines:

  • Guarantees safe preprocessing and feature engineering, so models never break on bad input.
  • Makes batch transformations and aggregations safe for production.

• APIs & Web Services:

  • Protects backend services from runtime math errors, returning fallback values instead of 500 errors.
  • Enables safe mathematical evaluation of user-provided expressions.

• Scientific Computing:

  • Handles edge cases in simulations, experiments, and research code without manual error checks.

• General Python Development:

  • Drop-in replacement for standard math operations to make any codebase more robust and production-ready.

---

✨ Features

• 🛡️ Never Crashes: Handles all edge cases with configurable fallbacks
• 📊 Universal Compatibility: Works with scalars, JAX arrays, Pandas Series/DataFrames
• ⚡ High Performance: Vectorized operations using JAX XLA compilation
• 🔧 Flexible Fallbacks: Global and per-function fallback strategies
• 🔗 Method Chaining: Fluent API with SafeNumber wrapper
• 📝 Expression Evaluation: Safe evaluation of mathematical expressions
• 🧪 Comprehensive Logging: Optional tracing for debugging and auditing

---

📚 Function Reference

Arithmetic Operations

sm.safe_add(a, b)
sm.safe_subtract(a, b)
sm.safe_multiply(a, b)
sm.safe_divide(a, b)
sm.safe_mod(a, b)
sm.safe_power(a, b)
sm.safe_abs(x)
sm.safe_negate(x)

Mathematical Functions

sm.safe_log(x)
sm.safe_log10(x)
sm.safe_sqrt(x)
sm.safe_sin(x)
sm.safe_cos(x)
sm.safe_tan(x)

---

💡 Usage Examples

Scalars & Arrays

sm.safe_divide(10, 0)                # nan (no ZeroDivisionError)
sm.safe_log(-5, fallback=0)          # 0 (no ValueError)
sm.safe_sqrt(jnp.array([1, -1, 0]))   # [1.0, nan, 0.0]

Pandas Integration

import pandas as pd

df = pd.DataFrame({
    'revenue': [100, 200, 0],
    'costs': [50, 250, 0],
    'temperature': [25, -10, 0]
})

# Safe operations on columns
df['profit_margin'] = sm.safe_divide(df['revenue'] - df['costs'], df['revenue'])
df['log_temp'] = sm.safe_log(df['temperature'] + 273.15)

# In-place operation
sm.safe_sqrt(df[['revenue', 'costs']], inplace=True)

Global Fallback Configuration

sm.set_global_fallback(0)
sm.safe_log(-1)  # Returns 0 instead of nan

sm.enable_trace()
sm.safe_divide(10, 0)  # Logs fallback usage

Method Chaining with SafeNumber

from safemathjax import SafeNumber

data = [16, -4, 0, 25]
result = SafeNumber(data).abs().sqrt().log().value_or(0)
# Equivalent to: log(sqrt(abs(data))) with fallback 0

Safe Expression Evaluation

sm.safe_eval("log(x) + sqrt(y)", {"x": 10, "y": 25})
sm.safe_eval("log(-1) + tan(pi/2)", fallback=0)

---

🧪 Error Handling Examples

All of these return sensible fallbacks instead of crashing:

sm.safe_divide(1, 0)         # nan
sm.safe_log(-1)              # nan
sm.safe_sqrt([-1, 4])        # [nan, 2.0]
sm.safe_tan(jnp.pi/2)         # finite value
sm.safe_power(10, 1000)      # handles overflow
sm.safe_add(None, 5)         # nan
sm.safe_log(0, fallback=-jnp.inf)  # -inf

---

🏗️ Advanced Usage

CLI

safemathjax "log(10) + sqrt(25)" --fallback=0
safemathjax "x^2 + y" --variables="x=3" --variables="y=4"

Integration

import jax.numpy as jnp
import safemathjax as sm

# Instead of:
result = jnp.log(data) / jnp.sqrt(data - 1)  # Can crash

# Use:
result = sm.safe_divide(sm.safe_log(data), sm.safe_sqrt(data - 1))

---

📊 Supported Data Types

Input Type	Example	Output Type
Scalar	5, 3.14, 1+2j	Same type
List/Tuple	[1, 2, 3]	JAX array
JAX Array	jnp.array([1, 2, 3])	JAX array
Pandas Series	pd.Series([1, 2, 3])	Pandas Series
Pandas DataFrame	Numeric columns only	Pandas DataFrame

---

🔧 Configuration

sm.set_global_fallback(0)      # Use 0 for errors
sm.set_global_fallback(None)   # Use None for errors
sm.enable_trace()              # Enable logging
sm.disable_trace()             # Disable logging
current_fallback = sm.get_global_fallback()

---

🧪 Testing

python -m pytest tests/
python -m pytest tests/ --cov=safemathjax
python -m pytest tests/test_core.py
python -m pytest tests/test_pandas.py

---

📋 Requirements

• Python >= 3.7
• JAX >= 0.4.0
• Pandas >= 1.3.0

---

📄 License

MIT License - see LICENSE for details.

---

🤝 Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

---

🔗 Links

• PyPI Package
• GitHub Repository
• Documentation

---

SafeMathJax – Making mathematical operations safe and reliable for production Python applications.