prompt-template 1.0.0

A simple but robust zero-dependency library for creating prompt templates

Prompt Template

This is a lightweight zero-dependency Python library for managing LLM prompt template. Its modelled on the stdlib string.Template but with more robust features.

Features

- Template validation
- Support for nested braces and JSON structures
- Automatic value serialization
- Incremental template population
- Clear error messages with detailed context
- Type hints for better IDE support
- Extensible design

Installation

Install using pip:

```bash
pip install prompt-template
```

Usage

Basic Usage

```python
from prompt_template import PromptTemplate

# Create a template
template = PromptTemplate("Hello ${name}! Welcome to ${location}.")

# Render the template
result = template.to_string(name="Alice", location="Wonderland")
print(result)  # Hello Alice! Welcome to Wonderland.
```

Working with JSON Templates

```python
template = PromptTemplate("""
{
    "user": "${username}",
    "settings": {
        "theme": "${theme}",
        "notifications": ${notifications}
    }
}
""")

# Values are automatically serialized
result = template.to_string(
    username="john_doe",
    theme="dark",
    notifications={"email": True, "push": False}
)
```

Incremental Template Population

You can populate template values incrementally using the substitute method:

```python
# Start with a base template
base = PromptTemplate("The ${animal} jumped over the ${obstacle} in ${location}.")

# Partially populate values
partial = base.substitute(animal="fox", obstacle="fence")

# Complete the template later
final = partial.to_string(location="garden")
print(final)  # The fox jumped over the fence in garden.
```

Validation Features

The library includes built-in validation to catch common issues:

```python
# Missing variables raise MissingTemplateValuesError
template = PromptTemplate("Hello ${name}!")
try:
    template.to_string()  # Raises MissingTemplateValuesError
except MissingTemplateValuesError as e:
    print(f"Missing values: {e.missing_values}")

# Invalid keys raise InvalidTemplateKeysError
try:
    template.to_string(name="World", invalid_key="value")  # Raises InvalidTemplateKeysError
except InvalidTemplateKeysError as e:
    print(f"Invalid keys: {e.invalid_keys}")
```

Automatic Value Serialization

The library handles various Python types automatically:

```python
from uuid import UUID
from decimal import Decimal

template = PromptTemplate("ID: ${id}, Amount: ${amount}, Data: ${data}")
result = template.to_string(
    id=UUID("550e8400-e29b-41d4-a716-446655440000"),
    amount=Decimal("45.67"),
    data={"key": "value"}
)
```

Custom Serialization

You can customize how values are serialized by subclassing PromptTemplate:

```python
from datetime import datetime
import json

class CustomTemplate(PromptTemplate):
    @staticmethod
    def serializer(value: Any) -> str:
        if isinstance(value, datetime):
            return value.isoformat()
        return json.dumps(value, default=str)

# Use custom serialization
template = CustomTemplate("Time: ${current_time}, Data: ${data}")
result = template.to_string(
    current_time=datetime.now(),
    data={"complex": "object"}
)
```

Error Handling

The library provides specific exception classes for different error cases:

• TemplateError: Base exception for all template-related errors
• InvalidTemplateKeysError: Raised when invalid keys are provided
• MissingTemplateValuesError: Raised when required template values are missing
• TemplateSerializationError: Raised when value serialization fails

License

MIT License