Implementation:Explodinggradients Ragas SimplePrompt Class

Field	Value
source	Explodinggradients_Ragas (GitHub)
domains	Prompts, Framework
last_updated	2026-02-10 00:00 GMT

Overview

The Prompt class provides a simple few-shot prompting system that uses Python string formatting with optional examples, response model validation, and JSON-based persistence with optional gzip compression.

Description

Prompt stores an instruction template string with f-string placeholders, an optional list of (input_dict, output_dict) example pairs for few-shot learning, and an optional Pydantic response model. The format() method fills placeholders with provided keyword arguments and appends formatted examples. The add_example() method validates that both input and output are dictionaries before appending. The save() method serializes to JSON (or gzip-compressed JSON with a .gz extension), warning that response_model cannot be serialized. The load() class method restores a prompt from file, requiring the response_model to be provided again if one was originally used. Schema validation ensures the provided model matches the saved schema.

Usage

Create a Prompt with an instruction template and optional examples, then call format() with keyword arguments to produce the final prompt string. Use save()/load() for persistence.

Code Reference

Item	Detail
Source Location	`src/ragas/prompt/simple_prompt.py` L17-303
Class Signature	`class Prompt`
Key Methods	`format(**kwargs)`, `add_example(input, output)`, `save(path)`, `load(path, response_model)`
Import	`from ragas.prompt.simple_prompt import Prompt`

I/O Contract

Inputs

Parameter	Type	Description
`instruction`	`str`	Template string with `{placeholder}` syntax
`examples`	`Optional[List[Tuple[Dict, Dict]]]`	Few-shot (input_dict, output_dict) pairs
`response_model`	`Optional[BaseModel]`	Pydantic model for response validation
`**kwargs` (format)	`Any`	Values to fill template placeholders

Outputs

Method	Return Type	Description
`format()`	`str`	Fully formatted prompt string with examples appended
`save()`	`None`	Writes JSON or gzip-compressed JSON file
`load()`	`Prompt`	Restored Prompt instance from file

Usage Examples

from ragas.prompt.simple_prompt import Prompt

# Basic prompt with placeholders
prompt = Prompt("Answer the question: {question}")
formatted = prompt.format(question="What is 2+2?")
print(formatted)  # "Answer the question: What is 2+2?"

# Prompt with few-shot examples
examples = [
    ({"question": "What is 1+1?"}, {"answer": "2"}),
    ({"question": "What is 3+3?"}, {"answer": "6"}),
]
prompt = Prompt("Answer: {question}", examples=examples)
formatted = prompt.format(question="What is 5+5?")

# Adding examples dynamically
prompt.add_example(
    {"question": "What is 7+7?"},
    {"answer": "14"},
)

# Save and load
prompt.save("my_prompt.json")
loaded = Prompt.load("my_prompt.json")

# With gzip compression
prompt.save("prompt.json.gz")
loaded = Prompt.load("prompt.json.gz")

Related Pages

DynamicFewShotPrompt_Class - Extends Prompt with similarity-based example selection
PydanticPrompt_Class - Structured alternative with Pydantic models
BasePrompt_Class - Abstract base in the prompt hierarchy

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment