Implementation:Explodinggradients Ragas InMemoryBackend Class

Knowledge Sources	Explodinggradients_Ragas
Domains	Storage, Backend
Last Updated	2026-02-10 00:00 GMT

Overview

Non-persistent in-memory backend for temporary storage of datasets and experiments with deep-copy isolation.

Description

InMemoryBackend implements BaseBackend to store datasets and experiments in Python dictionaries. Data is deep-copied on save and load to ensure instance isolation. This backend is suitable for testing, prototyping, and transient workflows where persistence is not required.

Usage

Use this backend when you need temporary storage during a single session, such as testing metrics or running experiments without disk I/O.

Code Reference

Source Location

• Repository: Explodinggradients_Ragas
• File: src/ragas/backends/inmemory.py
• Lines: 11-131

Signature

class InMemoryBackend(BaseBackend):
    def __init__(self) -> None:
        ...
    def load_dataset(self, name: str) -> List[Dict[str, Any]]:
        ...
    def load_experiment(self, name: str) -> List[Dict[str, Any]]:
        ...
    def save_dataset(self, name: str, data: List[Dict[str, Any]], data_model: Optional[Type[BaseModel]] = None) -> None:
        ...
    def save_experiment(self, name: str, data: List[Dict[str, Any]], data_model: Optional[Type[BaseModel]] = None) -> None:
        ...
    def list_datasets(self) -> List[str]:
        ...
    def list_experiments(self) -> List[str]:
        ...

Import

from ragas.backends.inmemory import InMemoryBackend

I/O Contract

Inputs

Name	Type	Required	Description
name	str	Yes	Name identifier for the dataset or experiment
data	List[Dict[str, Any]]	Yes	List of records to save (deep-copied)
data_model	Optional[Type[BaseModel]]	No	Pydantic model for validation context

Outputs

Name	Type	Description
load returns	List[Dict[str, Any]]	Deep copy of stored records
list returns	List[str]	Sorted list of stored names

Usage Examples

from ragas.backends.inmemory import InMemoryBackend

backend = InMemoryBackend()

# Save a dataset
backend.save_dataset("my_dataset", [{"question": "What is AI?", "answer": "Artificial Intelligence"}])

# Load it back
data = backend.load_dataset("my_dataset")
print(data)  # [{"question": "What is AI?", "answer": "Artificial Intelligence"}]

# List datasets
print(backend.list_datasets())  # ["my_dataset"]