Ryoma

AI Powered Data Agent framework, a comprehensive solution for data analysis, engineering, and visualization. This version is compatible with pydandic v2.

Tech Stack

Our platform leverages a combination of cutting-edge technologies and frameworks:

• Langchain: Facilitates the seamless integration of language models into application workflows, significantly enhancing AI interaction capabilities.
• Reflex: An open-source framework for quickly building beautiful, interactive web applications in pure Python
• Apache Arrow: A cross-language development platform for in-memory data that specifies a standardized language-independent columnar memory format for flat and hierarchical data, organized for efficient analytic operations on modern hardware like CPUs and GPUs.
• Jupyter Ai Magics: A JupyterLab extension that provides a set of magics for working with AI models.
• Amundsen: A data discovery and metadata platform that helps users discover, understand, and trust the data they use.
• Ibis: A Python data analysis framework that provides a pandas-like API for analytics on large datasets.
• Feast: An operational feature store for managing and serving machine learning features to models in production.

Installation

Simply install the package using pip:

pip install ryoma_ai

Or with extra dependencies:

pip install ryoma_ai[snowflake]

Basic Example

Below is an example of using SqlAgent to connect to a postgres database and ask a question. You can read more details in the documentation.

from ryoma_ai.agent.sql import SqlAgent
from ryoma_ai.datasource.postgres import PostgresDataSource

# Connect to a postgres catalog
datasource = PostgresDataSource("postgresql://user:password@localhost:5432/dbname")

# Create a SQL agent
sql_agent = SqlAgent("gpt-3.5-turbo").add_datasource(datasource)

# ask question to the agent
sql_agent.stream("I want to get the top 5 customers which making the most purchases", display=True)

The Sql agent will try to run the tool as shown below:

================================ Human Message =================================

I want to get the top 5 customers which making the most purchases
================================== Ai Message ==================================
Tool Calls:
  sql_database_query (call_mWCPB3GQGOTLYsvp21DGlpOb)
 Call ID: call_mWCPB3GQGOTLYsvp21DGlpOb
  Args:
    query: SELECT C.C_NAME, SUM(L.L_EXTENDEDPRICE) AS TOTAL_PURCHASES FROM CUSTOMER C JOIN ORDERS O ON C.C_CUSTKEY = O.O_CUSTKEY JOIN LINEITEM L ON O.O_ORDERKEY = L.L_ORDERKEY GROUP BY C.C_NAME ORDER BY TOTAL_PURCHASES DESC LIMIT 5
    result_format: pandas

Continue to run the tool with the following code:

sql_agent.stream(tool_mode=ToolMode.ONCE)

Output will look like after running the tool:

================================== Ai Message ==================================

The top 5 customers who have made the most purchases are as follows:

1. Customer#000143500 - Total Purchases: $7,154,828.98
2. Customer#000095257 - Total Purchases: $6,645,071.02
3. Customer#000087115 - Total Purchases: $6,528,332.52
4. Customer#000134380 - Total Purchases: $6,405,556.97
5. Customer#000103834 - Total Purchases: $6,397,480.12

Use Ryoma Lab

Ryoma lab is an application that allows you to interact with your data and AI models in UI. The ryoma lab is built with Reflex.

1. Create Ryoma lab configuration file rxconfig.py in your project:

import logging

import reflex as rx
from reflex.constants import LogLevel

config = rx.Config(
    app_name="ryoma_lab",
    loglevel=LogLevel.INFO,
)

# Setup basic configuration for logging
logging.basicConfig(level=logging.INFO, format="%(asctime)s - %(levelname)s - %(message)s")

2. You can start the ryoma lab by running the following command:

ryoma_lab run

the ryoma lab will be available at http://localhost:3000.

Supported Models

Model provider are supported by jupyter ai magics. Ensure the corresponding environment variables are set before using the Ryoma agent.

Supported Data Sources

• Snowflake
• Sqlite
• BigQuery
• Postgres
• MySQL
• File (CSV, Excel, Parquet, etc.)
• Redshift
• DynamoDB

Supported Engines

• Apache Spark
• Apache Flink
• Presto

🛡 License

This project is licensed under the terms of the Apache Software License 2.0 license. See LICENSE for more details.