Giter Club home page Giter Club logo

oomstore's Introduction

OomStore

Lightweight and Fast Feature Store Powered by Go

CICD Version Version Platform

Quickstart ยท Architecture ยท Benchmark

Overview

oomstore allows you to:

  • Define features with YAML.
  • Store features in databases of choice.
  • Retrieve features for both online serving and offline training, fast.

Please see our docs for more details.

Features

  • ๐Ÿผ Simple. Being serverless and CLI-friendly, users can be productive in hours, not months.
  • ๐Ÿ”Œ Composable. We support your preferred databases of choice.
  • โšก Fast. Benchmark shows oomstore performs QPS > 50k and latency < 0.3 ms with Redis.
  • ๐ŸŒŠ Streaming. We support streaming features to ensure your predictions are up-to-date.

Architecture

Architecture

See Architecture for more details.

Quickstart

  1. Install oomcli following the guide.

  2. oomcli init to initialize oomstore. Make sure there is a ~/.config/oomstore/config.yaml as below.

online-store:
  sqlite:
    db-file: /tmp/oomstore.db

offline-store:
  sqlite:
    db-file: /tmp/oomstore.db

metadata-store:
  sqlite:
    db-file: /tmp/oomstore.db
  1. oomcli apply -f metadata.yaml to register metadata. See metadata.yaml below.
kind: Entity
name: user
description: 'user ID'
groups:
- name: account
  category: batch
  description: 'user account'
  features:
  - name: state
    value-type: string
  - name: credit_score
    value-type: int64
  - name: account_age_days
    value-type: int64
  - name: 2fa_installed
    value-type: bool
- name: txn_stats
  category: batch
  description: 'user txn stats'
  features:
  - name: count_7d
    value-type: int64
  - name: count_30d
    value-type: int64
- name: recent_txn_stats
  category: stream
  snapshot-interval: 24h
  description: 'user recent txn stats'
  features:
  - name: count_10min
    value-type: int64
  1. Import CSV data to Offline Store, then sync from Offline to Online Store
oomcli import \
  --group account \
  --input-file account.csv \
  --description 'sample account data'
oomcli import \
  --group txn_stats \
  --input-file txn_stats.csv \
  --description 'sample txn stats data'
oomcli sync --group-name account --revision-id 2
oomcli sync --group-name txn_stats --revision-id 4
  1. Push stream data to both Online and Offline Store.
oomcli push --group recent_txn_stats --entity-key 1006 --feature count_10min=1
  1. Fetch features by key.
oomcli get online \
  --entity-key 1006 \
  --feature account.state,account.credit_score,account.account_age_days,account.2fa_installed,txn_stats.count_7d,txn_stats.count_30d,recent_txn_stats.count_10min
+------+---------------+----------------------+--------------------------+-----------------------+--------------------+---------------------+------------------------------+
| user | account.state | account.credit_score | account.account_age_days | account.2fa_installed | txn_stats.count_7d | txn_stats.count_30d | recent_txn_stats.count_10min |
+------+---------------+----------------------+--------------------------+-----------------------+--------------------+---------------------+------------------------------+
| 1006 | Louisiana     |                  710 |                       32 | false                 |                  8 |                  22 |                            1 |
+------+---------------+----------------------+--------------------------+-----------------------+--------------------+---------------------+------------------------------+
  1. Generate training datasets via Point-in-Time Join.
oomcli join \
	--feature account.state,account.credit_score,account.account_age_days,account.2fa_installed,txn_stats.count_7d,txn_stats.count_30d,recent_txn_stats.count_10min \
	--input-file label.csv
+------------+---------------+---------------+----------------------+--------------------------+-----------------------+--------------------+---------------------+------------------------------+
| entity_key |  unix_milli   | account.state | account.credit_score | account.account_age_days | account.2fa_installed | txn_stats.count_7d | txn_stats.count_30d | recent_txn_stats.count_10min |
+------------+---------------+---------------+----------------------+--------------------------+-----------------------+--------------------+---------------------+------------------------------+
|       1002 | 1950236233000 | Hawaii        |                  625 |                      861 | true                  |                 11 |                  36 |                              |
|       1003 | 1950411318000 | Arkansas      |                  730 |                      958 | false                 |                  0 |                  16 |                              |
|       1004 | 1950653614000 | Louisiana     |                  610 |                     1570 | false                 |                 12 |                  26 |                              |
|       1005 | 1950166137000 | South Dakota  |                  635 |                     1953 | false                 |                  7 |                  30 |                              |
|       1006 | 1950403162000 | Louisiana     |                  710 |                       32 | false                 |                  8 |                  22 |                            1 |
|       1007 | 1950160030000 | New Mexico    |                  645 |                       37 | true                  |                  5 |                  40 |                              |
|       1008 | 1950274859000 | Nevada        |                  735 |                     1627 | false                 |                 12 |                  51 |                              |
|       1009 | 1949958846000 | Kentucky      |                  650 |                       88 | true                  |                 11 |                  23 |                              |
|       1010 | 1949920686000 | Delaware      |                  680 |                     1687 | false                 |                  2 |                  39 |                              |
+------------+---------------+---------------+----------------------+--------------------------+-----------------------+--------------------+---------------------+------------------------------+

See Quickstart for more complete details.

Supported Databases

Online Store

  • Amazon DynamoDB
  • Redis
  • TiKV
  • Cassandra
  • PostgreSQL
  • MySQL
  • TiDB
  • SQLite

Offline Store

  • Snowflake
  • Amazon Redshift
  • Google BigQuery
  • PostgreSQL
  • MySQL
  • TiDB
  • SQLite

Metadata Store

  • PostgreSQL
  • MySQL
  • TiDB
  • SQLite

Community

Feel free to join the community for questions and feature requests!

Credits

oomstore is highly inspired by Feast.

oomstore's People

Contributors

jinghancc avatar wfxr avatar lianxmfor avatar yiksanchan avatar

Watchers

James Cloos avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.