personal-timeline  by facebookresearch

Digital timeline builder for personal data

created 2 years ago
369 stars

Top 77.7% on sourcepulse

GitHubView on GitHub
1 Expert Loves This Project
Project Summary

TimelineBuilder is a system for constructing and querying personal digital data timelines, targeting researchers and individuals interested in organizing and analyzing their digital footprint. It offers a unified interface to ingest data from various sources, visualize it, and perform natural language question answering.

How It Works

The system ingests data from multiple sources (e.g., Apple Health, Google Photos, Spotify) into a structured SQLite database. A ReactJS frontend provides visualization capabilities, while a retrieval-augmented language model (PostText) powers a question-answering engine. The QA engine supports ChatGPT for general knowledge, retrieval-based QA over personal data, and view-based QA for aggregate queries via SQL translation.

Quick Start & Requirements

  • Install: Clone the repository and run sh src/init.sh.
  • Prerequisites: Docker Desktop, git-lfs, Python 3.10.
  • API Keys: Google Maps API, Spotify API, OpenAI API are required for full functionality.
  • Data: Users must download and place their personal data into specific directories within ~/personal-data.
  • Setup: End-to-end pipeline can be started with docker-compose up -d --build.
  • Links: TimelineQA Benchmark

Highlighted Details

  • Supports ingestion from 9+ digital services including Apple Health, Google Photos/Location, Spotify, and Facebook.
  • Features a retrieval-based and view-based QA engine for querying personal data using natural language.
  • Includes TimelineQA, a synthetic benchmark for evaluating personal timeline QA systems.
  • Offers a ReactJS-based frontend for data visualization.

Maintenance & Community

The project is from Meta AI (facebookresearch). Key contributors include Tripti Singh (DB backend, importer, orchestrator) and Wang-Chiew Tan (PostText query engine).

Licensing & Compatibility

Licensed under the Apache 2.0 license, permitting commercial use and integration with closed-source projects.

Limitations & Caveats

Requires significant user effort to download and organize personal data from various services. API key setup is necessary for visualization and QA features. The system relies on external APIs (OpenAI, Google Maps, Spotify) which may incur costs or have usage limits.

Health Check
Last commit

11 months ago

Responsiveness

Inactive

Pull Requests (30d)
0
Issues (30d)
0
Star History
11 stars in the last 90 days

Explore Similar Projects

Starred by Chip Huyen Chip Huyen(Author of AI Engineering, Designing Machine Learning Systems) and Elie Bursztein Elie Bursztein(Cybersecurity Lead at Google DeepMind).

LightRAG by HKUDS

1.0%
19k
RAG framework for fast, simple retrieval-augmented generation
created 10 months ago
updated 20 hours ago
Feedback? Help us improve.