Skip to main content

sageFlow

sageFlow is a cutting-edge, vector-native stream processing engine designed specifically to maintain and materialize semantic state snapshots for real-time, LLM-based generation tasks. The engine offers a declarative API to compose stateful vector operations within temporal windows, enabling fast and efficient updates to semantic context for dynamically changing datasets.

Features

  • Vector-Native Stream Processing: At its core, sageFlow is built to handle high-dimensional vector streams efficiently.
  • Declarative API: Easily compose complex, stateful vector operations such as TopK, Filter, and Join within defined temporal windows.
  • Incremental Low-Latency Updates: Optimized for incremental computations, ensuring semantic states are updated with minimal delay.
  • Optimized Three-Phase Pipeline: Abstracts stream processing into three distinct phases—ingestion, state materialization, and snapshot exposure—unlocking significant optimization opportunities.
  • Stateful and Windowed Operations: Natively supports windowing to create time-bound semantic snapshots from continuous data streams.

Key Use Cases

  • Real-time LLM Generation: Provide large language models with fresh, stateful context snapshots for more accurate and timely responses.
  • Dynamic Context Maintenance: Ideal for conversational AI or interactive applications where the context evolves rapidly over time.
  • Streaming Data Analytics: Serve high-velocity data analysis use cases that require complex, stateful semantic queries on vector data.
  • Adaptive Recommendation Systems: Build systems that can update recommendations in real-time based on the most recent user interactions and streaming events.

Setup

To setup sageFlow and it's dependencies, begin by making sure that you have docker installed, or any Linux release version that contains apt, such as Ubuntu or Debian

We suggest first begin with docker before you are familiar with sageFlow.

Quick Installation (Ubuntu/Debian)

For a quick one-click installation of all dependencies including DiskANN support, run:

cd <PATH_TO_REPO>
sudo ./scripts/install-deps.sh

This script will install:

  • Build essentials (gcc, g++, cmake, etc.)
  • DiskANN dependencies (libaio, boost, etc.)
  • Intel MKL (Math Kernel Library)
  • Environment configuration

After installation, reload your environment:

source /etc/profile.d/mkl.sh

Docker

make sure you have installed Docker and Docker is running

Windows

cd <PATH_TO_REPO>/setup
./start_win.bat

Linux

cd <PATH_TO_REPO>/setup
./start.sh

Linux with apt

check the dependencies in <PATH_TO_REPO>/setup/Dockerfile, and build your env

sageFlow: examples

run the following commands to generate examples

cmake -B build
cmake --build build -j $(nproc)

This will generate the examples in the build/bin directory. You can run the examples with:

./build/bin/itopk

Metadata

Release files for isage-flow 0.1.3.1

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Built distribution (wheel)

Table of built distributions (wheels) for isage-flow 0.1.3.1
File Interpreter ABI Platform Actions
isage_flow-0.1.3.1-cp311-cp311-manylinux_2_34_x86_64.whl CPython 3.11 CPython 3.11 Linux glibc 2.34+ x86-64 Details for isage_flow-0.1.3.1-cp311-cp311-manylinux_2_34_x86_64.whl

Release files / isage_flow-0.1.3.1-cp311-cp311-manylinux_2_34_x86_64.whl

Download URL isage_flow-0.1.3.1-cp311-cp311-manylinux_2_34_x86_64.whl
Size 1.1 MB
Tags CPython 3.11 Linux glibc 2.34+ x86-64
SHA-256 checksum
How to use checksums
3ba97440acd1d79dafc9e5abdaefcdcbbd545c110534cfe97f42e71ca27fe82e
BLAKE2b-256 checksum
How to use checksums
898fc9bdb5226248c7d9511525ca742e4037bb3e66460065564553b03f9d11ef
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.2.0 CPython/3.11.11

Security for isage-flow 0.1.3.1

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page