Skip to main content

LangGraph Checkpoint KurrentDB: Experimental

KurrentDB checkpoint implementation for LangGraph. This is an experimental release and not be ready to be used in production, yet. This project is intended to show that KurrentDB has the right native capabilities to be used as a checkpointing mechanism for LangGraph.

If you would like to see a production version of the checkpointer, feel free to upvote this issue: https://github.com/kurrent-io/langgraph-checkpoint-kurrentdb/issues/3

Overview

LangGraph is an orchestration framework for complex agentic systems. This package provides a checkpointing mechanism for LangGraph using KurrentDB. We use KurrentDB's event sourcing capabilities to store the state of the LangGraph execution. The versioning of state is mapped to KurrentDB's native event versioning and every the channel value is kept into its own stream. Every step in the thread then only has a pointer (channel value name and version) to the actual channel value (event). (Original Langgraph's channel versioning is also kept)

graph TD
  subgraph ThreadStream["Thread's Stream"]
    Step1["Step -1 (event)"]
    Step2["Step 0 (event)"]
    Step3["Step 1... (event)"]
  end

  subgraph ChannelValueStreams["Channel Value Streams"]
    ChannelA["Channel A"]
    ChannelB["Channel B"]
    ChannelC["Channel C"]
  end

  subgraph ChannelValueEvents["Channel A Value Events"]
    Event0["Event V0"]
    Event1["Event V1"]
    Event2["Event V2..."]
  end

  Step1 -->|Points to| ChannelA
  Step1 -->|Points to| ChannelB
  Step2 -->|Points to| ChannelB
  Step2 -->|Points to| ChannelC
  Step3 -->|Points to| ChannelA
  Step3 -->|Points to| ChannelC
  ChannelA -->|Points to| ChannelValueEvents

Installation

pip install langgraph-checkpoint-kurrentdb

Or using Poetry:

poetry add langgraph-checkpoint-kurrentdb

Usage

Setting up the KurrentDB checkpointer for your LangGraph:

from langgraph.graph import StateGraph
from langgraph_checkpoint_kurrentdb import KurrentDBSaver
from kurrentdbclient import KurrentDBClient

# Initialize the KurrentDB client
client = KurrentDBClient(uri="esdb://localhost:2113?Tls=false")

# Create the KurrentDB checkpoint saver
saver = KurrentDBSaver(client)

# Create your LangGraph
builder = StateGraph(int)
builder.add_node("add_one", lambda x: x + 1)
builder.set_entry_point("add_one")
builder.set_finish_point("add_one")

# Compile with the KurrentDB checkpointer
graph = builder.compile(checkpointer=saver)

# Use thread ID to identify this execution
config = {"configurable": {"thread_id": "example-thread-1", "checkpoint_ns": ""}}
result = graph.invoke(42, config)

# Later, retrieve the state
state = graph.get_state(config)

Tests

poetry run pytest --cov --cov-report=term-missing

Metadata

Release files for langgraph_checkpoint_kurrentdb 0.1.3

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

Source distribution (sdist)

Source distribution for langgraph_checkpoint_kurrentdb 0.1.3
File Size Uploaded
langgraph_checkpoint_kurrentdb-0.1.3.tar.gz 9.6 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for langgraph_checkpoint_kurrentdb 0.1.3
File Interpreter ABI Platform
langgraph_checkpoint_kurrentdb-0.1.3-py3-none-any.whl Python 3 none any Details

Total release size: 19.3 kB

Release files / langgraph_checkpoint_kurrentdb-0.1.3.tar.gz

Download URL langgraph_checkpoint_kurrentdb-0.1.3.tar.gz
Size 9.6 kB
Tags Source
SHA-256 checksum
How to use checksums
b437f7a0819df3156bc32ae9726f30a55daa2216139a3e0a6fe193eb78655fbc
BLAKE2b-256 checksum
How to use checksums
8100a8b67b97a088e5d9eb0e9807ab1f3801fb7b5e72b17f439b04f790721e21
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via poetry/2.1.1 CPython/3.12.9 Windows/11

Release files / langgraph_checkpoint_kurrentdb-0.1.3-py3-none-any.whl

Download URL langgraph_checkpoint_kurrentdb-0.1.3-py3-none-any.whl
Size 9.7 kB
Tags Python 3
SHA-256 checksum
How to use checksums
70c1a202ec617c2a48b817e705971d681c8f8cd7e9da7399721d2a841e69c93d
BLAKE2b-256 checksum
How to use checksums
06ee9951a34cfbe1dfefacbd503d3f3d4b0a36566639a1e7f11c2b1c35b0bd3f
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via poetry/2.1.1 CPython/3.12.9 Windows/11

Release history Release notifications | RSS feed

This release

0.1.3 This release

2 release files

0.1.2

2 release files

0.1.1

2 release files

0.1.0

2 release files

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