Files
langgraph/libs/checkpoint-conformance/README.md
T
398d6cc59d chore(langgraph): add guide/conformance for delta channel checkpointer (#7736)
## Summary

Add a user-facing design doc and `get_delta_channel_keepset` helper for
third-party `BaseCheckpointSaver` authors who need to support graphs
using `DeltaChannel`.

**Deliverables:**

1. ~~**`docs/delta-channel-checkpointer-guide.md`** — comprehensive
guide covering~~:
moved to docs repo

2. **`BaseCheckpointSaver.get_delta_channel_keepset` /
`aget_delta_channel_keepset`** — returns the minimum set of ancestor
`checkpoint_id`s that must survive deletion for a given head's
`DeltaChannel` reconstruction to remain intact. Enables safe `prune`
implementations without silently corrupting delta history.

3. **Docstring warnings** on `prune`, `aprune`, `delete_for_runs`,
`adelete_for_runs`, `copy_thread`, `acopy_thread` explaining the
DeltaChannel pitfall (silent data loss if ancestor writes/snapshots are
deleted).

4. **Three new conformance capabilities** in
`libs/checkpoint-conformance`:
- `delta_channel_history` — validates the `aget_delta_channel_history`
walk contract
   - `delta_channel_keepset` — validates the keep-set contract
- `delta_channel_reconstruction` — end-to-end round-trip (aput +
aput_writes + history + reconstruct)

## Test plan

- [x] `make format lint` passes in `libs/checkpoint`,
`libs/checkpoint-conformance`
- [x] All three new conformance capabilities pass against
`InMemorySaver`
- [x] Run conformance against SQLite saver 
- [x] Run conformance against Postgres saver

---------

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-05-07 17:34:46 -07:00

3.1 KiB

langgraph-checkpoint-conformance

Conformance test suite for LangGraph checkpointer implementations.

Validates that a BaseCheckpointSaver subclass correctly implements the checkpoint storage contract — blob round-trips, metadata preservation, namespace isolation, incremental channel updates, and more.

Installation

pip install langgraph-checkpoint-conformance

Quick start

Register your checkpointer with @checkpointer_test and run validate():

import asyncio
from langgraph.checkpoint.conformance import checkpointer_test, validate

@checkpointer_test(name="MyCheckpointer")
async def my_checkpointer():
    saver = MyCheckpointer(...)
    yield saver
    # cleanup runs after yield

async def main():
    report = await validate(my_checkpointer)
    report.print_report()
    assert report.passed_all_base()

asyncio.run(main())

Or in a pytest test:

import pytest
from langgraph.checkpoint.conformance import checkpointer_test, validate

@checkpointer_test(name="MyCheckpointer")
async def my_checkpointer():
    yield MyCheckpointer(...)

@pytest.mark.asyncio
async def test_conformance():
    report = await validate(my_checkpointer)
    report.print_report()
    assert report.passed_all_base()

Capabilities

The suite tests base capabilities (required) and extended capabilities (optional, auto-detected):

Capability Required Method
put yes aput
put_writes yes aput_writes
get_tuple yes aget_tuple
list yes alist
delete_thread yes adelete_thread
delete_for_runs no adelete_for_runs
copy_thread no acopy_thread
prune no aprune
delta_channel_history no aget_delta_channel_history
delta_channel_keepset no aget_delta_channel_keepset
delta_channel_reconstruction no aput

Extended capabilities are detected by checking whether the method is overridden from BaseCheckpointSaver. If not overridden, those tests are skipped.

Options

Progress output

from langgraph.checkpoint.conformance.report import ProgressCallbacks

# Dot-style progress (. per pass, F per fail)
report = await validate(my_checkpointer, progress=ProgressCallbacks.default())

# Verbose (per-test names + stacktraces on failure)
report = await validate(my_checkpointer, progress=ProgressCallbacks.verbose())

Skip capabilities

@checkpointer_test(name="MyCheckpointer", skip_capabilities={"prune"})
async def my_checkpointer():
    yield MyCheckpointer(...)

Run specific capabilities

report = await validate(my_checkpointer, capabilities={"put", "list"})

Lifespan (one-time setup/teardown)

For expensive setup like database creation:

async def db_lifespan():
    await create_database()
    yield
    await drop_database()

@checkpointer_test(name="PostgresSaver", lifespan=db_lifespan)
async def pg_checkpointer():
    async with PostgresSaver.from_conn_string(CONN_STRING) as saver:
        yield saver