cmem-client

Next generation eccenca Corporate Memory client library.

cmem-client is the Python client library for eccenca Corporate Memory. Corporate Memory exposes its functionality through several services - DataIntegration (build), DataPlatform (explore), Keycloak and the Marketplace. This library hides that split behind a single typed and validated Python object: the client.

Features

• Validated responses. Every response is parsed into a Pydantic model, so a typo in a server field becomes an error instead of a KeyError three frames later.
• Proper typing. The whole package is type annotated and checked with mypy, so your editor can complete and check your code.
• Dict-like collections. Repositories such as client.graphs behave like read-only dictionaries: len(), in, [] and iteration all work, and data is fetched lazily on first access.
• Structured logging through the standard library, configured with one call on the client, including a TRACE level below DEBUG which logs the arguments and the result of the methods carrying the decorator.
• Built on httpx, which would also allow asynchronous calls. The exposed client API is synchronous.

Installation

pip install cmem-client

The package requires Python 3.13 or newer.

Getting started

Point the client at your Corporate Memory deployment and give it credentials:

export CMEM_BASE_URI="https://mycmem.example.com/"
export OAUTH_GRANT_TYPE="client_credentials"
export OAUTH_CLIENT_ID="cmem-service-account"
export OAUTH_CLIENT_SECRET="my-secret"

Then create a client and ask the deployment how it is doing:

from cmem_client.client import Client

client = Client.from_env()

print(client.store.self_information.type)      # GRAPHDB
print(client.deployment.get_status().health)   # UP
print(len(client.graphs))                      # 6

That is the whole setup. Besides Client.from_env() there are Client.from_dict() for settings you hold yourself, Client.from_context() for use inside a Corporate Memory Python plugin, and Client.from_cmempy() to reuse an already configured cmempy environment.

What you can reach

Everything hangs off the client:

client.graphs           # RDF graphs in the store, a read-only mapping
client.projects         # DataIntegration projects
client.store.sparql     # SPARQL query and update
client.marketplace      # Marketplace packages

Further repositories cover datasets, workflows, queries, files, variables, schedulers, vocabularies, graph imports and insights, validations, access conditions, user and client accounts, Python packages and installed marketplace packages. The documentation lists them all.

Coming from cmem-cmempy

cmem-client succeeds cmem-cmempy, but it is not a drop-in replacement. Where cmempy offers module-level functions returning parsed JSON, this library gives you a client object whose responses are validated models:

# cmem-cmempy
from cmem.cmempy.dp.proxy.graph import get_graphs_list
graphs = get_graphs_list()                # list of dicts

# cmem-client
graphs = list(client.graphs.values())     # list of Graph models

If a cmempy environment is already configured in your process, reuse it directly with Client.from_cmempy().

Documentation

The full API documentation, including the complete configuration reference and the migration notes, lives at documentation.eccenca.com.