stratum-client 0.1.0

Official Python client for the Stratum document-chunking API.

stratum-client (Python)

Official Python client for the Stratum document-chunking API.

pip install stratum-client

Quick start

from stratum_client import Client, JobConfig

client = Client(api_key="stratum_xxxxxxxx_...")  # or set STRATUM_API_KEY

job = client.submit_job(
    "report.pdf",
    config=JobConfig(target_size=500, enrichments=["summary"]),
)
result = client.wait_for_job(job.id)

for chunk in result["chunks"]:
    print(chunk["text"])

Client

Client(
    api_key: str | None = None,           # falls back to $STRATUM_API_KEY
    base_url: str = "https://api-stratum.pleias.dev",
    timeout: float = 60.0,
)

Methods:

Method	Description
healthz() / readyz()	Liveness / readiness probes
submit_job(file, *, config=None, project_id=None)	Upload a document and create a job
get_job(job_id)	Fetch a single job
list_jobs(*, page=1, page_size=20, status=None, project_id=None)	Paginated job list
iter_jobs(*, page_size=50, status=None, project_id=None)	Iterate every job, auto-paging
get_job_result(job_id)	Download the chunker output (dict matching docs/output-format.md)
cancel_job(job_id)	Delete a job and its storage artifacts
wait_for_job(job_id, *, poll_interval=2.0, timeout=None)	Poll until done; returns the result

Client is also a context manager: with Client(...) as c: ....

Errors

All errors derive from StratumError:

• AuthenticationError (401/403)
• NotFoundError (404)
• JobNotCompleteError (409 — result requested before job finished)
• ApiError (other non-2xx)
• JobFailedError (raised by wait_for_job when the job ends in failed)