Skip to content
SkeinRank SkeinRank

Python SDK

The Python SDK is the fastest way to see SkeinRank canonicalize messy company language. It runs locally and does not require Docker, Elasticsearch, OpenRouter, the Governance API, or a custom dictionary file for the first demo.

Install

Section titled “Install”
Terminal window
pip install skeinrank

From a source checkout:

Terminal window
cd packages/skeinrank-core
poetry install

Try the built-in demo dictionary

Section titled “Try the built-in demo dictionary”
import skeinrank


print(skeinrank.canonicalize("k8s pg timeout"))
# kubernetes postgresql timeout


print(skeinrank.extract("sev1 on kube after deploy"))
# ["critical incident", "kubernetes", "deployment"]

The built-in platform_ops_demo dictionary covers a compact set of platform operations, incident, CI/CD, search, RAG, and SkeinRank terms. It is designed for the first aha moment, not as a production vocabulary.

Use a small dictionary in code

Section titled “Use a small dictionary in code”
from skeinrank import SkeinRank


sr = SkeinRank({
    "kubernetes": ["k8s", "kube"],
    "postgresql": ["pg", "psql", "postgres"],
})


print(sr.canonicalize("kube timeout on pg"))
# kubernetes timeout on postgresql


print(sr.extract("kube timeout on pg"))
# ["kubernetes", "postgresql"]

Export the demo dictionary

Section titled “Export the demo dictionary”

Use the CLI when you want to inspect or adapt the built-in dictionary:

Terminal window
skeinrank demo-dictionary --compact
skeinrank demo-dictionary --output platform_ops_demo.dictionary.json

Load a dictionary file

Section titled “Load a dictionary file”
from skeinrank import SkeinRank


sr = SkeinRank.from_file("company_dictionary.json")
print(sr.canonicalize("k8s rollout uses pg database"))

The dictionary-first functions remain available when you want the lower-level API:

from skeinrank import load_dictionary, extract_terms, canonicalize_text


dictionary = load_dictionary("company_dictionary.json")


result = extract_terms("k8s rollout uses pg database", dictionary=dictionary)
print(result.canonical_values)


canonicalized = canonicalize_text("k8s rollout uses pg database", dictionary=dictionary)
print(canonicalized.text)

Extract from documents

Section titled “Extract from documents”
from skeinrank import load_dictionary, extract_terms_from_document


dictionary = load_dictionary("company_dictionary.json")
result = extract_terms_from_document("incident-runbook.md", dictionary=dictionary)


print(result.document.file_name)
print(result.extraction.canonical_values)

Public objects

Section titled “Public objects”

The stable local SDK surface includes:

  • SkeinRank(...)
  • SkeinRank.from_file(...)
  • canonicalize(...)
  • extract(...)
  • demo_dictionary(...)
  • demo_dictionary_payload(...)
  • Dictionary, DictionaryTerm, DictionaryAlias, DictionaryStopListEntry
  • load_dictionary(...)
  • validate_dictionary(...)
  • extract_terms(...)
  • canonicalize_text(...)
  • load_document_text(...)
  • extract_terms_from_document(...)
  • ExtractionResult, TermMatch, CanonicalizedText

Design direction

The SDK stays deterministic, local, and easy to test. The platform can govern terminology changes, but first-run canonicalization should not require a live database, model call, or search backend.