Google Knowledge Graph API in Python: Legacy Parsing and Migration

For an existing Google Knowledge Graph Search API integration, parse the legacy JSON-LD response explicitly. For a new project, assess Google’s current Enterprise Knowledge Graph guidance first. The legacy endpoint and the Cloud migration are different integration contracts; changing the URL alone does not migrate the client.

Identify the API before copying a Python example

Google’s legacy API overview directs new users toward Enterprise Knowledge Graph and cautions against a production-critical dependency on the legacy service. The legacy response contains individual matching entities; it is not an export of an interconnected knowledge graph.

The legacy request uses an API key. A Cloud migration has different authentication, identifiers and response details. Google’s migration documentation removes resultScore from the new contract. Keep the two parsers and credential configurations distinct.

Parse an owned legacy-response fixture

def legacy_entities(payload):
    rows = payload.get('itemListElement', [])
    if not isinstance(rows, list):
        raise ValueError('Expected itemListElement array')
    entities = []
    for item in rows:
        if not isinstance(item, dict) or not isinstance(item.get('result'), dict):
            continue
        result = item['result']
        entities.append({'id': result.get('@id'), 'name': result.get('name'),
                         'types': result.get('@type', []),
                         'score': item.get('resultScore')})
    return entities

fixture = {"itemListElement": [{"result": {"@id": "kg:/m/example",
    "name": "Fixture Entity", "@type": ["Organization"]}, "resultScore": 7.0}]}
assert legacy_entities(fixture)[0]["name"] == "Fixture Entity"
assert legacy_entities({}) == []

This example was tested with synthetic valid, empty and missing-field cases. The score is a legacy matching field, not a calibrated probability, an SEO ranking or a promise of a Knowledge Panel.

Make the migration decision explicit

Entity lookup is separate from a Google SERP request. Use the Google Search API tutorial when the task is to retrieve web result positions and URLs.

• Inventory callers that depend on the old identifier, resultScore or API-key flow.
• Read the current Cloud edition, authentication and response documentation before implementing a replacement.
• Retain fixtures for empty results and missing optional fields.
• Check actual project quota and service terms; do not rely on an unverified quota copied from a blog.

Reference documentation: developers.google.com — /knowledge-graph; developers.google.com — /knowledge-graph/reference/rest/v1; docs.cloud.google.com — /enterprise-knowledge-graph/docs/migrate.

Related implementation and validation guides

• Google Search API in Python: Make and Verify a JSON Request