evalops-google-generativeai 0.2.0

Drop-in Google Generative AI SDK shim for EvalOps llm-gateway routing

evalops-google-generativeai

Drop-in google-generativeai shim for stamping EvalOps organization scope, principal attribution, trace IDs, and provider references onto Google Generative AI calls.

This package is part of the G1 shim wedge from evalops/platform#1133. It ships the Python package and conformance surface while llm-gateway grows Google-native endpoint coverage.

Python

pip install evalops-google-generativeai

import evalops_google_generativeai as genai

genai.configure(
    organization_id="org_123",
    principal="user:[e-mail]",
)

model = genai.GenerativeModel("gemini-2.0-flash")
response = model.generate_content("hello")

Environment

• EVALOPS_API_KEY or GOOGLE_API_KEY: Platform-issued bearer token.
• EVALOPS_ORGANIZATION_ID: organization scope stamped into requests.
• EVALOPS_PRINCIPAL: optional actor string for audit attribution.
• EVALOPS_TRACE_ID: optional trace correlation ID.
• EVALOPS_LLM_GATEWAY_GOOGLE_URL: Google-compatible gateway base URL.
• EVALOPS_LLM_GATEWAY_URL: fallback gateway base URL.
• EVALOPS_PROVIDER_ENVIRONMENT: defaults to prod.
• EVALOPS_PROVIDER_CREDENTIAL_NAME: optional provider ref credential name.
• EVALOPS_PROVIDER_TEAM_ID: optional provider ref team ID.

Gateway Status

llm-gateway currently exposes OpenAI-compatible /v1/chat/completions and /v1/responses. This shim establishes the Google package contract and metadata injection layer; full import-only parity depends on gateway-side Google adapter support.