name: "Unit Tests"

on:
  pull_request:
    branches:
      - main
      - "litellm_**"
  push:
    branches:
      - main
  workflow_dispatch:

permissions:
  contents: read

concurrency:
  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.sha }}
  cancel-in-progress: ${{ github.event_name == 'pull_request' }}

# One caller for every tests/test_litellm shard, replacing the nine thin workflow
# files that each wrapped a single call to _test-unit-base.yml. Adding a shard is
# now one matrix entry rather than a new file.
#
# `name` is the shard id and nothing else, so each check reports as
# "<shard> / Run tests" exactly as it did when the shard had its own file. Those
# strings are the branch ruleset's required contexts, so they are load-bearing:
# renaming an entry renames a required check and the ruleset stops matching it.
#
# Every entry states its timeouts even when they equal the base workflow's
# defaults. An absent matrix key renders as an empty string, which is not a
# number, so a partially-specified entry would fail the call rather than fall
# back to the default.
#
# tests/unit/proxy keeps its own caller (test-unit-proxy-db.yml): it is already
# a matrix and carries a shard-coverage guard that reads that file by name.
# Folding it in here is a follow-up, together with generalising that guard into
# assert_ci_coverage.py.
#
# `unit-flag` names the `.circleci/tests.yml` job that now runs part of the
# shard under the same Codecov flag. That pipeline is manual-only while the
# tests migrate, so the shard also runs those files on every event.
jobs:
  unit:
    name: ${{ matrix.shard }}
    permissions:
      contents: read
      id-token: write
      pull-requests: write
    strategy:
      fail-fast: false
      matrix:
        include:
          - shard: mcp-integration
            artifact-name: mcp-integration
            test-path: "tests/mcp_tests"
            unit-flag: mcp-integration
            workers: 2
            reruns: 0
            timeout-minutes: 20
            job-timeout-minutes: 60

          - shard: core-utils
            artifact-name: core-utils
            test-path: ""
            unit-flag: core-utils
            workers: 2
            reruns: 1
            timeout-minutes: 20
            job-timeout-minutes: 60

          - shard: enterprise-routing
            artifact-name: enterprise-routing
            test-path: ""
            unit-flag: enterprise-routing
            workers: 2
            reruns: 2
            timeout-minutes: 20
            job-timeout-minutes: 60

          - shard: integrations
            artifact-name: integrations
            test-path: >-
              tests/test_litellm/integrations
              tests/test_litellm/tracing
            unit-flag: integrations
            workers: 2
            reruns: 3
            timeout-minutes: 20
            job-timeout-minutes: 60

          - shard: Vertex AI
            artifact-name: llm-vertex-ai
            test-path: ""
            unit-flag: llm-vertex-ai
            workers: 1
            reruns: 2
            timeout-minutes: 20
            job-timeout-minutes: 60

          - shard: All Other Providers
            artifact-name: llm-other-providers
            test-path: ""
            unit-flag: llm-other-providers
            workers: 2
            reruns: 2
            timeout-minutes: 20
            job-timeout-minutes: 60

          - shard: misc
            artifact-name: misc
            test-path: >-
              tests/test_litellm/test_*.py
            unit-flag: misc
            workers: 2
            reruns: 2
            timeout-minutes: 20
            job-timeout-minutes: 60

          - shard: proxy-auth
            artifact-name: proxy-auth
            test-path: >-
              tests/unit/proxy/auth
              tests/unit/proxy/hooks
              tests/unit/proxy/policy_engine
              tests/unit/proxy/client
              --ignore=tests/unit/proxy/auth/test_auth_checks.py
              --ignore=tests/unit/proxy/auth/test_user_api_key_auth.py
              --ignore=tests/unit/proxy/auth/test_default_end_user_budget_simple.py
              --ignore=tests/unit/proxy/auth/test_jwt.py
              --ignore=tests/unit/proxy/auth/test_models_fallback_endpoint.py
              --ignore=tests/unit/proxy/auth/test_multipart_bypass_repro.py
              --ignore=tests/unit/proxy/auth/test_proxy_routes.py
              --ignore=tests/unit/proxy/hooks/test_banned_keyword_list.py
              --ignore=tests/unit/proxy/hooks/test_unit_test_max_model_budget_limiter.py
            workers: 2
            reruns: 2
            timeout-minutes: 20
            job-timeout-minutes: 60

          - shard: proxy-endpoints
            artifact-name: proxy-endpoints
            test-path: >-
              tests/unit/proxy/analytics_endpoints
              tests/unit/proxy/management_endpoints
              tests/unit/proxy/list_api
              tests/unit/proxy/memory
              tests/unit/proxy/guardrails
              tests/unit/proxy/management_helpers
              --ignore=tests/unit/proxy/management_endpoints/test_jwt_key_mapping.py
              --ignore=tests/unit/proxy/management_endpoints/test_key_generate_prisma.py
              --ignore=tests/unit/proxy/management_endpoints/test_roi_calculator_endpoints.py
              --ignore=tests/unit/proxy/management_helpers/test_audit_logs_proxy.py
              --ignore=tests/unit/proxy/google_endpoints/test_gemini_agents_endpoints.py
              --ignore=tests/unit/proxy/google_endpoints/test_google_endpoint_routing.py
              --ignore=tests/unit/proxy/google_endpoints/test_google_gemini_proxy_request.py
              --ignore=tests/unit/proxy/public_endpoints/test_blog_posts_endpoint.py
              tests/unit/proxy/anthropic_endpoints
              tests/unit/proxy/google_endpoints
              tests/unit/proxy/openai_files_endpoint
              tests/unit/proxy/batches_endpoints
              tests/unit/proxy/container_endpoints
              tests/unit/proxy/fine_tuning_endpoints
              tests/unit/proxy/vector_store_files_endpoints
              tests/unit/proxy/video_endpoints
              tests/unit/proxy/response_api_endpoints
              tests/unit/proxy/image_endpoints
              tests/unit/proxy/ocr_endpoints
              tests/unit/proxy/vector_store_endpoints
              tests/unit/proxy/agent_endpoints
              tests/unit/proxy/a2a
              tests/unit/proxy/credential_endpoints
              tests/unit/proxy/discovery_endpoints
              tests/unit/proxy/health_endpoints
              tests/unit/proxy/shutdown
              tests/unit/proxy/public_endpoints
              tests/unit/proxy/prompts
              tests/unit/proxy/rag_endpoints
              tests/unit/proxy/rerank_endpoints
              tests/unit/proxy/realtime_endpoints
              tests/unit/proxy/ui_crud_endpoints
              tests/unit/proxy/config_resolvers
              tests/unit/proxy/utils
            workers: 4
            reruns: 2
            timeout-minutes: 20
            job-timeout-minutes: 60

          - shard: proxy-server
            artifact-name: proxy-server
            test-path: "tests/unit/proxy/proxy_server"
            workers: 4
            reruns: 2
            timeout-minutes: 60
            job-timeout-minutes: 100

          - shard: proxy-infra
            artifact-name: proxy-infra
            test-path: >-
              tests/unit/proxy/db
              --ignore=tests/unit/proxy/db/db_transaction_queue/test_e2e_pod_lock_manager.py
              --ignore=tests/unit/proxy/db/test_update_daily_tag_spend.py
              tests/unit/proxy/middleware
              --ignore=tests/unit/proxy/middleware/test_request_size_limit_middleware.py
              tests/unit/proxy/spend_tracking
              --ignore=tests/unit/proxy/spend_tracking/test_search_api_logging.py
              tests/unit/proxy/pass_through_endpoints
              tests/unit/proxy/_experimental
              --ignore=tests/unit/proxy/_experimental/mcp_server
              tests/unit/proxy/experimental
              tests/unit/proxy/common_utils
              --ignore=tests/unit/proxy/common_utils/test_cache_aware_routing.py
              --ignore=tests/unit/proxy/common_utils/test_check_batch_cost.py
              --ignore=tests/unit/proxy/common_utils/test_check_responses_cost.py
              --ignore=tests/unit/proxy/common_utils/test_proxy_encrypt_decrypt.py
              --ignore=tests/unit/proxy/common_utils/test_realtime_cache.py
              tests/unit/proxy/enterprise_billing
              tests/unit/proxy/types_utils
              tests/unit/proxy/logging_endpoints
            unit-flag: proxy-infra
            workers: 4
            reruns: 2
            timeout-minutes: 20
            job-timeout-minutes: 60

          - shard: proxy-infra-root
            artifact-name: proxy-infra-root
            test-path: >-
              tests/unit/proxy/test_*.py
              --ignore=tests/unit/proxy/test_aproxy_startup.py
              --ignore=tests/unit/proxy/test_credential_slot_registry.py
              --ignore=tests/unit/proxy/test_custom_callback_input.py
              --ignore=tests/unit/proxy/test_custom_logger_s3_gcs.py
              --ignore=tests/unit/proxy/test_custom_tokenizer_bug.py
              --ignore=tests/unit/proxy/test_db_schema_changes.py
              --ignore=tests/unit/proxy/test_deprecated_key_grace_period.py
              --ignore=tests/unit/proxy/test_get_favicon.py
              --ignore=tests/unit/proxy/test_get_image.py
              --ignore=tests/unit/proxy/test_prisma_client_backoff_retry.py
              --ignore=tests/unit/proxy/test_prompt_test_endpoint.py
              --ignore=tests/unit/proxy/test_proxy_config_unit_test.py
              --ignore=tests/unit/proxy/test_proxy_custom_auth.py
              --ignore=tests/unit/proxy/test_proxy_reject_logging.py
              --ignore=tests/unit/proxy/test_proxy_server.py
              --ignore=tests/unit/proxy/test_proxy_setting_guardrails.py
              --ignore=tests/unit/proxy/test_proxy_token_counter.py
              --ignore=tests/unit/proxy/test_proxy_utils.py
              --ignore=tests/unit/proxy/test_reducto_ocr_route.py
              --ignore=tests/unit/proxy/test_response_polling_pre_call_checks.py
              --ignore=tests/unit/proxy/test_server_root_path.py
              --ignore=tests/unit/proxy/test_ui_path_detection.py
              --ignore=tests/unit/proxy/test_unit_test_proxy_hooks.py
              --ignore=tests/unit/proxy/test_update_spend.py
              --ignore=tests/unit/proxy/test_zero_cost_model_budget_bypass.py
            workers: 4
            reruns: 2
            timeout-minutes: 20
            job-timeout-minutes: 60

          - shard: caching-local
            artifact-name: caching-local
            test-path: ""
            unit-flag: caching-local
            workers: 2
            reruns: 2
            timeout-minutes: 20
            job-timeout-minutes: 60

          - shard: proxy-extras
            artifact-name: proxy-extras
            test-path: ""
            unit-flag: proxy-extras
            workers: 2
            reruns: 2
            timeout-minutes: 20
            job-timeout-minutes: 60

          - shard: enterprise-package
            artifact-name: enterprise-package
            test-path: ""
            unit-flag: enterprise-package
            workers: 4
            reruns: 2
            timeout-minutes: 20
            job-timeout-minutes: 60

          - shard: responses-caching-types
            artifact-name: responses-caching-types
            test-path: ""
            unit-flag: responses-caching-types
            workers: 2
            reruns: 2
            timeout-minutes: 20
            job-timeout-minutes: 60
    uses: ./.github/workflows/_test-unit-base.yml
    with:
      test-path: ${{ matrix.test-path }}
      unit-flag: ${{ matrix.unit-flag || '' }}
      workers: ${{ matrix.workers }}
      reruns: ${{ matrix.reruns }}
      timeout-minutes: ${{ matrix.timeout-minutes }}
      job-timeout-minutes: ${{ matrix.job-timeout-minutes }}
      artifact-name: ${{ matrix.artifact-name }}
      legacy-mcp-peer: ${{ matrix.shard == 'mcp-integration' }}