Initial curated RAGcore source from c16667f0c9397e2bfc8dd0255b114fba23c8d05e

tests/unit/application/models/test_bootstrap.py

@@ -0,0 +1,118 @@
+from __future__ import annotations
+
+from uuid import uuid4
+
+import httpx
+import pytest
+
+from ragcore.application.models.bootstrap import (
+    EmbeddingBootstrapError,
+    EmbeddingBootstrapRequest,
+    ensure_embedding_profile,
+)
+from ragcore.application.models.registry import (
+    EmbeddingProfileNotFoundError,
+    EmbeddingProfileRegistry,
+)
+from ragcore.domain.control import DataPolicy
+from ragcore.domain.models import EmbeddingProfile, EmbeddingProvider, NormalizationMode
+from ragcore.infrastructure.models.memory import InMemoryEmbeddingProfileStore
+
+
+class FakeClient:
+    def __init__(self, responses: list[httpx.Response]) -> None:
+        self.responses = responses
+        self.requests: list[str] = []
+
+    async def post(self, url: str, **kwargs: object) -> httpx.Response:
+        self.requests.append(url)
+        return self.responses.pop(0)
+
+
+def _response(status: int, body: object) -> httpx.Response:
+    return httpx.Response(status, json=body, request=httpx.Request("POST", "http://test"))
+
+
+def _request() -> EmbeddingBootstrapRequest:
+    return EmbeddingBootstrapRequest(
+        provider=EmbeddingProvider.OLLAMA,
+        model_identifier="nomic-embed-text:1.5",
+        base_url="http://ollama",
+    )
+
+
+def _existing_profile() -> EmbeddingProfile:
+    return EmbeddingProfile(
+        id=uuid4(),
+        profile="hand-registered",
+        version="v1",
+        provider=EmbeddingProvider.OLLAMA,
+        model_identifier="mxbai-embed-large:1",
+        model_digest="a" * 64,
+        dimensions=3,
+        context_limit=4096,
+        tokenizer="sentencepiece-v1",
+        data_policy=DataPolicy.LOCAL_ONLY,
+        normalization=NormalizationMode.NONE,
+    )
+
+
+async def test_noop_when_a_profile_is_already_active() -> None:
+    store = InMemoryEmbeddingProfileStore()
+    registry = EmbeddingProfileRegistry(store)
+    existing = _existing_profile()
+    await registry.register(existing)
+    await registry.activate(existing.id)
+
+    client = FakeClient([])
+    await ensure_embedding_profile(registry, _request(), http_client=client)  # type: ignore[arg-type]
+
+    assert client.requests == []
+    assert (await registry.active()).id == existing.id
+
+
+async def test_noop_and_no_raise_when_no_request_and_no_active_profile() -> None:
+    registry = EmbeddingProfileRegistry(InMemoryEmbeddingProfileStore())
+    client = FakeClient([])
+
+    await ensure_embedding_profile(registry, None, http_client=client)  # type: ignore[arg-type]
+
+    with pytest.raises(EmbeddingProfileNotFoundError):
+        await registry.active()
+
+
+async def test_bootstraps_a_real_profile_from_provider_and_model_probes() -> None:
+    registry = EmbeddingProfileRegistry(InMemoryEmbeddingProfileStore())
+    client = FakeClient(
+        [
+            _response(200, {"digest": "sha256:" + "b" * 64}),
+            _response(200, {"embeddings": [[0.1, 0.2, 0.3, 0.4]]}),
+        ]
+    )
+
+    await ensure_embedding_profile(registry, _request(), http_client=client)  # type: ignore[arg-type]
+
+    active = await registry.active()
+    assert active.model_identifier == "nomic-embed-text:1.5"
+    assert active.model_digest == "b" * 64
+    assert active.dimensions == 4
+    assert client.requests == ["http://ollama/api/show", "http://ollama/api/embed"]
+
+
+async def test_bootstrap_raises_when_the_model_is_unreachable() -> None:
+    registry = EmbeddingProfileRegistry(InMemoryEmbeddingProfileStore())
+    client = FakeClient([_response(500, {"error": "down"})])
+
+    with pytest.raises(EmbeddingBootstrapError):
+        await ensure_embedding_profile(registry, _request(), http_client=client)  # type: ignore[arg-type]
+
+    with pytest.raises(EmbeddingProfileNotFoundError):
+        await registry.active()
+
+
+async def test_bootstrap_raises_on_malformed_digest() -> None:
+    registry = EmbeddingProfileRegistry(InMemoryEmbeddingProfileStore())
+    client = FakeClient([_response(200, {"digest": "not-a-digest"})])
+
+    with pytest.raises(EmbeddingBootstrapError):
+        await ensure_embedding_profile(registry, _request(), http_client=client)  # type: ignore[arg-type]