Spaces:
Sleeping
Sleeping
File size: 5,128 Bytes
299567b 3c0d20e 299567b 3c0d20e 299567b 3c0d20e 299567b 3c0d20e 299567b 3c0d20e c4fe3e6 5e96f99 c4fe3e6 3c0d20e 299567b 3e47cef 84098e5 299567b 84098e5 299567b 3e47cef 84098e5 3e47cef 3c0d20e 299567b 3c0d20e 5e96f99 3c0d20e | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 | """Endpoint tests using FastAPI's TestClient with the model stubbed out.
The `client` fixture patches `app.main.load_model` so the lifespan loads the FakeModel
instead of downloading MiniLM, while still exercising the real batcher and endpoint code.
"""
import pytest
from fastapi.testclient import TestClient
import app.main as main_module
from app.batching import QueueFullError, RequestTimeoutError
from app.config import settings
from app.main import app
@pytest.fixture
def client(monkeypatch, make_model):
"""A TestClient whose lifespan loads a FakeModel instead of downloading MiniLM."""
monkeypatch.setattr(main_module, "load_model", lambda *args, **kwargs: make_model())
with TestClient(app) as test_client:
yield test_client
def test_landing(client):
"""GET / is an HTML landing that lists the public endpoints."""
resp = client.get("/")
assert resp.status_code == 200
assert "text/html" in resp.headers["content-type"]
body = resp.text
for path in ("/health", "/metrics", "/embed", "/docs", "/redoc", "/openapi.json"):
assert path in body
assert 'id="runtime"' in body
assert "Server status" in body
assert "queue_depth" in body
assert "ttft_ms" in body
def test_health(client):
resp = client.get("/health")
assert resp.status_code == 200
body = resp.json()
assert body["status"] == "ready"
assert body["model_loaded"] is True
assert body["worker_alive"] is True
assert body["queue_depth"] == 0
assert body["inflight"] == 0
assert body["max_queue_size"] == settings.max_queue_size
assert body["max_batch_size"] == settings.max_batch_size
assert "model" in body
assert "device" in body
assert "backend" in body
def test_health_not_ready_returns_503(client, monkeypatch):
"""/health returns 503 when the batch worker is not alive."""
monkeypatch.setattr(app.state.batcher, "is_running", lambda: False)
resp = client.get("/health")
assert resp.status_code == 503
assert resp.json()["status"] == "not ready"
def test_metrics(client):
"""/metrics reuses the /health snapshot and stays 200 when ready."""
resp = client.get("/metrics")
assert resp.status_code == 200
body = resp.json()
assert body["status"] == "ready"
assert body["queue_depth"] == 0
assert body["inflight"] == 0
assert body["max_queue_size"] == settings.max_queue_size
assert body["max_batch_size"] == settings.max_batch_size
assert "backend" in body
assert "device" in body
def test_metrics_stays_200_when_not_ready(client, monkeypatch):
"""/metrics is a scrape endpoint, not a readiness probe."""
monkeypatch.setattr(app.state.batcher, "is_running", lambda: False)
resp = client.get("/metrics")
assert resp.status_code == 200
assert resp.json()["status"] == "not ready"
def test_request_id_generated(client):
"""Every response carries an X-Request-ID header."""
resp = client.get("/health")
assert resp.headers.get("X-Request-ID")
def test_request_id_echoed(client):
"""An inbound X-Request-ID is echoed back on the response."""
resp = client.get("/health", headers={"X-Request-ID": "abc123"})
assert resp.headers.get("X-Request-ID") == "abc123"
def test_embed(client, make_model):
resp = client.post("/embed", json={"texts": ["hello", "world"]})
assert resp.status_code == 200
body = resp.json()
assert body["num_texts"] == 2
assert body["dim"] == 4
assert len(body["embeddings"]) == 2
assert body["embeddings"] == make_model().encode(["hello", "world"]).tolist()
assert body["ttfr_ms"] >= body["ttft_ms"] >= 0
def test_embed_rejects_empty_texts(client):
"""An empty texts list fails validation (422) before reaching the model."""
assert client.post("/embed", json={"texts": []}).status_code == 422
def test_embed_rejects_too_many_texts(client, monkeypatch):
"""More than max_texts_per_request texts is rejected with 422."""
monkeypatch.setattr(settings, "max_texts_per_request", 2)
assert client.post("/embed", json={"texts": ["a", "b", "c"]}).status_code == 422
def test_embed_rejects_long_text(client, monkeypatch):
"""A text longer than max_chars_per_text is rejected with 422."""
monkeypatch.setattr(settings, "max_chars_per_text", 5)
assert client.post("/embed", json={"texts": ["this is too long"]}).status_code == 422
def test_embed_queue_full_returns_503(client, monkeypatch):
"""A QueueFullError from the batcher is mapped to 503 by the endpoint."""
async def reject(texts):
raise QueueFullError("full")
monkeypatch.setattr(app.state.batcher, "submit", reject)
assert client.post("/embed", json={"texts": ["x"]}).status_code == 503
def test_embed_timeout_returns_504(client, monkeypatch):
"""A RequestTimeoutError from the batcher is mapped to 504 by the endpoint."""
async def time_out(texts):
raise RequestTimeoutError("timeout")
monkeypatch.setattr(app.state.batcher, "submit", time_out)
assert client.post("/embed", json={"texts": ["x"]}).status_code == 504
|