File size: 10,551 Bytes
0962ea2 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 | """Tests for the v3 API surface — POST /extract/stream (SSE) and
POST /extract/batch + GET /extract/batch/{id}.
Same fake-extractor pattern as test_api.py — no OpenAI key needed.
"""
from __future__ import annotations
import json
import time
from typing import Any
from fastapi.testclient import TestClient
from src.api.batch_store import reset_job_store
from src.api.deps import get_extractor
from src.api.main import create_app
from src.schemas import ExtractionResult, Receipt
from src.utils.cost_tracker import ExtractionMetrics
# --- Fakes ---------------------------------------------------------------
class _FakeExtractor:
"""Fake extractor. Sleep hook lets us test progress ordering."""
def __init__(self, sleep_ms: int = 0, raise_type: type[Exception] | None = None):
self.sleep_ms = sleep_ms
self.raise_type = raise_type
self.calls = 0
def extract(self, file_bytes, filename, doc_type, *, model_override=None, render_images=True):
self.calls += 1
if self.sleep_ms:
time.sleep(self.sleep_ms / 1000.0)
if self.raise_type is not None:
raise self.raise_type("simulated failure")
data = Receipt(merchant=f"MERCHANT-{filename}", total=1.23, currency="USD")
result = ExtractionResult(
document_type=doc_type,
data=data,
field_confidences=[],
overall_confidence=0.9,
warnings=[],
raw_text_snippet=None,
)
metrics = ExtractionMetrics(
input_tokens=100, output_tokens=50, latency_ms=10.0,
model=model_override or "fake-model",
)
return result, metrics
def _build_client(fake) -> TestClient:
reset_job_store() # fresh in-memory store per test
app = create_app()
app.dependency_overrides[get_extractor] = lambda: fake
return TestClient(app)
def _fake_pdf_bytes() -> bytes:
# A real-enough PDF header so content-type sniffing accepts it.
return b"%PDF-1.4\n%fake pdf bytes for tests\n"
# --- SSE parser ----------------------------------------------------------
def _parse_sse(body: bytes) -> list[dict[str, Any]]:
"""Parse a Server-Sent-Events body into a list of {event, data}."""
events: list[dict[str, Any]] = []
for block in body.decode("utf-8").split("\n\n"):
block = block.strip()
if not block:
continue
event = None
data_lines: list[str] = []
for line in block.split("\n"):
if line.startswith("event:"):
event = line.split(":", 1)[1].strip()
elif line.startswith("data:"):
data_lines.append(line.split(":", 1)[1].strip())
payload = json.loads("\n".join(data_lines)) if data_lines else {}
events.append({"event": event, "data": payload})
return events
# =========================================================================
# /extract/stream
# =========================================================================
def test_stream_emits_progress_result_and_done_in_order():
client = _build_client(_FakeExtractor())
resp = client.post(
"/extract/stream",
files={"file": ("r.pdf", _fake_pdf_bytes(), "application/pdf")},
data={"doc_type": "receipt"},
)
assert resp.status_code == 200
assert resp.headers["content-type"].startswith("text/event-stream")
events = _parse_sse(resp.content)
names = [e["event"] for e in events]
assert names[0] == "progress"
assert names[-1] == "done"
assert "result" in names
# Progress must appear at least twice — starting + model_call at minimum.
assert names.count("progress") >= 2
def test_stream_result_carries_full_extraction_result_and_metrics():
client = _build_client(_FakeExtractor())
resp = client.post(
"/extract/stream",
files={"file": ("x.pdf", _fake_pdf_bytes(), "application/pdf")},
data={"doc_type": "receipt"},
)
events = _parse_sse(resp.content)
result_events = [e for e in events if e["event"] == "result"]
assert len(result_events) == 1
payload = result_events[0]["data"]
assert "result" in payload and "metrics" in payload
assert payload["result"]["data"]["merchant"].startswith("MERCHANT-")
assert payload["metrics"]["model"] == "fake-model"
def test_stream_rejects_unknown_doc_type_with_400():
client = _build_client(_FakeExtractor())
resp = client.post(
"/extract/stream",
files={"file": ("r.pdf", _fake_pdf_bytes(), "application/pdf")},
data={"doc_type": "hieroglyph"},
)
# Validation happens *before* the stream opens, so we get a real HTTP status.
assert resp.status_code == 400
body = resp.json()
assert body["error"]["code"] == "unsupported_doc_type"
def test_stream_reports_extractor_failure_via_sse_error_event():
client = _build_client(_FakeExtractor(raise_type=RuntimeError))
resp = client.post(
"/extract/stream",
files={"file": ("r.pdf", _fake_pdf_bytes(), "application/pdf")},
data={"doc_type": "receipt"},
)
# SSE has already started -> 200 OK, error is in-band.
assert resp.status_code == 200
events = _parse_sse(resp.content)
err_events = [e for e in events if e["event"] == "error"]
assert len(err_events) == 1
assert err_events[0]["data"]["code"] == "extraction_failed"
# Even on error, the stream must close with a `done` event.
assert events[-1]["event"] == "done"
# =========================================================================
# /extract/batch + /extract/batch/{id}
# =========================================================================
def test_batch_returns_202_and_job_id():
client = _build_client(_FakeExtractor())
resp = client.post(
"/extract/batch",
files=[
("files", ("a.pdf", _fake_pdf_bytes(), "application/pdf")),
("files", ("b.pdf", _fake_pdf_bytes(), "application/pdf")),
("files", ("c.pdf", _fake_pdf_bytes(), "application/pdf")),
],
data={"doc_type": "receipt"},
)
assert resp.status_code == 202
body = resp.json()
assert "job_id" in body
assert body["progress"]["total"] == 3
def test_batch_polls_progress_and_reports_done():
client = _build_client(_FakeExtractor())
resp = client.post(
"/extract/batch",
files=[
("files", ("a.pdf", _fake_pdf_bytes(), "application/pdf")),
("files", ("b.pdf", _fake_pdf_bytes(), "application/pdf")),
],
data={"doc_type": "receipt"},
)
job_id = resp.json()["job_id"]
# TestClient runs BackgroundTasks synchronously after the response —
# the follow-up GET should already see the job as done.
snap = client.get(f"/extract/batch/{job_id}").json()
assert snap["status"] == "done"
assert snap["progress"]["done"] == 2
assert snap["progress"]["errors"] == 0
assert all(item["status"] == "done" for item in snap["items"])
# Each item carries a validated result + metrics.
for item in snap["items"]:
assert item["result"]["data"]["merchant"].startswith("MERCHANT-")
assert item["metrics"]["model"] == "fake-model"
def test_batch_records_per_item_errors_without_failing_the_job():
"""If one extractor call raises, the item is marked errored — others must still finish."""
client = _build_client(_FakeExtractor(raise_type=RuntimeError))
resp = client.post(
"/extract/batch",
files=[
("files", ("bad.pdf", _fake_pdf_bytes(), "application/pdf")),
],
data={"doc_type": "receipt"},
)
job_id = resp.json()["job_id"]
snap = client.get(f"/extract/batch/{job_id}").json()
assert snap["status"] == "done"
assert snap["progress"]["errors"] == 1
err_item = snap["items"][0]
assert err_item["status"] == "error"
assert "RuntimeError" in err_item["error"]
def test_batch_get_returns_404_for_unknown_job_id():
client = _build_client(_FakeExtractor())
resp = client.get("/extract/batch/does-not-exist")
assert resp.status_code == 404
assert resp.json()["error"]["code"] == "job_not_found"
def test_batch_rejects_unknown_doc_type_up_front():
client = _build_client(_FakeExtractor())
resp = client.post(
"/extract/batch",
files=[("files", ("a.pdf", _fake_pdf_bytes(), "application/pdf"))],
data={"doc_type": "napkin"},
)
assert resp.status_code == 400
assert resp.json()["error"]["code"] == "unsupported_doc_type"
def test_batch_rejects_unsupported_extension_with_415():
client = _build_client(_FakeExtractor())
resp = client.post(
"/extract/batch",
files=[("files", ("resume.docx", b"binary", "application/octet-stream"))],
data={"doc_type": "receipt"},
)
assert resp.status_code == 415
assert resp.json()["error"]["code"] == "unsupported_media_type"
def test_batch_snapshot_shape_matches_contract():
"""Guardrail against accidental field drift in the JSON snapshot."""
client = _build_client(_FakeExtractor())
resp = client.post(
"/extract/batch",
files=[("files", ("a.pdf", _fake_pdf_bytes(), "application/pdf"))],
data={"doc_type": "receipt", "model": "gpt-5-nano"},
)
job_id = resp.json()["job_id"]
snap = client.get(f"/extract/batch/{job_id}").json()
# Top-level keys we\'ve documented in LEARN.md / the code.
for k in ("job_id", "doc_type", "model", "status", "progress", "created_at", "items"):
assert k in snap, f"missing top-level key {k!r}"
assert snap["doc_type"] == "receipt"
assert snap["model"] == "gpt-5-nano"
# Progress keys.
for k in ("total", "done", "errors", "pending", "running"):
assert k in snap["progress"], f"missing progress key {k!r}"
# Item shape.
it = snap["items"][0]
for k in ("index", "filename", "status", "error", "result", "metrics"):
assert k in it, f"missing item key {k!r}"
def test_batch_stores_model_override_on_the_job():
client = _build_client(_FakeExtractor())
resp = client.post(
"/extract/batch",
files=[("files", ("a.pdf", _fake_pdf_bytes(), "application/pdf"))],
data={"doc_type": "receipt", "model": "gpt-5-mini"},
)
job_id = resp.json()["job_id"]
snap = client.get(f"/extract/batch/{job_id}").json()
assert snap["model"] == "gpt-5-mini"
assert snap["items"][0]["metrics"]["model"] == "gpt-5-mini"
|