Buckets:

download
raw
3.09 kB
# Pickling is only supported when the Context is connected to a remote server
# via actual HTTP/TCP, not to an internal app via ASGI.
# We try connecting out to the demo deployment.
import pickle
import httpx
import pytest
import uvicorn
from tiled.client import from_context
from tiled.client.cache import Cache
from tiled.client.context import Context
from tiled.config import Authentication
from tiled.server.app import build_app
from tiled.utils import import_object
from .utils import Server
@pytest.fixture(scope="module")
def server_url():
EXAMPLE = "tiled.examples.generated:tree"
tree = import_object(EXAMPLE)
app = build_app(tree, authentication=Authentication(single_user_api_key="secret"))
config = uvicorn.Config(app, host="127.0.0.1", port=0, log_level="info")
server = Server(config)
with server.run_in_thread() as url:
yield url
def test_pickle_context(server_url):
try:
httpx.get(server_url).raise_for_status()
except Exception:
raise pytest.skip(f"Could not connect to {server_url}")
with Context.from_any_uri(server_url, api_key="secret")[0] as context:
pickle.loads(pickle.dumps(context))
def test_pickle_context_preserves_max_connections(server_url):
"""max_connections survives a pickle/unpickle round-trip."""
try:
httpx.get(server_url).raise_for_status()
except Exception:
raise pytest.skip(f"Could not connect to {server_url}")
with Context.from_any_uri(server_url, api_key="secret", max_connections=3)[
0
] as context:
assert context.max_connections == 3
restored = pickle.loads(pickle.dumps(context))
assert restored.max_connections == 3
@pytest.mark.parametrize("structure_clients", ["numpy", "dask"])
def test_pickle_clients(server_url, structure_clients, tmpdir):
try:
httpx.get(server_url).raise_for_status()
except Exception:
raise pytest.skip(f"Could not connect to {server_url}")
cache = Cache(tmpdir / "http_response_cache.db")
with Context.from_any_uri(server_url, api_key="secret", cache=cache)[0] as context:
client = from_context(context, structure_clients)
pickle.loads(pickle.dumps(client))
for segments in [
[],
["nested", "images", "small_image"],
["tables", "short_table"],
]:
original = client
for segment in segments:
original = original[segment]
roundtripped = pickle.loads(pickle.dumps(original))
roundtripped_twice = pickle.loads(pickle.dumps(roundtripped))
assert roundtripped.uri == roundtripped_twice.uri == original.uri
def test_lock_round_trip(tmpdir):
cache = Cache(tmpdir / "http_response_cache.db")
cache_round_tripped = pickle.loads(pickle.dumps(cache))
cache_round_tripped_twice = pickle.loads(pickle.dumps(cache_round_tripped))
# implementation detail!
assert (
cache._lock.lock
is cache_round_tripped._lock.lock
is cache_round_tripped_twice._lock.lock
)

Xet Storage Details

Size:
3.09 kB
·
Xet hash:
03c6d7da91e1982aa4d1892832a8b8c23a590fa8d7ab7d0c3ed74fdb6e1e1d3c

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.