mirror of
https://github.com/p2p-ld/nwb-linkml.git
synced 2024-11-10 16:54:29 +00:00
145 lines
4.9 KiB
Python
145 lines
4.9 KiB
Python
import h5py
|
|
import networkx as nx
|
|
import numpy as np
|
|
import pytest
|
|
|
|
from nwb_linkml.io.hdf5 import HDF5IO, filter_dependency_graph, hdf_dependency_graph, truncate_file
|
|
from nwb_linkml.maps.hdf5 import resolve_hardlink
|
|
|
|
|
|
@pytest.mark.skip()
|
|
@pytest.mark.parametrize("dset", ["aibs.nwb", "aibs_ecephys.nwb"])
|
|
def test_hdf_read(data_dir, dset):
|
|
NWBFILE = data_dir / dset
|
|
io = HDF5IO(path=NWBFILE)
|
|
# the test for now is just whether we can read it lol
|
|
_ = io.read()
|
|
|
|
|
|
def test_truncate_file(tmp_output_dir):
|
|
source = tmp_output_dir / "truncate_source.hdf5"
|
|
|
|
# create a dang ol hdf5 file with a big dataset and some softlinks and make sure
|
|
# we truncate the dataset and preserve softlink
|
|
|
|
h5f = h5py.File(str(source), "w")
|
|
data_group = h5f.create_group("data")
|
|
dataset_contig = h5f.create_dataset(
|
|
"/data/dataset_contig",
|
|
data=np.zeros((1000, 30, 40), dtype=np.float64),
|
|
compression="gzip",
|
|
compression_opts=9,
|
|
)
|
|
dataset_chunked = h5f.create_dataset(
|
|
"/data/dataset_chunked",
|
|
data=np.zeros((1000, 40, 50), dtype=np.float64),
|
|
compression="gzip",
|
|
compression_opts=9,
|
|
chunks=True,
|
|
)
|
|
dataset_contig.attrs["reference_other"] = dataset_chunked.ref
|
|
dataset_chunked.attrs["reference_other"] = dataset_contig.ref
|
|
dataset_contig.attrs["anattr"] = 1
|
|
|
|
link_group = h5f.create_group("link/child")
|
|
link_group.attrs["reference_contig"] = dataset_contig.ref
|
|
link_group.attrs["reference_chunked"] = dataset_chunked.ref
|
|
h5f.flush()
|
|
h5f.close()
|
|
|
|
source_size = source.stat().st_size
|
|
|
|
# do it without providing target to check that we make filename correctly
|
|
n = 10
|
|
target_output = truncate_file(source, n=n)
|
|
assert target_output == source.parent / (source.stem + "_truncated.hdf5")
|
|
# check that we actually made it smaller
|
|
target_size = target_output.stat().st_size
|
|
# empirically, the source dataset is ~125KB and truncated is ~17KB
|
|
assert target_size < source_size / 5
|
|
|
|
# then check that we have what's expected in the file
|
|
target_h5f = h5py.File(target_output, "r")
|
|
|
|
# truncation happened
|
|
assert target_h5f["data"]["dataset_contig"].shape == (n, 30, 40)
|
|
assert target_h5f["data"]["dataset_chunked"].shape == (n, 40, 50)
|
|
# references still work
|
|
# can't directly assess object identity equality with "is"
|
|
# so this tests if the referenced dereference and that they dereference to the right place
|
|
assert (
|
|
target_h5f[target_h5f["data"]["dataset_contig"].attrs["reference_other"]].name
|
|
== target_h5f["data"]["dataset_chunked"].name
|
|
)
|
|
assert (
|
|
target_h5f[target_h5f["data"]["dataset_chunked"].attrs["reference_other"]].name
|
|
== target_h5f["data"]["dataset_contig"].name
|
|
)
|
|
assert (
|
|
target_h5f[target_h5f["link"]["child"].attrs["reference_contig"]].name
|
|
== target_h5f["data"]["dataset_contig"].name
|
|
)
|
|
assert (
|
|
target_h5f[target_h5f["link"]["child"].attrs["reference_chunked"]].name
|
|
== target_h5f["data"]["dataset_chunked"].name
|
|
)
|
|
assert target_h5f["data"]["dataset_contig"].attrs["anattr"] == 1
|
|
|
|
|
|
def test_dependencies_hardlink(nwb_file):
|
|
"""
|
|
Test that hardlinks are resolved (eg. from /processing/ecephys/LFP/ElectricalSeries/electrodes
|
|
to /acquisition/ElectricalSeries/electrodes
|
|
Args:
|
|
nwb_file:
|
|
|
|
Returns:
|
|
|
|
"""
|
|
parent = "/processing/ecephys/LFP/ElectricalSeries"
|
|
source = "/processing/ecephys/LFP/ElectricalSeries/electrodes"
|
|
target = "/acquisition/ElectricalSeries/electrodes"
|
|
|
|
# assert that the hardlink exists in the test file
|
|
with h5py.File(str(nwb_file), "r") as h5f:
|
|
node = h5f.get(source)
|
|
linked_node = resolve_hardlink(node)
|
|
assert linked_node == target
|
|
|
|
graph = hdf_dependency_graph(nwb_file)
|
|
# the parent should link to the target as a child
|
|
assert (parent, target) in graph.edges([parent])
|
|
assert graph.edges[parent, target]["label"] == "child"
|
|
|
|
|
|
@pytest.mark.dev
|
|
def test_dependency_graph_images(nwb_file, tmp_output_dir):
|
|
"""
|
|
Generate images of the dependency graph
|
|
"""
|
|
graph = hdf_dependency_graph(nwb_file)
|
|
A_unfiltered = nx.nx_agraph.to_agraph(graph)
|
|
A_unfiltered.draw(tmp_output_dir / "test_nwb_unfiltered.png", prog="dot")
|
|
graph = filter_dependency_graph(graph)
|
|
A_filtered = nx.nx_agraph.to_agraph(graph)
|
|
A_filtered.draw(tmp_output_dir / "test_nwb_filtered.png", prog="dot")
|
|
|
|
|
|
@pytest.mark.parametrize(
|
|
"dset",
|
|
[
|
|
{"name": "aibs.nwb", "source": "sub-738651046_ses-760693773.nwb"},
|
|
{
|
|
"name": "aibs_ecephys.nwb",
|
|
"source": "sub-738651046_ses-760693773_probe-769322820_ecephys.nwb",
|
|
},
|
|
],
|
|
)
|
|
@pytest.mark.dev
|
|
def test_make_truncated_datasets(tmp_output_dir, data_dir, dset):
|
|
input_file = tmp_output_dir / dset["source"]
|
|
output_file = data_dir / dset["name"]
|
|
if not input_file.exists():
|
|
return
|
|
|
|
truncate_file(input_file, output_file, 10)
|