Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Use GET /datasets endpoint for builder #378

Merged
merged 2 commits into from
Apr 17, 2023
Merged

Use GET /datasets endpoint for builder #378

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
a766137
Use GET /datasets endpoint for builder
ebezzi Apr 13, 2023
9ea3714
various changes
ebezzi Apr 14, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
106 changes: 33 additions & 73 deletions tools/cellxgene_census_builder/src/cellxgene_census_builder/build_soma/manifest.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,3 @@
import concurrent.futures
import csv
import io
import logging
Expand Down Expand Up @@ -53,81 +52,42 @@ def load_manifest_from_CxG() -> List[Dataset]:
logging.info("Loading manifest from CELLxGENE data portal...")

# Load all collections and extract dataset_id
collections = fetch_json(f"{CXG_BASE_URI}curation/v1/collections")
assert isinstance(collections, list), "Unexpected REST API response, /curation/v1/collections"
datasets = {
dataset["dataset_id"]: {
"collection_id": collection["collection_id"],
"collection_name": null_to_empty_str(collection["name"]),
"collection_doi": null_to_empty_str(collection["doi"]),
"dataset_id": dataset["dataset_id"],
}
for collection in collections
for dataset in collection["datasets"]
}
logging.info(f"Found {len(datasets)} datasets, in {len(collections)} collections")

# load per-dataset schema version
with concurrent.futures.ThreadPoolExecutor() as tp:
dataset_metadata = tp.map(
lambda d: fetch_json(
f"{CXG_BASE_URI}curation/v1/collections/{d['collection_id']}/datasets/{d['dataset_id']}"
),
datasets.values(),
)
for d in dataset_metadata:
assert (
isinstance(d, dict) and "dataset_id" in d
), "Unexpected REST API response, /curation/v1/collections/.../datasets/..."
datasets[d["dataset_id"]].update(
{
"schema_version": d["schema_version"],
"dataset_title": null_to_empty_str(d["title"]),
}
)
datasets = fetch_json(f"{CXG_BASE_URI}curation/v1/datasets")
assert isinstance(datasets, list), "Unexpected REST API response, /curation/v1/datasets"

# Remove any datasets that don't match our target schema version
obsolete_dataset_ids = [id for id in datasets if datasets[id]["schema_version"] not in CXG_SCHEMA_VERSION_IMPORT]
if len(obsolete_dataset_ids) > 0:
logging.warning(f"Dropping {len(obsolete_dataset_ids)} datasets due to unsupported schema version")
for id in obsolete_dataset_ids:
logging.info(f"Dropping dataset_id {id} due to schema version.")
datasets.pop(id)

# Grab the asset URI for each dataset
with concurrent.futures.ThreadPoolExecutor(max_workers=16) as tp:
dataset_assets = tp.map(
lambda d: (
d["dataset_id"],
fetch_json(
f"{CXG_BASE_URI}curation/v1/collections/{d['collection_id']}/datasets/{d['dataset_id']}/assets"
),
),
datasets.values(),
)
no_asset_found = []
for dataset_id, assets in dataset_assets:
assert isinstance(
assets, list
), "Unexpected REST API response, /curation/v1/collections/.../datasets/.../assets"
response = []

for dataset in datasets:
dataset_id = dataset["dataset_id"]
schema_version = dataset["schema_version"]

if schema_version not in CXG_SCHEMA_VERSION_IMPORT:
logging.warning(f"Dropping dataset {dataset_id} due to unsupported schema version")
continue

assets = dataset.get("assets", [])
assets_h5ad = [a for a in assets if a["filetype"] == "H5AD"]
if len(assets_h5ad) == 0:
if not assets_h5ad:
logging.error(f"Unable to find H5AD asset for dataset id {dataset_id} - ignoring this dataset")
no_asset_found.append(dataset_id)
else:
asset = assets_h5ad[0]
datasets[dataset_id].update(
{
"corpora_asset_h5ad_uri": asset["presigned_url"],
"asset_h5ad_filesize": asset["filesize"],
}
)

# drop any datasets where we could not find an asset
for id in no_asset_found:
datasets.pop(id, None)

return [Dataset(**d) for d in datasets.values()]
continue
asset_h5ad_uri = assets[0]["url"]
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

worth asserting that len(assets) == 1 as a sanity check? Or at least logging this unexpected case as a warning.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I added a logging.error and a continue.

asset_h5ad_filesize = assets[0]["filesize"]

d = Dataset(
dataset_id=dataset_id,
corpora_asset_h5ad_uri=asset_h5ad_uri,
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@danieljhegeman title is a required (non-null) field, right?

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Seems like a good time to rename corpora to dataset

dataset_title=null_to_empty_str(dataset["title"]),
collection_id=dataset["collection_id"],
collection_name=null_to_empty_str(dataset["collection_name"]),
collection_doi=null_to_empty_str(dataset["collection_doi"]),
asset_h5ad_filesize=asset_h5ad_filesize,
schema_version=schema_version,
)
response.append(d)

logging.info(f"Found {len(datasets)} datasets")
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

out-of-scope, but it would be nice to report all the warning type counts here (datasets excluded for schema, missing h5ad asset)


return response


def load_manifest(manifest_fp: Optional[Union[str, io.TextIOBase]] = None) -> List[Dataset]:
Expand Down
120 changes: 61 additions & 59 deletions tools/cellxgene_census_builder/tests/test_manifest.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,8 +1,7 @@
import pathlib
import re
from unittest.mock import patch

from cellxgene_census_builder.build_soma.manifest import CXG_BASE_URI, load_manifest
from cellxgene_census_builder.build_soma.manifest import load_manifest


def test_load_manifest_from_file(tmp_path: pathlib.Path, manifest_csv: str) -> None:
Expand Down Expand Up @@ -42,23 +41,26 @@ def test_load_manifest_from_cxg() -> None:
If no parameters are specified, `load_manifest` should load the dataset list from Discover API.
"""
with patch("cellxgene_census_builder.build_soma.manifest.fetch_json") as m:

def mock_call_fn(uri): # type: ignore
if uri == f"{CXG_BASE_URI}curation/v1/collections":
return [
{
"collection_id": "collection_1",
"doi": None,
"name": "1",
"datasets": [{"dataset_id": "dataset_id_1"}, {"dataset_id": "dataset_id_2"}],
}
]
elif m := re.match(rf"{CXG_BASE_URI}curation/v1/collections/(\w+)/datasets/(\w+)$", uri):
return {"dataset_id": m[2], "schema_version": "3.0.0", "title": f"dataset #{m[2]}"}
elif m := re.match(rf"{CXG_BASE_URI}curation/v1/collections/(\w+)/datasets/(\w+)/assets$", uri):
return [{"filetype": "H5AD", "filesize": 1024, "presigned_url": f"https://fake.url/{m[2]}.h5ad"}]

m.side_effect = mock_call_fn
m.return_value = [
{
"dataset_id": "dataset_id_1",
"collection_id": "collection_1",
"collection_name": "1",
"collection_doi": None,
"title": "dataset #1",
"schema_version": "3.0.0",
"assets": [{"filesize": 123, "filetype": "H5AD", "url": "https://fake.url/dataset_id_1.h5ad"}],
},
{
"dataset_id": "dataset_id_2",
"collection_id": "collection_1",
"collection_name": "1",
"collection_doi": None,
"title": "dataset #2",
"schema_version": "3.0.0",
"assets": [{"filesize": 456, "filetype": "H5AD", "url": "https://fake.url/dataset_id_2.h5ad"}],
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

worth adding a non-H5AD asset as well, either here or in an appropriate test

},
]

manifest = load_manifest(None)
assert len(manifest) == 2
Expand All @@ -73,27 +75,26 @@ def test_load_manifest_from_cxg_excludes_datasets_with_old_schema() -> None:
`load_manifest` should exclude datasets that do not have a current schema version.
"""
with patch("cellxgene_census_builder.build_soma.manifest.fetch_json") as m:

def mock_call_fn(uri): # type: ignore
if uri == f"{CXG_BASE_URI}curation/v1/collections":
return [
{
"collection_id": "collection_1",
"doi": None,
"name": "1",
"datasets": [{"dataset_id": "dataset_id_1"}, {"dataset_id": "dataset_id_2"}],
}
]
elif m := re.match(rf"{CXG_BASE_URI}curation/v1/collections/(\w+)/datasets/(\w+)$", uri):
return {
"dataset_id": m[2],
"schema_version": "3.0.0" if m[2] == "dataset_id_1" else "2.0.0",
"title": f"dataset #{m[2]}",
}
elif m := re.match(rf"{CXG_BASE_URI}curation/v1/collections/(\w+)/datasets/(\w+)/assets$", uri):
return [{"filetype": "H5AD", "filesize": 1024, "presigned_url": f"https://fake.url/{m[2]}.h5ad"}]

m.side_effect = mock_call_fn
m.return_value = [
{
"dataset_id": "dataset_id_1",
"collection_id": "collection_1",
"collection_name": "1",
"collection_doi": None,
"title": "dataset #1",
"schema_version": "3.0.0",
"assets": [{"filesize": 123, "filetype": "H5AD", "url": "https://fake.url/dataset_id_1.h5ad"}],
},
{
"dataset_id": "dataset_id_2",
"collection_id": "collection_1",
"collection_name": "1",
"collection_doi": None,
"title": "dataset #2",
"schema_version": "2.0.0", # Old schema version
"assets": [{"filesize": 456, "filetype": "H5AD", "url": "https://fake.url/dataset_id_2.h5ad"}],
},
]

manifest = load_manifest(None)
assert len(manifest) == 1
Expand All @@ -106,25 +107,26 @@ def test_load_manifest_from_cxg_excludes_datasets_with_no_assets() -> None:
`load_manifest` should exclude datasets that do not have assets
"""
with patch("cellxgene_census_builder.build_soma.manifest.fetch_json") as m:

def mock_call_fn(uri): # type: ignore
if uri == f"{CXG_BASE_URI}curation/v1/collections":
return [
{
"collection_id": "collection_1",
"doi": None,
"name": "1",
"datasets": [{"dataset_id": "dataset_id_1"}, {"dataset_id": "dataset_id_2"}],
}
]
elif m := re.match(rf"{CXG_BASE_URI}curation/v1/collections/(\w+)/datasets/(\w+)$", uri):
return {"dataset_id": m[2], "schema_version": "3.0.0", "title": f"dataset #{m[2]}"}
elif m := re.match(rf"{CXG_BASE_URI}curation/v1/collections/(\w+)/datasets/dataset_id_1/assets$", uri):
return [{"filetype": "H5AD", "filesize": 1024, "presigned_url": "https://fake.url/dataset_id_1.h5ad"}]
elif m := re.match(rf"{CXG_BASE_URI}curation/v1/collections/(\w+)/datasets/dataset_id_2/assets$", uri):
return []

m.side_effect = mock_call_fn
m.return_value = [
{
"dataset_id": "dataset_id_1",
"collection_id": "collection_1",
"collection_name": "1",
"collection_doi": None,
"title": "dataset #1",
"schema_version": "3.0.0",
"assets": [{"filesize": 123, "filetype": "H5AD", "url": "https://fake.url/dataset_id_1.h5ad"}],
},
{
"dataset_id": "dataset_id_2",
"collection_id": "collection_1",
"collection_name": "1",
"collection_doi": None,
"title": "dataset #2",
"schema_version": "3.0.0",
"assets": [],
},
]

manifest = load_manifest(None)
assert len(manifest) == 1
Expand Down