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

sciarg: fix partitioning #159

Merged
merged 3 commits into from
Nov 1, 2024
Merged

sciarg: fix partitioning #159

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
c4f14b8
test label counts for labeled_partitions in sciarg
ArneBinder Nov 1, 2024
99505bd
fix and simplify tests
ArneBinder Nov 1, 2024
3086c6a
fix allowing newlines between matching tags (important for abstract)
ArneBinder Nov 1, 2024
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
3 changes: 2 additions & 1 deletion dataset_builders/pie/sciarg/sciarg.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -123,7 +123,8 @@ def _generate_document(self, example, **kwargs):
def document_converters(self) -> DocumentConvertersType:
regex_partitioner = RegexPartitioner(
partition_layer_name="labeled_partitions",
pattern="<([^>/]+)>.*</\\1>",
# find matching tags, allow newlines in between (s flag) and capture the tag name
pattern="<([^>/]+)>(?s:.)*?</\\1>",
label_group_id=1,
label_whitelist=["Title", "Abstract", "H1"],
skip_initial_partition=True,
Expand Down
34 changes: 21 additions & 13 deletions tests/dataset_builders/pie/sciarg/test_sciarg.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -56,6 +56,21 @@
"spans": {"background_claim": 2752, "data": 4093, "own_claim": 5450},
},
}
CONVERTED_LAYER_MAPPING = {
"default": {
"spans": "labeled_spans",
"relations": "binary_relations",
},
"resolve_parts_of_same": {
"spans": "labeled_multi_spans",
"relations": "binary_relations",
},
}
FULL_LABEL_COUNTS_CONVERTED = {
variant: {CONVERTED_LAYER_MAPPING[variant][ln]: value for ln, value in counts.items()}
for variant, counts in FULL_LABEL_COUNTS.items()
}
LABELED_PARTITION_COUNTS = {"Abstract": 40, "H1": 340, "Title": 40}


def resolve_annotation(annotation: Annotation) -> Any:
Expand Down Expand Up @@ -257,32 +272,25 @@ def converted_dataset(dataset, target_document_type) -> Optional[DatasetDict]:
return dataset.to_document_type(target_document_type)


def test_converted_datasets(converted_dataset, dataset_variant):
def test_converted_datasets(converted_dataset, dataset_variant, target_document_type):
if converted_dataset is not None:
split_sizes = {name: len(ds) for name, ds in converted_dataset.items()}
assert split_sizes == SPLIT_SIZES
if dataset_variant == "default":
expected_document_type = TextDocumentWithLabeledSpansAndBinaryRelations
layer_name_mapping = {
"spans": "labeled_spans",
"relations": "binary_relations",
}
elif dataset_variant == "resolve_parts_of_same":
expected_document_type = TextDocumentWithLabeledMultiSpansAndBinaryRelations
layer_name_mapping = {
"spans": "labeled_multi_spans",
"relations": "binary_relations",
}
else:
raise ValueError(f"Unknown dataset variant: {dataset_variant}")

assert issubclass(converted_dataset.document_type, expected_document_type)
assert isinstance(converted_dataset["train"][0], expected_document_type)

if TEST_FULL_DATASET:
expected_label_counts = {
layer_name_mapping[ln]: value
for ln, value in FULL_LABEL_COUNTS[dataset_variant].items()
}
# copy to avoid modifying the original dict
expected_label_counts = {**FULL_LABEL_COUNTS_CONVERTED[dataset_variant]}
if issubclass(target_document_type, TextDocumentWithLabeledPartitions):
expected_label_counts["labeled_partitions"] = LABELED_PARTITION_COUNTS
assert_dataset_label_counts(converted_dataset, expected_label_counts)


Expand Down
Loading