Skip to content

Commit 2e0bf61

Browse files
new: Add prodigy entrypoint tests
1 parent c6aee2a commit 2e0bf61

File tree

5 files changed

+129
-12
lines changed

5 files changed

+129
-12
lines changed

tests/prodigy/common.py

Lines changed: 12 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -13,3 +13,15 @@ def get_path(p):
1313
TEST_REF_TOKENS = get_path("test_data/test_reference_to_token_tokens.jsonl")
1414
TEST_REF_SPANS = get_path("test_data/test_reference_to_token_spans.jsonl")
1515
TEST_REF_EXPECTED_SPANS = get_path("test_data/test_reference_to_token_expected.jsonl")
16+
17+
# Prodigy format document containing numbered reference section
18+
19+
TEST_NUMBERED_REFERENCES = get_path("test_data/test_numbered_references.jsonl")
20+
21+
# Prodigy format document with spans annotating every token in the document
22+
23+
TEST_TOKEN_LABELLED = get_path("test_data/test_token_labelled_references.jsonl")
24+
25+
# Reference section in Reach format
26+
27+
TEST_REACH = get_path("test_data/test_reach.jsonl")

tests/prodigy/test_data/test_numbered_references.jsonl

Lines changed: 2 additions & 0 deletions
Large diffs are not rendered by default.

tests/prodigy/test_data/test_reach.jsonl

Lines changed: 1 addition & 0 deletions
Large diffs are not rendered by default.

tests/prodigy/test_data/test_token_labelled_references.jsonl

Lines changed: 72 additions & 0 deletions
Large diffs are not rendered by default.
Lines changed: 42 additions & 12 deletions
Original file line numberDiff line numberDiff line change
@@ -1,20 +1,50 @@
1-
from deep_reference_parser.prodigy import annotate_numbered_references
2-
from deep_reference_parser.prodigy import prodigy_to_tsv
3-
from deep_reference_parser.prodigy import reach_to_prodigy
4-
from deep_reference_parser.prodigy import reference_to_token_annotations
1+
"""Simple tests that entrypoints run. Functionality is tested in other more
2+
specific tests
3+
"""
54

5+
import os
66

7-
def test_annotate_numbered_references_entrypoint():
8-
pass
7+
import pytest
98

9+
from deep_reference_parser.prodigy import (
10+
annotate_numbered_references,
11+
prodigy_to_tsv,
12+
reach_to_prodigy,
13+
reference_to_token_annotations,
14+
)
1015

11-
def test_prodigy_to_tsv():
12-
pass
16+
from .common import TEST_NUMBERED_REFERENCES, TEST_TOKEN_LABELLED, TEST_REACH
1317

1418

15-
def test_reach_to_prodigy():
16-
pass
19+
@pytest.fixture(scope="session")
20+
def tmpdir(tmpdir_factory):
21+
return tmpdir_factory.mktemp("data")
1722

1823

19-
def test_reference_to_token_annotations():
20-
pass
24+
def test_annotate_numbered_references_entrypoint(tmpdir):
25+
annotate_numbered_references(
26+
TEST_NUMBERED_REFERENCES, os.path.join(tmpdir, "references.jsonl")
27+
)
28+
29+
30+
def test_prodigy_to_tsv(tmpdir):
31+
prodigy_to_tsv(
32+
TEST_TOKEN_LABELLED,
33+
os.path.join(tmpdir, "tokens.tsv"),
34+
respect_lines=False,
35+
respect_docs=True,
36+
)
37+
38+
39+
def test_reach_to_prodigy(tmpdir):
40+
reach_to_prodigy(
41+
TEST_REACH,
42+
os.path.join(tmpdir, "prodigy.jsonl")
43+
)
44+
45+
46+
def test_reference_to_token_annotations(tmpdir):
47+
reference_to_token_annotations(
48+
TEST_NUMBERED_REFERENCES,
49+
os.path.join(tmpdir, "tokens.jsonl")
50+
)

0 commit comments

Comments
 (0)