From 4f880822e73197236c289c2c294a5f5c742766bd Mon Sep 17 00:00:00 2001
From: Eva Bardou <bardou@teklia.com>
Date: Tue, 23 Jan 2024 11:52:33 +0000
Subject: [PATCH] Allow extracting elements holding transcriptions without
entities
---
dan/datasets/extract/arkindex.py | 6 +++++-
tests/test_extract.py | 26 ++++++++++++++++++++++++++
2 files changed, 31 insertions(+), 1 deletion(-)
diff --git a/dan/datasets/extract/arkindex.py b/dan/datasets/extract/arkindex.py
index d8dcc1f6..12dff600 100644
--- a/dan/datasets/extract/arkindex.py
+++ b/dan/datasets/extract/arkindex.py
@@ -106,9 +106,10 @@ class ArkindexExtractor:
raise NoTranscriptionError(element.id)
transcription = random.choice(transcriptions)
+ stripped_text = transcription.text.strip()
if not self.tokens:
- return transcription.text.strip()
+ return stripped_text
entities = get_transcription_entities(
transcription.id,
@@ -116,6 +117,9 @@ class ArkindexExtractor:
supported_types=list(self.tokens),
)
+ if not entities.count():
+ return stripped_text
+
return self.translate(
entities_to_xml(
transcription.text, entities, entity_separators=self.entity_separators
diff --git a/tests/test_extract.py b/tests/test_extract.py
index 29132e86..1bf92c3a 100644
--- a/tests/test_extract.py
+++ b/tests/test_extract.py
@@ -425,6 +425,32 @@ def test_empty_transcription(allow_empty, mock_database):
extractor.extract_transcription(element_no_transcription)
+@pytest.mark.parametrize("tokens", (None, EXTRACTION_DATA_PATH / "tokens.yml"))
+def test_extract_transcription_no_translation(mock_database, tokens):
+ extractor = ArkindexExtractor(
+ element_type=["text_line"],
+ entity_separators=None,
+ tokens=tokens,
+ )
+
+ element = Element.get_by_id("test-page_1-line_1")
+ # Deleting one of the two transcriptions from the element
+ Transcription.get(
+ Transcription.element == element,
+ Transcription.worker_version_id == "worker_version_id",
+ ).delete_instance(recursive=True)
+
+ # Deleting all entities on the element remaining transcription while leaving the transcription intact
+ if tokens:
+ TranscriptionEntity.delete().where(
+ TranscriptionEntity.transcription
+ == Transcription.select().where(Transcription.element == element).get()
+ ).execute()
+
+ # Early return with only the element transcription text instead of a translation
+ assert extractor.extract_transcription(element) == "Coupez Bouis 7.12.14"
+
+
@pytest.mark.parametrize(
"nestation, xml_output, separators",
(
--
GitLab