source = "This is the first part. This is the second part."
pretranslation = "Esta es la primera parte. Esta es la segunda parte."
align_info = PlaceMarkersAlignmentInfo(
source_tokens=[t for t in TOKENIZER.tokenize(source)],
translation_tokens=[t for t in TOKENIZER.tokenize(pretranslation)],
alignment=to_word_alignment_matrix("0-0 1-1 2-2 3-3 4-4 5-5 6-6 7-7 8-8 9-9 10-10 11-11"),
paragraph_behavior=UpdateUsfmMarkerBehavior.PRESERVE,
style_behavior=UpdateUsfmMarkerBehavior.STRIP,
)
empty_align_info = PlaceMarkersAlignmentInfo(
source_tokens=[],
translation_tokens=[],
alignment=to_word_alignment_matrix(""),
paragraph_behavior=UpdateUsfmMarkerBehavior.PRESERVE,
style_behavior=UpdateUsfmMarkerBehavior.STRIP,
)
usfm = r"""\id MAT
\c 1
\v 1-2 This is the first part.
\p This is the second part.
"""
result = r"""\id MAT
\c 1
\v 1-2 Esta es la primera parte.
\p Esta es la segunda parte.
"""
rows = [
UpdateUsfmRow(scr_ref("MAT 1:1"), str(pretranslation), metadata={"alignment_info": align_info}),
UpdateUsfmRow(scr_ref("MAT 1:2"), "", metadata={"alignment_info": empty_align_info}),
]
target = update_usfm(rows, usfm, update_block_handlers=[PlaceMarkersUsfmUpdateBlockHandler()])
assert_usfm_equals(target, result)
Verse ranges are represented as multiple rows in a
UsfmUpdateBlock, only the first of which is non-empty. The marker placement algorithm uses only one alignment matrix, defaulting to the last, which is empty in the case of a verse range. This results in markers not being placed for verse ranges.Here is a failing test for this case: