Expand test coverage across nodes, lowfat, and TSV

[jonathanrobie]

Summary

Adds tests that catch the bugs fixed by PRs #136, #137, and #140, and enforce data quality constraints going forward.

New tests

test_nodes.py

• test_file_is_nfc — all text must be Unicode NFC
• test_no_cgj_anywhere — no CGJ (U+034F) anywhere in file
• test_m_xml_id_format — xml:id must be o<digits> (with optional ה suffix for subsumed definite articles)
• test_m_lemma_not_empty — every <m> has non-empty @lemma
• test_m_morph_not_empty — every <m> has non-empty @morph
• test_m_after_valid_values — @after must be from the known valid set
• test_non_final_word_last_m_has_after — last morpheme of each non-final orthographic word has non-empty @after

test_lowfat.py

• test_file_is_nfc — all text must be Unicode NFC
• test_no_cgj_anywhere — no CGJ anywhere in file
• test_w_lemma_not_empty — every <w> has non-empty @lemma
• test_w_after_not_missing — every <w> has an @after attribute

test_tsv.py

• test_tsv_is_nfc — all text must be Unicode NFC
• test_tsv_no_cgj — no CGJ in the TSV
• test_tsv_xml_id_format — xml:id matches expected format
• test_tsv_ref_format — ref matches USFM pattern
• test_tsv_after_valid_values — after column from known valid set
• test_tsv_lemma_not_empty — every row has non-empty lemma

Expected failures until other PRs merge

Test	Passes after
test_file_is_nfc, test_no_cgj_anywhere (nodes)	PR #136 merges
test_file_is_nfc, test_no_cgj_anywhere, test_w_after_not_missing (lowfat)	PRs #136 + #137 merge + lowfat regenerated
test_tsv_is_nfc, test_tsv_no_cgj	PRs #136 + #137 merge + TSV regenerated
test_non_final_word_last_m_has_after (nodes)	PR #137 merges

🤖 Generated with Claude Code