Found during adversarial review of #192.
After #192, every current structure record has at least one evidence-backed causal graph. The next scoring frontier is graph-kind breadth: 12 of 19 records still have exactly one graph and are flagged with single_causal_graph by scripts/score_causal_graphs.py.
The poorest-first list starts with these five records, all tied at 76 after #192:
data/structures/division_machinery/divisome_complex.yaml
data/structures/ribonucleoprotein/ribosome.yaml
data/structures/envelope/peptidoglycan_based_cell_wall.yaml
data/structures/nucleoid/nucleoid.yaml
data/structures/division_machinery/elongasome.yaml
Suggested fix
Work down the scorer output and add a complementary mechanistic graph where the record currently covers only one axis:
- FUNCTION-only records usually need an ASSEMBLY graph.
- ASSEMBLY-only records usually need a FUNCTION or REGULATION graph.
- Add new source-backed edges through
write_validated_structure so YAML serialization remains canonical.
- Keep regenerating
pages/ after each tranche.
This is a follow-on curation issue, not a blocker for #192, whose purpose is to make the ranking and the remaining gaps explicit.
Found during adversarial review of #192.
After #192, every current structure record has at least one evidence-backed causal graph. The next scoring frontier is graph-kind breadth: 12 of 19 records still have exactly one graph and are flagged with
single_causal_graphbyscripts/score_causal_graphs.py.The poorest-first list starts with these five records, all tied at 76 after #192:
Suggested fix
Work down the scorer output and add a complementary mechanistic graph where the record currently covers only one axis:
write_validated_structureso YAML serialization remains canonical.pages/after each tranche.This is a follow-on curation issue, not a blocker for #192, whose purpose is to make the ranking and the remaining gaps explicit.