Second review pass on #260. Cosmetic, filed for the record.
missing_symbols compares the record's declared symbols against the
gene_symbol written on each example, which is UniProt's primary name. Where
the two differ — rodA declared, mrdB recorded — the declared symbol always
looks unrepresented, so the adapter re-queries it on every run. The result is
still correct and idempotent (deduplication by accession drops it, and a second
dry run proposes 0 additions), it is just six wasted queries per run:
roda, pbp2, bmc_h x2, secy_channel_subunit, sece_clamp_subunit
A fix would need to remember which symbol was queried, and gene_symbol is
single-valued. The role text names the queried symbol but matching on prose
would be worse than the waste. Probably not worth fixing.
Second review pass on #260. Cosmetic, filed for the record.
missing_symbolscompares the record's declared symbols against thegene_symbolwritten on each example, which is UniProt's primary name. Wherethe two differ —
rodAdeclared,mrdBrecorded — the declared symbol alwayslooks unrepresented, so the adapter re-queries it on every run. The result is
still correct and idempotent (deduplication by accession drops it, and a second
dry run proposes 0 additions), it is just six wasted queries per run:
A fix would need to remember which symbol was queried, and
gene_symbolissingle-valued. The
roletext names the queried symbol but matching on prosewould be worse than the waste. Probably not worth fixing.