All issues

Search results

MIM and CultureMech both split L-cysteine HCl monohydrate by punctuation: 'x H2O' vs '-HCl x H2O'

Status: Open.

#271 In CultureBotAI/CultureMech;

· realmarcin opened on Aug 15, 2026

Ingredient synonyms are modelled but effectively unpopulated — 22 distinct strings corpus-wide

Status: Open.

#265 In CultureBotAI/CultureMech;

· realmarcin opened on Aug 15, 2026

34 records carry YAML anchors/aliases from the pre-deepcopy split path

Status: Open.

#263 In CultureBotAI/CultureMech;

· realmarcin opened on Aug 15, 2026

Resolve ingredient groundings from MIM's published label index instead of local term.id

Status: Open.

#260 In CultureBotAI/CultureMech;

· realmarcin opened on Aug 14, 2026

term.label often carries the ingredient string instead of the ontology label

Status: Open.

#259 In CultureBotAI/CultureMech;

· realmarcin opened on Aug 14, 2026

One ingredient name, several CHEBI ids: the glucose family is split four ways

Status: Open.

#258 In CultureBotAI/CultureMech;

· realmarcin opened on Aug 14, 2026

Our groundings override MIM downstream: 6,680 divergent rows, some outright wrong compounds (MIM#138)

Status: Open.

#256 In CultureBotAI/CultureMech;

· realmarcin opened on Aug 14, 2026

Corpus repair for the implausible-concentration audit (unfinished half of #118)

Status: Open.

#150 In CultureBotAI/CultureMech;

· realmarcin opened on Jul 31, 2026