MediaIngredientMech mediaingredientmech

Also known as: MIM

MediaIngredientMech is an LLM-assisted curation system for culture-media ingredient ontology mappings, with full audit trails. It manages the ingredient records aggregated from media recipes in CultureMech and owns ingredient identity for the X-Mech fleet. An ingredient record denotes a specific chemical someone can order. Where sources disagree about which form a recipe means, the record picks one by convention, preferring the term that carries the commercial CAS number, and keeps every raw form as a resolvable synonym.

The repository carries no LICENSE file as of 2026-09-10. The CC0 license recorded here is the one the X-Mech suite page states for the fleet.

One record is

a culture-media ingredient

2953 records as of 2026-10-12

Counted from data/ingredients/*.yaml at 1742bfd. Refreshed weekly.

Links

License

Records: CC0 1.0

Domains

chemistry and biochemistry nutrition microbiology

Curation model

Agent-curated	Yes. AI agents generate or maintain most of the content.
Human review	`pull_request`. Agent changes arrive as pull requests that a person reviews and merges.
Agents and tools	Claude Code, culturebotai-claw
Evidence policy	Every mapping carries a MappingEvidence record and every curation action carries an event noting whether an LLM assisted.
Validation in CI	schema ontology_terms cross_field
Append-only history	Yes

Features

ontology_grounding Entities bind to public ontology terms with labels.
append_only_history Every change is an event in an append-only history directory.
knowledge_gaps Records carry structured open questions, controversies and curation to-dos.
static_browser A static site lets a reader browse every record.
sssom_export Mappings export as SSSOM.

Grounding and sources

Ontologies and vocabularies: CHEBI CAS NCIT FOODON ENVO MESH

Source	Relation	What is taken
CultureMech	prov:wasDerivedFrom	Unmapped ingredient strings and occurrence counts from recipes become the curation backlog.
ChEBI	prov:used	Primary mapping target for chemical ingredients.
FOODON	prov:used	Mapping target for complex and food-derived ingredients.

Products

Product	Kind	Format	Description
Ingredient records `mediaingredientmech.records`	RecordCorpus	yaml	Mapped and unmapped ingredient records under data/ingredients/.
MediaIngredientMech LinkML schema `mediaingredientmech.schema`	DataModelProduct	linkml linkml mech_shared
Ingredient mappings (SSSOM) `mediaingredientmech.sssom`	MappingProduct	sssom sssom	Ingredient to ontology mappings exported as SSSOM.
MediaIngredientMech browser `mediaingredientmech.browser`	GraphicalInterface	html

Relations to other Mechs

follows_pattern_of dismech
consumes culturemech: Unmapped ingredient strings and occurrence counts become the curation backlog. Records link back to recipes by stable id with the CultureMechReference class.
provides_to culturemech: Curated ChEBI and FOODON mappings sync back to recipes. CultureMech vendors the ingredient-role enums and consumes the immutable label index.

Referenced by

Contacts

CultureBotAI (Organization) CultureBotAI
Marcin P. Joachimiak (Individual) ORCID 0000-0001-8175-045X mjoachimiak@lbl.gov realmarcin

Registry entry

Created 2026-09-10. Last modified 2026-09-10. Edit this entry on GitHub.

The record count is the sum of the mapped and unmapped ingredient YAML files under data/ingredients/.