---

name: entity-labeling-and-ner description: "Use THULAC, the agricultural label taxonomy, and the legacy KNN/fastText label workflow for entity recognition and category prediction." disable-model-invocation: true metadata: disco-role: operating license: GPL 3.0

Entity Labeling and NER

Use this sub-skill for agricultural entity labeling, THULAC-based NER, and the legacy KNN label-prediction workflow.

Read when

• You need to validate or interpret term label files with labels 0-16.
• You need the entity-recognition flow that combines THULAC tags, predicted labels, and Neo4j entity presence checks.
• You need to inspect the KNN classifier interface or its fastText prerequisites.
• You need to work with manual seed labels or predicted label files.

Do

• Start with label taxonomy for the accepted label ids and file formats.
• Use NER and labeling workflow for the recognition flow, seed-label generation, and label file lifecycle.
• Use KNN classifier API when a task depends on the legacy classifier interface or its prerequisites.
• Run label file check before trusting a label file.
• Run KNN feature probe to inspect the non-model prerequisites without downloading the fastText model.

Do not

• Do not treat this sub-skill as the Django UI route owner.
• Do not treat this sub-skill as the Neo4j import/query owner.
• Do not treat this sub-skill as the relation-sentence labeling owner.
• Do not claim live KNN prediction is verified unless the fastText model file and backing graph data are actually present.

Outputs this sub-skill owns

• Numeric agricultural label taxonomy 0-16.
• THULAC + predicted-label NER behavior.
• Label-file validation.
• KNN feature prerequisites and legacy classifier interface notes.
• Manual and seed label workflows.

Quick checks

• python scripts/label_file_check.py --demo
• python scripts/knn_feature_probe.py --demo