---

name: graph-query-and-data-management description: "Operate the Agriculture KnowledgeGraph Neo4j graph import/query, CSV schemas, hierarchy tree, and vector utilities." disable-model-invocation: true metadata: disco-role: operating license: GPL 3.0

Graph Query and Data Management

Use this sub-skill when the task is about the Agriculture KnowledgeGraph data model rather than the web UI or model-training pipelines.

Use this for

• Planning or debugging Neo4j imports for Hudong, Wikidata, attribute, weather, and city-climate CSV artifacts.
• Understanding graph labels, relationship types, and relation-search semantics.
• Translating or modernizing the repo's Neo4j wrapper query methods safely.
• Inspecting CSV field meanings, delimiter quirks, and hierarchy/vector utility input formats.
• Running a tiny non-repo smoke check of the hierarchy tree behavior.

Route elsewhere

• Django service startup, URL/view behavior, form fields, and preload side effects: use ../web-app-service/.
• THULAC entity recognition, predicted labels, KNN, or fastText files: use ../entity-labeling-and-ner/.
• Scrapy crawlers, Wikidata conversion scripts, and weather data acquisition: use ../crawlers-and-wikidata-pipelines/.
• PCNN relation extraction datasets, Mongo annotation data, or TensorFlow training: use ../relation-extraction-pipeline/.

Read first

• Neo4j import and query guide
• Data format guide
• Tree and vector API guide
• Workflow troubleshooting

Bundled helpers

• Tree API smoke check creates temporary edge/leaf fixtures and asserts source-compatible hierarchy behavior without importing the original checkout.
• Cypher import templates provides adaptable LOAD CSV templates that use Neo4j import-directory file URLs, not machine-local absolute paths.

Operating cautions

• Treat live Neo4j, MongoDB, the Django service, large word-vector files, and network crawls as external prerequisites unless a task explicitly provisions them.
• The historical source concatenates Cypher strings. When writing new code, prefer parameterized queries and preserve exact label/relationship semantics.
• The source is old and un-packaged; use these references as operating guidance and keep runtime work self-contained.