Use when working with medical entity types, Neo4j graph model, ICD codes, Vietnamese medical terminology, disease naming conventions, or knowledge graph schema. Triggers on 'entity types', 'Neo4j schema', 'ICD code', 'medical ontology', 'disease naming', 'Vietnamese medical terms'.
| Label | Purpose | Key Properties |
|---|---|---|
:Chunk | PDF-ingested content block | disease_name, block_id, page_number, embedding, content, title, level, section_path |
:Disease | Disease entity | name, icd_code, aliases (list) |
:Protocol | Umbrella for multi-disease PDF | name |
:Page | Web-crawled content (V1 legacy) | title, , , |
descriptionurlembedding| Label | Examples | Properties |
|---|---|---|
:Drug | Amoxicillin, Corticoid, PPI | name, dosage, route |
:Symptom | Đau tai, Chảy mủ tai, Ù tai | name |
:LabTest | CT scan, Thính lực đồ, Nội soi | name |
:Procedure | Phẫu thuật nội soi, Mở khí quản | name |
:Complication | Liệt mặt, Viêm màng não | name |
:OntologyClass | Schema definition | name, type |
(:Chunk)-[:ABOUT_DISEASE]->(:Disease) # Disease scoping
(:Chunk)-[:NEXT_CHUNK]->(:Chunk) # Reading order
(:Chunk)-[:HAS_CHILD]->(:Chunk) # Hierarchy
(:Protocol)-[:COVERS_DISEASE]->(:Disease) # Multi-disease grouping
(:Chunk)-[:MENTIONS]->(:Drug|:Symptom|:LabTest) # Entity extraction
(:Page)-[:LINKS_TO|HAS_REFERENCE]->(:Page) # V1 web links
| Index Name | Label | Property | Dimensions | Similarity |
|---|---|---|---|---|
chunk_vector_index | Chunk | embedding | 1536 | cosine |
clinical_vector_index | Page | embedding | 1536 | cosine |
| Index Name | Label | Properties |
|---|---|---|
chunk_fulltext | Chunk | content, title |
| Official Name | Common Aliases |
|---|---|
| Viêm tai giữa cấp tính trẻ em | viêm tai giữa cấp, viêm tai giữa, VTG cấp |
| Viêm mũi xoang mạn tính | viêm xoang mạn, viêm xoang mạn tính |
| Ung thư vòm mũi họng | ung thư vòm họng, K vòm, NPC |
| Sốt xuất huyết Dengue | sốt xuất huyết, SXH, SXHD, dengue |
ĐỊNH NGHĨA, ĐẠI CƯƠNG, NGUYÊN NHÂN, TRIỆU CHỨNG,
CHẨN ĐOÁN, ĐIỀU TRỊ, TIÊN LƯỢNG, BIẾN CHỨNG,
PHÒNG BỆNH, DỰ PHÒNG, THEO DÕI, TÀI LIỆU THAM KHẢO
Given this medical text chunk, extract entities into these categories:
- Drug: medication names with dosage if mentioned
- Symptom: clinical signs and symptoms
- LabTest: diagnostic tests, imaging, lab work
- Procedure: surgical or clinical procedures
- Complication: adverse outcomes or complications
Return JSON: {"drugs": [...], "symptoms": [...], "lab_tests": [...], "procedures": [...], "complications": [...]}
Text: {chunk_content}
resolve_disease_name(query) in medical_agent.py: