# Neo4j Graph Database Import Guide

This folder contains Neo4j-import-ready CSV files containing:
* **Nodes**: Papers, Authors, and deduplicated Datasets.
* **Relationships**: Authorship (`:AUTHORED`) and dataset citations (`:MENTIONS`).

---

## CSV File List

1. **`nodes_papers.csv`**: Contains Policy Research Working Papers.
2. **`nodes_datasets.csv`**: Contains canonicalized datasets.
3. **`nodes_authors.csv`**: Contains unique authors.
4. **`edges_authored.csv`**: Maps Authors to Papers.
5. **`edges_mentions.csv`**: Maps Papers to Datasets with page numbers, context, and model confidence scores.

---

## Import Cypher Queries

To import these files into your Neo4j instance, place the CSV files in your Neo4j project's `import/` directory, open the Neo4j Browser, and execute the following queries:

### 1. Create Constraints
```cypher
CREATE CONSTRAINT UNIQUE_paper FOR (p:Paper) REQUIRE p.id IS UNIQUE;
CREATE CONSTRAINT UNIQUE_dataset FOR (d:Dataset) REQUIRE d.id IS UNIQUE;
CREATE CONSTRAINT UNIQUE_author FOR (a:Author) REQUIRE a.id IS UNIQUE;
```

### 2. Load Nodes
```cypher
// Load Papers
LOAD CSV WITH HEADERS FROM 'file:///nodes_papers.csv' AS row
MERGE (p:Paper {id: row.id})
SET p.title = row.title,
    p.year = toInteger(row.year);

// Load Datasets
LOAD CSV WITH HEADERS FROM 'file:///nodes_datasets.csv' AS row
MERGE (d:Dataset {id: row.id})
SET d.name = row.name,
    d.acronym = row.acronym,
    d.typology = row.typology,
    d.producer = row.producer,
    d.geography = row.geography;

// Load Authors
LOAD CSV WITH HEADERS FROM 'file:///nodes_authors.csv' AS row
MERGE (a:Author {id: row.id})
SET a.name = row.name;
```

### 3. Load Relationships
```cypher
// Load AUTHORED
LOAD CSV WITH HEADERS FROM 'file:///edges_authored.csv' AS row
MATCH (a:Author {id: row.author_id})
MATCH (p:Paper {id: row.paper_id})
MERGE (a)-[:AUTHORED]->(p);

// Load MENTIONS
LOAD CSV WITH HEADERS FROM 'file:///edges_mentions.csv' AS row
MATCH (p:Paper {id: row.paper_id})
MATCH (d:Dataset {id: row.dataset_id})
MERGE (p)-[:MENTIONS {
    pages: split(row.pages, ';'),
    context: row.context,
    confidence: toFloat(row.confidence)
}]->(d);
```