# szpitale-graph

Graph DB project. Polish hospitals data in Neo4j.

Two-layer architecture. Ingest sources → canonical JSON → load into Neo4j.

## Stack

- Java 21
- Maven (use `./mvnw` wrapper)
- Spring Boot 3.4.4 + Spring Shell
- Neo4j 5
- JUnit 5 + Testcontainers

## Key commands

From `szpitale-graph/`:

```bash
./mvnw -q compile       # compile
./mvnw -q test          # all tests
./mvnw -q test -Dtest=ClassName  # single test
./mvnw -q verify        # full verify
java -jar target/szpitale-graph.jar ingest --voivodeship <voiv>   # layer 1
java -jar target/szpitale-graph.jar load   --voivodeship <voiv>   # layer 2
```

Neo4j:

```bash
docker compose up -d                    # start Neo4j (UI :7474, Bolt :7687)
docker compose down -v                  # stop + remove volumes
```

## Architecture

```
ingest → canonical/*.json → load → Neo4j
```

**Layer 1 (ingest):** Source parsing, normalization, deduplication. No Neo4j dependency.
**Layer 2 (load):** Canonical JSON → UNWIND/MERGE into Neo4j. No network dependency.

Canonical output: `canonical/<voivodeship>.json` or `canonical/all.json`.

## Source data

Input markdown files: `research-<voivodeship>.md`. Table format:

```
| Funkcja | Imię i nazwisko | Afiliacja partyjna | Źródło (URL) |
```

Every cell must have an `https://...` URL.

## Code layout

`src/main/java/com/developx/szpitale/`

| Package | Purpose |
|---|---|
| `model/` | `CanonicalDataset`, `Hospital`, `Person`, `Affiliation`, `Role`, `Mandate`, `RawHospitalRecord` |
| `model/enums/` | `LegalForm`, `RoleType`, `MandateType`, `ConfidenceLevel`, `RoleStatus`, `AffiliationType`, `OrganType`, `SupervisoryBodyType` |
| `ingest/` | `IngestCommand`, `CanonicalWriter`, `IngestResults` |
| `ingest/source/` | `MarkdownRegistrySource`, `VoivodeshipSource` |
| `ingest/normalize/` | `NameNormalizer`, `PartyNormalizer`, `Deduplicator` |
| `load/` | `LoadCommand`, `CanonicalReader`, `GraphSchema`, `GraphLoader`, `GraphValidator` |
| `SzpitaleGraphApplication.java` | `@SpringBootApplication` entry point |

## 16 voivodeships

Supported regions: all Polish voivodeships. Start with `research-<voiv>.md` files, then run `ingest` + `load`.

## Dev

Test-driven development. TDD cycle: red → green → refactor. Tests mirror main package structure.

Integrations tests use Testcontainers with Neo4j container.