feat: implement get_datasets endpoint

Author: tillywoodfield

README.md

@@ -14,12 +14,22 @@ source .venv/bin/activate
 pip install -r requirements_dev.txt
 ```
 
+### Set database enrivonment variable
+
+```
+export DATABASE_URL="postgresql://oc4ids_datastore_read_only@localhost/oc4ids_datastore"
+```
+
 ### Run app
 
 ```
 fastapi dev oc4ids_datastore_api/main.py
 ```
 
+### View the OpenAPI schema
+
+While the app is running, go to `http://127.0.0.1:8000/docs/`
+
 ### Run linting and type checking
 
 ```

oc4ids_datastore_api/database.py

@@ -0,0 +1,13 @@
+import os
+from typing import Sequence
+
+from sqlmodel import Session, create_engine, select
+
+from oc4ids_datastore_api.models import DatasetSQLModel
+
+engine = create_engine(os.environ["DATABASE_URL"])
+
+
+def fetch_all_datasets() -> Sequence[DatasetSQLModel]:
+    with Session(engine) as session:
+        return session.exec(select(DatasetSQLModel)).all()

oc4ids_datastore_api/main.py

@@ -1,8 +1,11 @@
 from fastapi import FastAPI
 
+from oc4ids_datastore_api.schemas import Dataset
+from oc4ids_datastore_api.services import get_all_datasets
+
 app = FastAPI()
 
 
-@app.get("/")
-def index() -> str:
-    return "Hello, World!"
+@app.get("/datasets")
+def get_datasets() -> list[Dataset]:
+    return get_all_datasets()

oc4ids_datastore_api/models.py

@@ -0,0 +1,17 @@
+import datetime
+
+from sqlmodel import Field, SQLModel
+
+
+class DatasetSQLModel(SQLModel, table=True):
+    __tablename__ = "dataset"
+
+    dataset_id: str = Field(primary_key=True)
+    source_url: str
+    publisher_name: str
+    license_url: str | None
+    license_name: str | None
+    json_url: str | None
+    csv_url: str | None
+    xlsx_url: str | None
+    updated_at: datetime.datetime

oc4ids_datastore_api/schemas.py

@@ -0,0 +1,25 @@
+import datetime
+
+from pydantic import BaseModel
+
+
+class Publisher(BaseModel):
+    name: str
+
+
+class License(BaseModel):
+    url: str | None
+    name: str | None
+
+
+class Download(BaseModel):
+    format: str
+    url: str
+
+
+class Dataset(BaseModel):
+    loaded_at: datetime.datetime
+    source_url: str
+    publisher: Publisher
+    license: License
+    downloads: list[Download]

oc4ids_datastore_api/services.py

@@ -0,0 +1,23 @@
+from oc4ids_datastore_api.database import fetch_all_datasets
+from oc4ids_datastore_api.models import DatasetSQLModel
+from oc4ids_datastore_api.schemas import Dataset, Download, License, Publisher
+
+
+def _transform_dataset(dataset: DatasetSQLModel) -> Dataset:
+    downloads = [
+        Download(format="json", url=(dataset.json_url or "")),
+        Download(format="csv", url=(dataset.csv_url or "")),
+        Download(format="xlsx", url=(dataset.xlsx_url or "")),
+    ]
+    return Dataset(
+        loaded_at=dataset.updated_at,
+        source_url=dataset.source_url,
+        publisher=Publisher(name=dataset.publisher_name),
+        license=License(url=dataset.license_url, name=dataset.license_name),
+        downloads=downloads,
+    )
+
+
+def get_all_datasets() -> list[Dataset]:
+    datasets = fetch_all_datasets()
+    return [_transform_dataset(dataset) for dataset in datasets]

pyproject.toml

@@ -8,7 +8,9 @@ description = "OC4IDS Datastore API"
 version = "0.1.0"
 readme = "README.md"
 dependencies = [
-  "fastapi[standard]"
+  "fastapi[standard]",
+  "psycopg2",
+  "sqlmodel",
 ]
 
 [project.optional-dependencies]

requirements_dev.txt

@@ -76,10 +76,14 @@ pathspec==0.12.1
     # via black
 platformdirs==4.3.6
     # via black
+psycopg2==2.9.10
+    # via oc4ids-datastore-api (pyproject.toml)
 pycodestyle==2.12.1
     # via flake8
 pydantic==2.10.6
-    # via fastapi
+    # via
+    #   fastapi
+    #   sqlmodel
 pydantic-core==2.27.2
     # via pydantic
 pyflakes==3.2.0
@@ -102,6 +106,10 @@ shellingham==1.5.4
     # via typer
 sniffio==1.3.1
     # via anyio
+sqlalchemy==2.0.37
+    # via sqlmodel
+sqlmodel==0.0.22
+    # via oc4ids-datastore-api (pyproject.toml)
 starlette==0.45.3
     # via fastapi
 typer==0.15.1
@@ -114,6 +122,7 @@ typing-extensions==4.12.2
     #   pydantic
     #   pydantic-core
     #   rich-toolkit
+    #   sqlalchemy
     #   typer
 uvicorn[standard]==0.34.0
     # via