add a simple flask api which uses duckdb

Author: nirdosh17

.gitignore

@@ -0,0 +1,2 @@
+# created by virtualenv automatically
+local/

Dockerfile

@@ -0,0 +1,17 @@
+FROM python:3.10.8-slim
+
+RUN apt-get update && apt-get upgrade -y
+
+WORKDIR /app
+
+COPY requirements.txt /app/requirements.txt
+
+RUN pip install -r requirements.txt
+
+COPY server.py /app/server.py
+
+COPY test.db /app/test.db
+
+EXPOSE 8080
+
+CMD ["python", "server.py"]

Makefile

@@ -0,0 +1,11 @@
+build:
+	pip3 install -r requirements.txt
+
+run: build
+	python3 server.py
+
+docker.build:
+	./build_image.sh
+
+docker.run:
+	docker run --rm -it --name python-api -p 8080:8080 python-api:latest

Readme.md

@@ -0,0 +1,40 @@
+# Containerized Python API using DuckDB
+Sample Containerized Python API using [DuckDB](https://duckdb.org/)
+
+## Flask API
+The service exposes `GET http://localhost:8080/stats` api which returns aggregated user count from DuckDB database `test.duckdb`.
+
+**API response:**
+```json
+[
+	{
+		"date": "2021-02-20",
+		"users_joined": 2598
+	},
+	{
+		"date": "2021-02-21",
+		"users_joined": 2578
+	}
+]
+```
+
+## Running without docker
+```bash
+make run
+```
+
+## Running as a container
+```bash
+# builds docker image
+make docker.build
+
+# runs docker image
+make docker.run
+```
+
+## Test Data
+Test data is coming from database file `test.duckdb`. It contains a table called `users` which has following fields with ~1 million randomly generated rows:
+
+| id (int32)| joined_date (date) | name (varchar)|    email (varchar)      |
+|-----------|--------------------|---------------|-------------------------|
+|      1    |     2021-09-14     |  Jarret Kuhn  |  carsondooley@wolf.name |

build_image.sh

@@ -0,0 +1,10 @@
+#!/bin/sh
+
+if [[ $(uname -m) == *"arm"* ]]; then
+  BUILD_PLATFORM="--platform linux/amd64"
+fi
+
+docker build $BUILD_PLATFORM -t python-api .
+
+echo "cleaning up old docker images..."
+docker system prune -f

requirements.txt

@@ -0,0 +1,11 @@
+click==8.1.3
+distlib==0.3.6
+duckdb==0.7.0
+filelock==3.9.0
+Flask==2.2.3
+itsdangerous==2.1.2
+Jinja2==3.1.2
+MarkupSafe==2.1.2
+platformdirs==3.0.0
+virtualenv==20.19.0
+Werkzeug==2.2.3

server.py

@@ -0,0 +1,29 @@
+from flask import Flask
+import duckdb, json
+
+app = Flask(__name__)
+db = duckdb.connect(database='test.duckdb', read_only=True)
+
+@app.route('/stats', methods=['GET'])
+def stats():
+  db.execute("""
+    SELECT
+        CAST(joined_date AS VARCHAR) as date,
+        MAX(row_number) as users_joined
+      FROM (
+        SELECT
+          joined_date,
+          ROW_NUMBER() OVER (PARTITION BY joined_date) as row_number
+        FROM users
+        WHERE joined_date BETWEEN (now()::TIMESTAMP - INTERVAL '2 years') AND now()
+      ) GROUP BY joined_date
+      ORDER BY joined_date ASC;
+  """)
+  stats = db.fetchall()
+
+  columns = ['date', 'users_joined']
+  items = [dict(zip(columns, row)) for row in stats]
+  return json.dumps(items)
+
+if __name__ == '__main__':
+  app.run(port=8080, host='0.0.0.0')