feat(cdk-ops): add libdatawarehouse

- Introduces a new Lambda layer `lambda/libdatawarehouse` that provides
  utilities to handle the data warehouse for CloudFront access logs. A
  new CDK construct `LibdatawarehouseLayer` provisions it.

issue codemonger-io#30

cdk-ops/lambda/libdatawarehouse/.gitignore

@@ -0,0 +1,2 @@
+/build
+*.egg-info

cdk-ops/lambda/libdatawarehouse/pyproject.toml

@@ -0,0 +1,3 @@
+[build-system]
+requires = ["setuptools", "wheel"]
+build-backend = "setuptools.build_meta"

cdk-ops/lambda/libdatawarehouse/setup.cfg

@@ -0,0 +1,13 @@
+[metadata]
+name = libdatawarehouse
+version = attr: libdatawarehouse.VERSION
+description = Library for Codemonger Data Warehouse
+
+[options]
+packages =
+	libdatawarehouse
+package_dir =
+	= src
+
+[options.package_data]
+libdatawarehouse = *.pyi, py.typed

cdk-ops/lambda/libdatawarehouse/src/libdatawarehouse/__init__.py

@@ -0,0 +1,8 @@
+# -*- coding: utf-8 -*-
+
+"""Library for Codemonger Data Warehouse.
+"""
+
+VERSION = '0.1.0'
+
+ACCESS_LOGS_DATABASE_NAME = 'access_logs'

cdk-ops/lambda/libdatawarehouse/src/libdatawarehouse/data_api.py

@@ -0,0 +1,37 @@
+# -*- coding: utf-8 -*-
+
+"""Provides utilities to access the Redshift Data API.
+"""
+
+import time
+from typing import Dict, Optional, Tuple
+
+
+RUNNING_STATUSES = ['SUBMITTED', 'PICKED', 'STARTED']
+
+
+def wait_for_results(
+    client,
+    statement_id: str,
+    polling_interval: float = 0.05,
+    polling_timeout: int = round(300 / 0.05),
+) -> Tuple[Optional[str], Dict]:
+    """Waits for a given statement to finish.
+
+    :param RedshiftDataAPIService.Client client: Redshift Data API client.
+
+    :param float polling_interval: interval in seconds between two consecutive
+    pollings.
+
+    :param int polling_timeout: timeout represented as the number of pollings.
+    """
+    polling_counter = 0
+    while True:
+        res = client.describe_statement(Id=statement_id)
+        status = res['Status']
+        if status not in RUNNING_STATUSES:
+            return status, res
+        polling_counter += 1
+        if polling_counter >= polling_timeout:
+            return None, res
+        time.sleep(polling_interval)

cdk-ops/lambda/libdatawarehouse/src/libdatawarehouse/exceptions.py

@@ -0,0 +1,22 @@
+# -*- coding: utf-8 -*-
+
+"""Common exceptions.
+"""
+
+class DataWarehouseException(Exception):
+    """Base exception raised when a data warehouse operation fails.
+    """
+    def __init__(self, message: str):
+        """Initializes with a message.
+        """
+        self.message = message
+
+
+    def __str__(self) -> str:
+        classname = type(self).__name__
+        return f'{classname}({self.message})'
+
+
+    def __repr__(self) -> str:
+        classname = type(self).__name__
+        return f'{classname}({repr(self.message)})'

cdk-ops/lambda/libdatawarehouse/src/libdatawarehouse/tables.py

@@ -0,0 +1,16 @@
+# -*- coding: utf-8 -*-
+
+"""Tables in the data warehouse.
+"""
+
+ACCESS_LOG_TABLE_NAME = 'access_log'
+
+REFERER_TABLE_NAME = 'referer'
+
+PAGE_TABLE_NAME = 'page'
+
+USER_AGENT_TABLE_NAME = 'user_agent'
+
+EDGE_LOCATION_TABLE_NAME = 'edge_location'
+
+RESULT_TYPE_TABLE_NAME = 'result_type'

cdk-ops/lib/libdatawarehouse-layer.ts

@@ -0,0 +1,20 @@
+import * as path from 'path';
+import { aws_lambda as lambda } from 'aws-cdk-lib';
+import { Construct } from 'constructs';
+
+import { PythonLibraryLayer } from 'cdk2-python-library-layer';
+
+/** CDK construct that provisions a Lambda layer of `libdatawarehouse`. */
+export class LibdatawarehouseLayer extends Construct {
+  /** Lambda layer of `libdatawarehouse`. */
+  readonly layer: lambda.ILayerVersion;
+
+  constructor(scope: Construct, id: string) {
+    super(scope, id);
+
+    this.layer = new PythonLibraryLayer(this, 'Layer', {
+      runtime: lambda.Runtime.PYTHON_3_8,
+      entry: path.join('lambda', 'libdatawarehouse'),
+    });
+  }
+}