innoldb 1.0.6

a ORM model for AWS QLDB

Makpar Innovation Lab

innoldb

A simple Object-Relation-Mapping for a serverless AWS Quantum Ledger Database backend, and a command line utility for querying tables on those ledgers.

NOTE: The user or process using this library must have an IAM policy that allows access to QLDB.

from innoldb.qldb import Document

document = Document('my-table')
document.field = 'my field'
document.save()

Setup

Overview

1. (Optional) Configure environment
2. Create QLDB Ledger
3. Configure IAM user/role permissions for ledger
4. Install library

Steps

1. (Optional) Configure Environment

This library needs to be informed of the QLDB Ledger to write against on your AWS environment. There are several ways to configure the Ledger setting.

Before you start a Python shell, export theLEDGER environment variable,

export LEDGER='ledger-name'
python

Alternatively, configure the variable directly in a Python script

import os

os.environ['LEDGER'] = 'ledger=name'

The environment variable LEDGER should point to the QLDB ledger so the application knows to which ledger to write. If you do not configure the LEDGER environment variable, you will need to pass in the ledger name to the Document object. See below for more information.

2. Create a QLDB Ledger

Boto3 Client

The easiest way to create a ledger is through the boto3 wrapper function in innoldb.qldb.Driver,

from innoldb.qldb import Driver

Driver.ledger('my-ledger')

NOTE: You must install the innoldb package from PyPi before creating a ledger this way. See Step 4 for more information.

CloudFormation

A QLDB CloudFormation template is also available in the cf directory of this project's Github. A script has been provided to post this template to CloudFormation, assuming your AWS CLI has been authenticated and configured. Clone the repository and then from the project root, execute the following script and specify the <ledger-name> to create a ledger on the QLDB service,

./scripts/cf-stack --ledger <ledger-name>

NOTE: The <ledger-name> must match the value of the LEDGER environment variable. The name of the ledger that is stood up on AWS is passed to the library through this environment variable.

NOTE: This script has other optional arguments detailed in the comments of the script itself.

3. Configure User Permissions

In production, you will want to limit the permissions of the application client to the ledger and table to which it is authorized to read and write. For the purposes of using this library locally, you can add a blanket policy to your user account by following the instructions here.

If you are configuring an application role to use this library for a particular ledger and table, you will need to scope the permissions using this reference.

4. Install innoldb

pip install innoldb

Command Line Usage

Installing the innoldb package puts a command line utility on your path. This tool allows you to query the QLDB ledger directly from the command line.

The --table argument is required. Queries can be constructed against this table by passing in other arguments. See below for examples of the different queries.

Be sure to export the LEDGER environment variable before executing any of these commands,

export LEDGER='ledger-name'

Find Document By ID

innoldb --table <table-name> --id <id>

Find All Documents

innoldb --table <table-name> --all

Generate New Mock Document

innoldb --table <table-name> --mock

Update Field in Document

innoldb --table <table-name> --id <id> --update <field>=<value> <field>=<value> ...

Insert Document

innoldb --table <table-name> --insert <field>=<value> <field>=<value> ...

You can optionally specify the Document ID if you wish to overwrite the auto-generated UUID.

Document Object Model

This library abstracts much of the QLDB implementation away from its user. All the user has to do is create a Document object, add fields to it and then call save(). Under the hood, the library will translate the Document fields into PartiQL queries and use the pyqldb Driver to post the queries to the QLDB instance on AWS.

NOTE: All documents are indexed through the key field id.

Saving

If you have the LEDGER environment variable set, all that is required is to create a Document object and pass it the table name from the QLDB ledger. If the following lines are feed into an interactive Python shell or copied into a script,

from innoldb.qldb import Document

my_document = Document('table-name')
my_document.property_one = 'property 1'
my_document.property_two = 'property 2'
my_document.save()

Then a document will be inserted into the QLDB ledger table. If you do not have the LEDGER environment variable set, you must pass in the ledger name along with the table name through named arguments,

from innoldb.qldb import Document

my_document = Document(table='table-name', ledger='ledger-name')
my_document.property_one = 'property 1'
my_document.property_two = 'property 2'
my_document.save()

The Document class will auto-generate a UUID for each document inserted into the ledger table. If you prefer a more memorable id, you can override the auto-generated ID through the constructor,

from innoldb.qldb import Document

my_document = Document(table='table-name', id='12345')
my_document.property_one = 'this is a test'
my_document.save()

Congratulations! You have saved a document to QLDB!

Loading

To load a document that exists in the ledger table already, pass in the id of the Document when creating a new instance,

from innoldb.qldb import Document

my_document(table='table-name', id='12345')
print(my_document.property_one)

Updating

Updating and saving are different operations, in terms of the PartiQL queries that implement these operations, but from the Document's perspective, they are the same operation; the same method is called in either case. The following script will save a value of test 1 to field and then overwrite it with a value of test 2,

from innoldb.qldb import Document

my_document = Document('table-name')
my_document.field = 'test 1'
my_document.save()
my_document.field = 'test 2'
my_document.save()

Behind the scenes, whenever the save() method is called, a query is run to check for the existence of the given Document. If the Document doesn't exist, the library will create a new one. If the Document does exist, the library will overwrite the existing Document.

Fields

The document fields can be returned as a dict through the fields() method. The following script will loop through the fields on an existing document with id=test and print their corresponding values,

from innoldb.qldb import Document

my_document = Document(table='table-name', id='test')

for key, value in my_document.fields().items():
  print(key, '=', value)

Query Object Model

Queries are represented as an object, Query. Each Query must be initialized with a table that it will run PartialQL queries against. All queries return a list of Document objects.

All

The following script queries the ledger table for all documents and prints a JSON representation of each document to screen,

from innoldb.qldb import Document, Query

all_documents = Query('table-name').all()
for document in all_documents:
  print(document.fields())

Configuration

Log Level

The log level can be set through the environment variable LOG_LEVEL to the values: NOTSET, INFO, DEBUG and ERROR,

export LOG_LEVEL='INFO'
innoldb --table table-name --all

import os
from innoldb.qldb import Query

os.environ['LOG_LEVEL'] = 'DEBUG'
Query('table-name').all()

Build From Source

The innoldb library can be built from source with the following script,

git clone https://github.com/Makpar-Innovation-Laboratory/innoldb
cd innoldb
python -m build
VERSION=$(cat version.txt)
cd dist
pip install innoldb-${VERSION}-py3-none-any.whl

Or use the pre-packaged helper script,

git clone https://github.com/Makpar-Innovation-Laboratory/innoldb
./innoldb/scripts/install

References

• AWS QLDB Documentation
• QLDB Python Driver Documentation
• Boto3 QLDB Client Documentation
• PartiQL Documentation