[Python 3] Boto3 and Webscrape for List of AWS Instances Based on The Nitro System

get_nitro_instance_types.py

import requests
import re
import json
import boto3


def get_all_ec2_types():
    '''
    Modified version of get_instances() from
    https://github.com/powdahound/ec2instances.info/blob/master/ec2.py
    AWS provides no simple API for retrieving all instance types
    '''
    all_instances = []
    pricing_client = boto3.client('pricing', region_name='us-east-1')
    product_pager = pricing_client.get_paginator('get_products')

    product_iterator = product_pager.paginate(
        ServiceCode='AmazonEC2', Filters=[
            # We're gonna assume N. Virginia has all the available types
            {'Type': 'TERM_MATCH', 'Field': 'location',
             'Value': 'US East (N. Virginia)'},
        ]
    )
    for product_item in product_iterator:
        for offer_string in product_item.get('PriceList'):
            offer = json.loads(offer_string)
            product = offer.get('product')

            # Check if it's an instance
            if product.get('productFamily') not in ['Compute Instance', 'Compute Instance (bare metal)', 'Dedicated Host']:
                continue

            product_attributes = product.get('attributes')
            instance_type = product_attributes.get('instanceType')
            if instance_type in ['u-6tb1', 'u-9tb1', 'u-12tb1']:
                # API return the name without the .metal part
                instance_type = instance_type + '.metal'
            all_instances.append(instance_type)

    return set(all_instances)


def get_nitro_prefix_list():
    '''
    Webscrape of latest AWS docs source seems to be only
    way to retrieve a list of Nitro system instance types.

    WARNING: Will break if formatting of target page is
      changed enough to break the prefix_list string parse
    '''
    scrape_raw = requests.get(
        'https://raw.githubusercontent.com/awsdocs/amazon-ec2-user-guide/master/doc_source/instance-types.md')
    prefix_list = re.sub(
        r"(\n)|(and)|(Bare metal:)|`|\ |\+",
        "",
        scrape_raw.text.split(
            "based on the Nitro system:\n+ ")[1].split("\n\n**Resources")[0]
    ).split(',')

    return prefix_list


def filter_nitro_ec2_types(ec2_instances, prefix_list):
    nitro_types = []
    for nitro in prefix_list:
        if '.' in nitro:
            if nitro.lower() in ec2_instances:
                nitro_types.append(nitro)
        else:
            for instance_type in ec2_instances:
                if f"{nitro.lower()}." in instance_type:
                    nitro_types.append(instance_type)
    
    return nitro_types


def get_all_nitro():
    return filter_nitro_ec2_types(
        get_all_ec2_types(),
        get_nitro_prefix_list()
    )


if __name__ == "__main__":
    for instance_type in get_all_nitro():
        print(instance_type)