Ardy is a toolkit to work with AWS Lambas and implement Continuous Integration. AWS Lambda is a serverless compute service that runs your code in response to events and automatically manages the underlying compute resources for you. Alas, AWS Lambda has a very bad GUI interfaces, especially if you work with teams and releases. You can't see at a glance the triggers you have active, the resources of your AWS Lambda or have a version control.
With Ardy you can manage your AWS Lambda with a JSON config file stored in your VCS.
IMPORTANT NOTE: If you want to work with AWS Lambda, it's recommended read about it. Ardy helps and support you to manage your environments but doesn't performs "The black magic" for you.
Install the latest Ardy release via pip:
pip install ardy
You may also install a specific version:
pip install ardy==0.0.1
See the documentation
This project is build with Git Flow. If you want to commit some code use this pattern please:
import operator from nltk import FreqDist from nltk.tokenize import RegexpTokenizer from nltk.book import text6 # Book Monty Python Holy Grail import requests tokens = [f.lower() for f in text6] result_holygrail = FreqDist(tokens) # result_holygrail.most_common(42) holygrail_top = [s[0] for s in sorted([(w, result_holygrail[w]) for w in set(tokens) if len(w) > 4 and result_holygrail[w] > 20], key=operator.itemgetter(1), reverse=True)] tokenizer = RegexpTokenizer(r'\w+') response = requests.get("http://www.angelfire.com/movies/closedcaptioned/meanlife.txt") meanlife = response.text tokens = tokenizer.tokenize(meanlife) result_meanlife = FreqDist(tokens) # result_meanlife.most_common(42) meanlife_top = [s[0] for s in sorted([(w, result_meanlife[w]) for w in set(tokens) if len(w) > 4 and result_meanlife[w] > 20], key=operator.itemgetter(1), reverse=True)] for i in range(0, 30): print("{}: {} {}".format(i+1, holygrail_top[i], meanlife_top[i])) print("{}: {}{}".format(i+1, holygrail_top[i][:2], meanlife_top[i][-2:]))