chalice-pywb-app.py

"""
FIXME: Cannot use C extensions. See below:
http://www.perrygeo.com/running-python-with-compiled-code-on-aws-lambda.html
http://stackoverflow.com/questions/34749806/using-moviepy-scipy-and-numpy-in-amazon-lambda
"""


from datetime import datetime
from urlparse import urlparse
from os import path

import boto3
from chalice import Chalice

from pywb.rewrite.rewrite_live import LiveRewriter
from pywb.rewrite.url_rewriter import UrlRewriter


app = Chalice(app_name='archive')

app.debug = True

S3 = boto3.client('s3', region_name='us-east-1')
BUCKET = 'archive.do'
ENDPOINT = 'archive.do.s3-website-us-east-1.amazonaws.com'


@app.route('/{url+}')
def index():
	url = app.current_request.uri_params['url']
	parsed_url = urlparse(url)
	wburl = path.join(
		datetime.now().strftime('%Y/%m/%d/%H/%M'),
		'%s%s' % (parsed_url.netloc, parsed_url.path),
		'index.html',
	)
	urlrewriter = UrlRewriter(wburl)
	liverewriter = LiveRewriter()
	status_headers, buff = liverewriter.get_rewritten(url, urlrewriter)
	S3.put_object(
		Bucket=BUCKET,
		Key=wburl,
		Body=buff,
		ContentType='text/html',
	)
	return {
		'url': '%s/%s' % (ENDPOINT, wburl)
	}