DDPB-2847: Use a maintained PDF generator (#197)

* Port wkhtmltopdf to maintained image

Copied from opg-docker repo, but now using Alpine and Python 3.x

* Switch from Alpine to Debian

The Alpine version doesn't include the QT patch, which is essential for rendering custom fonts.

Debian is the default OS used by `python:3`

* Build wkhtmltopdf image in build process

* Specify version of wkhtmltopdf to use

* Use apt-get rather than apt

* Add cronable script to clean files from /tmp

* Install and run cron on container

We need `service cron restart` to ensure that the cron process starts and acknowledges our new file in `cron.hourly`

* Increase cleanup delay to five minutes

Just to give some breathing extra room for currently-running processes

* Don't expose wkhtmltopdf port

It's not necessary and wasn't the case before

docker-compose.ci.yml

@@ -12,3 +12,6 @@ services:
   test:
     image: ${AWS_REGISTRY}/digideps/test:${VERSION}
     build: ./behat
+  wkhtmltopdf:
+    image: ${AWS_REGISTRY}/digideps/wkhtmltopdf:${VERSION}
+    build: ./wkhtmltopdf

docker-compose.yml

@@ -121,6 +121,12 @@ services:
             - ./api/docker/env/api.env
         restart: always
 
+    wkhtmltopdf:
+        build: ./wkhtmltopdf
+        volumes:
+            - ./wkhtmltopdf/app.py:/app.py
+        restart: always
+
     postgres:
         image: postgres:9.6
         ports:
@@ -144,10 +150,6 @@ services:
         image: redis:2.8.21
         restart: always
 
-    wkhtmltopdf:
-        image: registry.service.opg.digital/opguk/wkhtmlpdf:0.1.209
-        restart: always
-
     file-scanner-server:
         image: mkodockx/docker-clamav
 

environment/ecr.tf

@@ -5,7 +5,7 @@ locals {
     file_scanner = "${data.aws_ecr_repository.images["file-scanner"].repository_url}:${var.OPG_DOCKER_TAG}"
     sync         = "${data.aws_ecr_repository.images["sync"].repository_url}:${var.OPG_DOCKER_TAG}"
     test         = "${data.aws_ecr_repository.images["test"].repository_url}:${var.OPG_DOCKER_TAG}"
-    wkhtmltopdf  = "${data.aws_ecr_repository.images["wkhtmltopdf"].repository_url}:latest"
+    wkhtmltopdf  = "${data.aws_ecr_repository.images["wkhtmltopdf"].repository_url}:${var.OPG_DOCKER_TAG}"
   }
 
   repositories = [

wkhtmltopdf/Dockerfile

@@ -0,0 +1,23 @@
+FROM python:3
+
+# Download and install wkhtmltopdf and bash (because executor can't use 'sh')
+RUN apt-get update \
+    && apt-get -y install cron wget xfonts-base xfonts-75dpi \
+    && wget https://github.com/wkhtmltopdf/wkhtmltopdf/releases/download/0.12.5/wkhtmltox_0.12.5-1.stretch_amd64.deb \
+    && dpkg -i wkhtmltox_0.12.5-1.stretch_amd64.deb
+
+# Install dependencies for running web service
+RUN pip install werkzeug executor gunicorn
+
+ADD docker-entrypoint.sh /docker-entrypoint.sh
+ADD app.py /app.py
+ADD clean-tmp /etc/cron.hourly/clean-tmp
+EXPOSE 80
+
+# Make commands executable
+RUN ["chmod", "+x", "/etc/cron.hourly/clean-tmp"]
+RUN ["chmod", "+x", "/docker-entrypoint.sh"]
+
+ENTRYPOINT ["/docker-entrypoint.sh"]
+
+CMD ["-b", "0.0.0.0:80", "--log-file", "-", "app:application"]

wkhtmltopdf/app.py

@@ -0,0 +1,73 @@
+#! /usr/bin/env python
+"""
+    WSGI APP to convert wkhtmltopdf As a webservice
+
+    :copyright: (c) 2013 by Openlabs Technologies & Consulting (P) Limited
+    :license: BSD, see LICENSE for more details.
+"""
+import base64
+import json
+import tempfile
+
+from werkzeug.wsgi import wrap_file
+from werkzeug.wrappers import Request, Response
+from executor import execute
+
+
+@Request.application
+def application(request):
+    """
+    To use this application, the user must send a POST request with
+    base64 or form encoded encoded HTML content and the wkhtmltopdf Options in
+    request data, with keys 'base64_html' and 'options'.
+    The application will return a response with the PDF file.
+    """
+    if request.method != 'POST':
+        return
+
+    request_is_json = request.content_type.endswith('json')
+
+    with tempfile.NamedTemporaryFile(suffix='.html') as source_file:
+
+        if request_is_json:
+            # If a JSON payload is there, all data is in the payload
+            payload = json.loads(request.data)
+            source_file.write(base64.b64decode(payload['contents']))
+            options = payload.get('options', {})
+        elif request.files:
+            # First check if any files were uploaded
+            source_file.write(request.files['file'].read())
+            # Load any options that may have been provided in options
+            options = json.loads(request.form.get('options', '{}'))
+
+        source_file.flush()
+
+        # Evaluate argument to run with subprocess
+        args = ['wkhtmltopdf']
+
+        # Add Global Options
+        if options:
+            for option, value in options.items():
+                args.append('--%s' % option)
+                if value:
+                    args.append('"%s"' % value)
+
+        # Add source file name and output file name
+        file_name = source_file.name
+        args += [file_name, file_name + ".pdf"]
+
+        # Execute the command using executor
+        execute(' '.join(args))
+
+        return Response(
+            wrap_file(request.environ, open(file_name + '.pdf')),
+            mimetype='application/pdf',
+            direct_passthrough=True,
+        )
+
+
+if __name__ == '__main__':
+    from werkzeug.serving import run_simple
+    run_simple(
+        '127.0.0.1', 5000, application, use_debugger=True, use_reloader=True
+    )

wkhtmltopdf/clean-tmp

@@ -0,0 +1,3 @@
+#!/bin/sh
+
+find /tmp -mmin +5 -type f -delete

wkhtmltopdf/docker-entrypoint.sh

@@ -0,0 +1,4 @@
+#!/bin/sh
+service cron restart
+
+usr/local/bin/gunicorn $@