Merge pull request #696 from Pchelolo/pdf

Added PDF endpoint

config.example.wikimedia.yaml

@@ -33,6 +33,11 @@ default_project: &default_project
           cache_control: s-maxage=86400, max-age=86400
         # 10 days Varnish caching, one day client-side
         purged_cache_control: s-maxage=864000, max-age=86400
+        pdf:
+          # Cache PDF for 5 minutes since it's not purged
+          cache_control: s-maxage=600, max-age=600
+          uri: https://pdf-electron.wmflabs.org
+          secret: secret
         skip_updates: false
 
 # A different project template, sharing configuration options.

config.test.yaml

@@ -38,6 +38,11 @@ default_project: &default_project
           host: http://appservice.wmflabs.org
         events: {}
         purged_cache_control: test_purged_cache_control
+        pdf:
+          # Cache PDF for 5 minutes since it's not purged
+          cache_control: s-maxage=600, max-age=600
+          uri: https://pdf-electron.wmflabs.org
+          secret: secret
         skip_updates: false
 
 labs_project: &labs_project

projects/wmf_default.yaml

@@ -54,6 +54,8 @@ paths:
                 - path: v1/random.yaml
                   options: '{{merge({"random_cache_control": "s-maxage=2, max-age=1"},
                                 options.mobileapps)}}'
+                - path: v1/pdf.yaml
+                  options: '{{options.pdf}}'
             /feed:
               x-modules:
                 - path: v1/feed.js

projects/wmf_wiktionary.yaml

@@ -50,6 +50,8 @@ paths:
                   options:
                     response_cache_control: '{{options.purged_cache_control}}'
                     host: '{{options.mobileapps.host}}'
+                - path: v1/pdf.yaml
+                  options: '{{options.pdf}}'
             /transform:
               x-modules:
                 - path: v1/transform.yaml

test/features/feed.js

@@ -3,7 +3,6 @@
 const assert = require('../utils/assert.js');
 const server = require('../utils/server.js');
 const preq   = require('preq');
-const P = require('bluebird');
 
 function assertStorageRequest(requests, method, bucket, expected) {
     const storageRequests = requests.filter((log) =>

test/features/pdf.js

@@ -0,0 +1,25 @@
+'use strict';
+
+const assert = require('../utils/assert.js');
+const server = require('../utils/server.js');
+const preq   = require('preq');
+
+describe('Feed', () => {
+
+    before(() => server.start());
+
+    it('Should get PDF for a page', () => {
+        return preq.get({
+            uri: `${server.config.hostPort}/ru.wikipedia.org/v1/page/pdf/%D0%94%D0%B0%D1%80%D1%82_%D0%92%D0%B5%D0%B9%D0%B4%D0%B5%D1%80`
+        })
+        .then((res) => {
+            assert.deepEqual(res.status, 200);
+            assert.deepEqual(res.headers['content-disposition'],
+                'attachment; filename=%D0%94%D0%B0%D1%80%D1%82_%D0%92%D0%B5%D0%B9%D0%B4%D0%B5%D1%80.pdf;'
+                    + ' filename*=%D0%94%D0%B0%D1%80%D1%82_%D0%92%D0%B5%D0%B9%D0%B4%D0%B5%D1%80.pdf');
+            assert.deepEqual(res.headers['content-type'], 'application/pdf');
+            assert.ok(/"\d+\/[\d\w-]+"/.test(res.headers.etag));
+            assert.ok(res.body.length !== 0);
+        });
+    });
+});

v1/pdf.yaml

@@ -0,0 +1,89 @@
+swagger: '2.0'
+info:
+  version: '1.0.0-beta'
+  title: MediaWiki PDF API
+  description: Page PDF Render API
+  termsOfService: https://github.com/wikimedia/restbase#restbase
+  contact:
+    name: Services
+    email: services@lists.wikimedia.org
+    url: https://www.mediawiki.org/wiki/Services
+  license:
+    name: Apache licence, v2
+    url: https://www.apache.org/licenses/LICENSE-2.0
+paths:
+  /pdf/{title}:
+    x-route-filters:
+      - path: ./lib/revision_table_access_check_filter.js
+        options:
+          redirect_cache_control: '{{options.cache_control}}'
+    get:
+      tags:
+        - Page content
+      summary: Get a page as PDF
+      description: |
+        Renders the page content as PDF.
+
+        Stability: [experimental](https://www.mediawiki.org/wiki/API_versioning#Experimental)
+      produces:
+        - application/pdf
+      parameters:
+        - name: title
+          in: path
+          description: "Page title. Use underscores instead of spaces. Example: `Main_Page`."
+          type: string
+          required: true
+      responses:
+        '200':
+          description: The PDF render of an article
+          schema:
+            type: file
+          headers:
+            ETag:
+              description: >
+                Syntax: "{revision}/{tid}".
+                Example: "701384379/154d7bca-c264-11e5-8c2f-1b51b33b59fc"
+        '404':
+          description: Unknown page title
+          schema:
+            $ref: '#/definitions/problem'
+        default:
+          description: Error
+          schema:
+            $ref: '#/definitions/problem'
+
+      x-request-handler:
+        - get_latest_revision:
+            request:
+              method: get
+              uri: /{domain}/sys/page_revisions/page/{title}
+        - get_pdf_from_backend:
+            request:
+              method: get
+              uri: '{{options.uri}}/pdf?accessKey={options.secret}&url=https://{{domain}}/wiki/{title}'
+            return:
+              status: 200
+              headers:
+                # Firefox supports filename*= syntax while Chrome respect filename=. Safari is stupid and understands neither
+                content-disposition: 'attachment; filename={request.params.title}.pdf; filename*={request.params.title}.pdf'
+                content-type: '{{get_pdf_from_backend.headers.content-type}}'
+                content-length: '{{get_pdf_from_backend.headers.content-length}}'
+                cache-control: '{{default(options.cache_control, "s-maxage=600, max-age=600")}}'
+                etag: '{{get_latest_revision.headers.etag}}'
+              body: '{{get_pdf_from_backend.body}}'
+      x-monitor: false # PDF generation is expensive and it's not stored, so don't run checker script
+
+definitions:
+  # A https://tools.ietf.org/html/draft-nottingham-http-problem
+  problem:
+    required:
+      - type
+    properties:
+      type:
+        type: string
+      title:
+        type: string
+      detail:
+        type: string
+      instance:
+        type: string