Merge e07a770 into 9522e9f

cnxpublishing/db.py

@@ -35,8 +35,11 @@
     ResourceFileExceededLimitError,
     UserFetchError,
     )
-from .utils import parse_archive_uri, parse_user_uri
+from .models import inject_mathml_svgs
 from .publish import publish_model, republish_binders
+from .utils import (
+    parse_archive_uri, parse_user_uri,
+    )
 
 
 __all__ = (
@@ -582,6 +585,9 @@ def mark_invalid_reference(reference):
                     mark_invalid_reference(reference)
             # else, it's a remote or cnx.org reference ...Do nothing.
 
+        # Generate SVGs for MathML
+        model.content = inject_mathml_svgs(model.content)
+
         args = (psycopg2.Binary(model.content.encode('utf-8')),
                 publication_id, model.id,)
         stmt = """\

cnxpublishing/models.py

@@ -0,0 +1,69 @@
+# -*- coding: utf-8 -*-
+# ###
+# Copyright (c) 2015, Rice University
+# This software is subject to the provisions of the GNU Affero General
+# Public License version 3 (AGPLv3).
+# See LICENCE.txt for details.
+# ###
+import requests
+from lxml import etree
+from pyramid.settings import asbool
+from pyramid.threadlocal import get_current_registry
+
+
+MATHML_NAMESPACE = "http://www.w3.org/1998/Math/MathML"
+NSMAP = {'m': MATHML_NAMESPACE}
+
+
+def _find_or_create_semantics_block(math_block):
+    """Finds or creates the MathML semantics tag."""
+    try:
+        semantics_block = math_block.xpath(
+            '/m:samantics',
+            namespaces=NSMAP)[0]
+    except IndexError:
+        # Move all the contents of the math_block into a semantics wrapper.
+        children = []
+        for child in math_block.getchildren():
+            children.append(child)
+            math_block.remove(child)  # why no pop?
+        semantics_block = etree.SubElement(
+            math_block,
+            '{{{}}}semantics'.format(MATHML_NAMESPACE))
+        for child in children:
+            semantics_block.append(child)
+    return semantics_block
+
+
+def inject_mathml_svgs(content):
+    """Inject MathML SVG annotations into HTML content."""
+    settings = get_current_registry().settings
+    is_enabled = asbool(settings.get('mathml2svg.enabled?', False))
+    url = settings.get('mathml2svg.url')
+
+    # Bailout when svg generation is disabled.
+    if not is_enabled:
+        return content
+
+    xml = etree.fromstring(content)
+    math_blocks = xml.xpath(
+        '//m:math[not(/m:annotation-xml[@encoding="image/svg+xml"])]',
+        namespaces=NSMAP)
+    for math_block in math_blocks:
+        # Submit the MathML block to the SVG generation service.
+        payload = {'MathML': etree.tostring(math_block)}
+        response = requests.post(url, data=payload)
+        # Inject the SVG into the MathML as an annotation
+        # only if the resposne was good, otherwise skip over it.
+        if response.status_code == 200:
+            semantics_wrapper = _find_or_create_semantics_block(math_block)
+            svg = response.text
+            content_type = response.headers['content-type']
+            # Insert the svg into the content
+            annotation = etree.SubElement(
+                semantics_wrapper,
+                '{{{}}}annotation-xml'.format(MATHML_NAMESPACE))
+            annotation.set('encoding', content_type)
+            annotation.append(etree.fromstring(svg))
+    modified_content = etree.tostring(xml)
+    return modified_content

cnxpublishing/tests/test_db.py

@@ -22,6 +22,7 @@
 from cnxarchive import config as archive_config
 from cnxarchive.database import initdb as archive_initdb
 from cnxarchive.utils import join_ident_hash, split_ident_hash
+from lxml import etree
 from pyramid import testing
 
 from . import use_cases
@@ -1137,6 +1138,59 @@ def test_add_pending_document_w_invalid_references(self, cursor):
         self.assertEqual(len(state_messages), 2)
         self.assertEqual(state_messages[-1], expected_state_message)
 
+    @db_connect
+    def test_add_pending_document_w_mathml(self, cursor):
+        """Add a pending document with mathml that generates SVG."""
+        publication_id = self.make_publication()
+
+        # Insert a valid module for referencing...
+        cursor.execute("""\
+INSERT INTO abstracts (abstract) VALUES ('abstract')
+RETURNING abstractid""")
+        cursor.execute("""\
+INSERT INTO modules
+(module_ident, portal_type, name,
+ created, revised, abstractid, licenseid,
+ doctype, submitter, submitlog, stateid, parent, parentauthors,
+ language, authors, maintainers, licensors,
+ google_analytics, buylink)
+VALUES
+(1, 'Module', 'mathml module',
+ DEFAULT, DEFAULT, 1, 1,
+ 0, 'admin', 'log', NULL, NULL, NULL,
+ 'en', '{admin}', NULL, '{admin}',
+ DEFAULT, DEFAULT) RETURNING uuid || '@' || major_version""")
+        doc_ident_hash = cursor.fetchone()[0]
+
+        # Create and add a document for the publication.
+        metadata = {
+            'title': 'Document Title',
+            'summary': 'Document Summary',
+            'authors': [{u'id': u'able', u'type': u'cnx-id'}],
+            'publishers': [{'id': 'able', 'type': 'cnx-id'}],
+            'license_url': VALID_LICENSE_URL,
+            }
+        content = """\
+<div class="equation">
+<math xmlns="http://www.w3.org/1998/Math/MathML"><semantics><mrow> <mi>x</mi> <mo>=</mo> <mfrac> <mrow> <mo>&#8722;<!-- &#8722; --></mo> <mi>b</mi> <mo>&#177;<!-- &#177; --></mo> <msqrt> <msup> <mi>b</mi> <mn>2</mn> </msup> <mo>&#8722;<!-- &#8722; --></mo> <mn>4</mn> <mi>a</mi> <mi>c</mi> </msqrt> </mrow> <mrow> <mn>2</mn> <mi>a</mi> </mrow> </mfrac> </mrow></semantics></math>
+</div>"""
+        document = self.make_document(content=content, metadata=metadata)
+
+        # Here we are testing the function of add_pending_document.
+        from ..db import add_pending_model, add_pending_model_content
+        document_ident_hash = add_pending_model(
+            cursor, publication_id, document)
+
+        # Enable the mathml2svg service for this test.
+        from pyramid.threadlocal import get_current_registry
+        get_current_registry().settings['mathml2svg.enabled?'] = 'on'
+
+        # Mock the inject_mathml_svgs function.
+        with mock.patch('cnxpublishing.db.inject_mathml_svgs') as func:
+            func.return_value = content
+            add_pending_model_content(cursor, publication_id, document)
+            self.assertEqual(func.call_count, 1)
+
     @db_connect
     def test_add_pending_binder_w_document_pointers(self, cursor):
         """Add a pending binder with document pointers."""

cnxpublishing/tests/test_models.py

@@ -0,0 +1,79 @@
+# -*- coding: utf-8 -*-
+# ###
+# Copyright (c) 2015, Rice University
+# This software is subject to the provisions of the GNU Affero General
+# Public License version 3 (AGPLv3).
+# See LICENCE.txt for details.
+# ###
+import unittest
+try:
+    from unittest import mock
+except ImportError:
+    import mock
+
+from lxml import etree
+
+
+class SVGInjectionUtilTestCase(unittest.TestCase):
+
+    def setUp(self):
+        # Mock the registry, since this is a utility test.
+        self.registry_patcher = mock.patch(
+            'cnxpublishing.models.get_current_registry')
+        self.mock_registry = self.registry_patcher.start()
+        self.mock_registry.return_value.settings = {
+            'mathml2svg.enabled?': 'on',
+            }
+
+        # Mock the external communication with the cnx-mathml2svg service.
+        self.requests_patcher = mock.patch(
+            'requests.post')
+        self.mocked_post = self.requests_patcher.start()
+        self.mocked_post.return_value.status_code = 200
+        self.mocked_post.return_value.headers = {
+            'content-type': 'image/svg+xml',
+            }
+        self.mocked_post.return_value.text = '<svg>mocked</svg>'
+
+    def tearDown(self):
+        self.registry_patcher.stop()
+        self.requests_patcher.stop()
+
+    @property
+    def target(self):
+        from ..models import inject_mathml_svgs
+        return inject_mathml_svgs
+
+    def test_mathml2svg(self):
+        content = """\
+<div class="equation">
+<math xmlns="http://www.w3.org/1998/Math/MathML"><semantics><mrow> <mi>x</mi> <mo>=</mo> <mfrac> <mrow> <mo>&#8722;<!-- &#8722; --></mo> <mi>b</mi> <mo>&#177;<!-- &#177; --></mo> <msqrt> <msup> <mi>b</mi> <mn>2</mn> </msup> <mo>&#8722;<!-- &#8722; --></mo> <mn>4</mn> <mi>a</mi> <mi>c</mi> </msqrt> </mrow> <mrow> <mn>2</mn> <mi>a</mi> </mrow> </mfrac> </mrow></semantics></math>
+</div>"""
+
+        # Call the target function.
+        result = self.target(content)
+
+        elms = etree.fromstring(result)
+        annotation = elms.xpath(
+            '/div/m:math//m:annotation-xml[@encoding="image/svg+xml"]',
+            namespaces={'m': "http://www.w3.org/1998/Math/MathML"})[0]
+        expected = """<annotation-xml xmlns="http://www.w3.org/1998/Math/MathML" encoding="image/svg+xml"><svg>mocked</svg></annotation-xml>"""
+
+        self.assertEqual(etree.tostring(annotation), expected)
+
+    def test_missing_semantics_wrapper(self):
+        content = """\
+<div class="equation">
+<math xmlns="http://www.w3.org/1998/Math/MathML"><mrow> <mi>x</mi> <mo>=</mo> <mfrac> <mrow> <mo>&#8722;<!-- &#8722; --></mo> <mi>b</mi> <mo>&#177;<!-- &#177; --></mo> <msqrt> <msup> <mi>b</mi> <mn>2</mn> </msup> <mo>&#8722;<!-- &#8722; --></mo> <mn>4</mn> <mi>a</mi> <mi>c</mi> </msqrt> </mrow> <mrow> <mn>2</mn> <mi>a</mi> </mrow> </mfrac> </mrow></math>
+</div>"""
+
+        # Call the target function.
+        result = self.target(content)
+
+        elms = etree.fromstring(result)
+        annotation = elms.xpath(
+            '/div/m:math/m:semantics/m:annotation-xml[@encoding="image/svg+xml"]',
+            namespaces={'m': "http://www.w3.org/1998/Math/MathML"})[0]
+        expected = """<annotation-xml xmlns="http://www.w3.org/1998/Math/MathML" encoding="image/svg+xml"><svg>mocked</svg></annotation-xml>"""
+
+        self.assertEqual(etree.tostring(annotation), expected)

development.ini

@@ -21,6 +21,11 @@ db-connection-string = dbname=cnxarchive user=cnxarchive password=cnxarchive
 # size limit of file uploads in MB
 file-upload-limit = 50
 
+# mathml2svg is disabled by default.
+# To enable it use ``mathml2svg.enabled? = yes`` and set the url.
+mathml2svg.url = http://localhost:5689
+mathml2svg.enabled? = no
+
 session_key = 'somkindaseekret'
 # Application API keys with authentication information.
 # This information is organized in the following form:

setup.py

@@ -14,6 +14,7 @@
     'pyramid>=1.5',
     'pyramid_jinja2',
     'pyramid_multiauth',
+    'requests',
     )
 tests_require = [
     'webtest',

testing.ini

@@ -29,6 +29,9 @@ openstax_accounts.login_path = /login
 openstax_accounts.callback_path = /callback
 openstax_accounts.logout_path = /logout
 
+mathml2svg.url = http://localhost:5689
+mathml2svg.enabled? = no
+
 [server:main]
 use = egg:waitress#main
 host = 0.0.0.0