From 4b01408393c70f7c1372cf277ea64dc65c0787ea Mon Sep 17 00:00:00 2001
From: Bilal Shaikh <bilalshaikh42@gmail.com>
Date: Wed, 13 Apr 2022 15:54:12 -0400
Subject: [PATCH] fix(cellml): tweak cellml 1.1 validation to allow RDF and
 numbered ids

---
 biosimulators_utils/model_lang/cellml/cellml_1_1.xsd      | 5 ++++-
 .../model_lang/cellml/common/common-attribs.xsd           | 2 +-
 biosimulators_utils/model_lang/cellml/validation.py       | 8 ++++++--
 3 files changed, 11 insertions(+), 4 deletions(-)

diff --git a/biosimulators_utils/model_lang/cellml/cellml_1_1.xsd b/biosimulators_utils/model_lang/cellml/cellml_1_1.xsd
index 3ade21ef..b3f124e0 100644
--- a/biosimulators_utils/model_lang/cellml/cellml_1_1.xsd
+++ b/biosimulators_utils/model_lang/cellml/cellml_1_1.xsd
@@ -2,9 +2,11 @@
 <schema xmlns="http://www.w3.org/2001/XMLSchema"
         xmlns:cellml="http://www.cellml.org/cellml/1.1#"
         xmlns:mathml="http://www.w3.org/1998/Math/MathML"
+        xmlns:rdf= "http://www.w3.org/1999/02/22-rdf-syntax-ns#"
         xmlns:xlink="http://www.w3.org/1999/xlink"
         elementFormDefault="qualified"
         targetNamespace="http://www.cellml.org/cellml/1.1#">
+ <import namespace ="http://www.w3.org/1999/02/22-rdf-syntax-ns#"/>
  <import namespace="http://www.w3.org/1998/Math/MathML"
          schemaLocation="mathml2.xsd"/>
  <import namespace="http://www.w3.org/1999/xlink" schemaLocation="common/xlink-href.xsd"/>
@@ -109,7 +111,8 @@
       <element name="units" type="cellml:units"/>
       <element name="variable" type="cellml:variable"/>
       <element name="reaction" type="cellml:reaction"/>
-      <element ref="mathml:math"/>
+      <element ref="mathml:math"/>      
+      <any namespace="http://www.w3.org/1999/02/22-rdf-syntax-ns#"  processContents="skip" />
     </choice>
     <attribute name="name" type="cellml:cellMLIdentifier" use="required"/>
     <anyAttribute namespace="##other" processContents="lax"/>
diff --git a/biosimulators_utils/model_lang/cellml/common/common-attribs.xsd b/biosimulators_utils/model_lang/cellml/common/common-attribs.xsd
index 33698b8f..2a3011f6 100644
--- a/biosimulators_utils/model_lang/cellml/common/common-attribs.xsd
+++ b/biosimulators_utils/model_lang/cellml/common/common-attribs.xsd
@@ -26,7 +26,7 @@
   <xs:attribute name="class" type="xs:NMTOKENS"/>
   <xs:attribute name="style" type="xs:string"/>
   <xs:attribute name="xref" type="xs:IDREF"/>
-  <xs:attribute name="id" type="xs:ID"/>
+  <xs:attribute name="id" type="xs:string"/>
   <xs:attribute ref="xlink:href"/>
   <!-- allow attributes from foreign namespaces, and don't check them -->
   <xs:anyAttribute namespace="##other" processContents="skip"/>
diff --git a/biosimulators_utils/model_lang/cellml/validation.py b/biosimulators_utils/model_lang/cellml/validation.py
index e1e96135..bac5045f 100644
--- a/biosimulators_utils/model_lang/cellml/validation.py
+++ b/biosimulators_utils/model_lang/cellml/validation.py
@@ -142,13 +142,17 @@ def validate_model_version_1_1(filename, doc, resolve_imports=True, config=None)
     """
     # ``cellml_1_1_original.xsd`` was obtained from https://www.cellml.org/tools/cellml_1_1_schema
     # `cellml_1_1.xsd`` was created by were directly downloading all referenced namspaces from
-    # https://www.cellml.org/tools/cellml_1_1_schema/common, 
-    # https://www.cellml.org/tools/cellml_1_1_schema/content, 
+    # https://www.cellml.org/tools/cellml_1_1_schema/common,
+    # https://www.cellml.org/tools/cellml_1_1_schema/content,
     # https://www.cellml.org/tools/cellml_1_1_schema/presentation
     # The original ``cellml_1_1.xsd`` was modified to change the http sources to local sources.
     # Additionally, the following two lines were removed from the ``cellml_1_1_original.xsd``: 113, 309
     # These lines added the following to the definitions of "Component" and "Role":  <any namespace="##other" processContents="skip"/>
     # This caused lxml to fail to parse the schema due to it being nondeterministic.
+    # In its place, <any namespace="http://www.w3.org/1999/02/22-rdf-syntax-ns#"  processContents="skip" />
+    # was added to allow using RDF syntax in the schema. The RDF is not validated.
+    # Finally, in common/common-attribs.xsd the type of the attribute "id" was changed from "ID" to "string".
+    # This was done to allow using numbers as ids for equations as described in the CellML primer.
 
     schema_filename = pkg_resources.resource_filename('biosimulators_utils',
                                                       os.path.join("model_lang", "cellml", "cellml_1_1.xsd"))