Flatten comp models when we read them from a string.

source/rrSBMLReader.cpp

@@ -96,9 +96,11 @@ bool SBMLReader::is_sbml(const std::string& str)
 /**
  * extract the <sbml...> tag from an sbml document.
  * This regex should stop at the first match.
+ * RE_DOTALL lets '.' match newlines, since the tag's attributes may be
+ * spread across multiple lines.
  */
 static const Poco::RegularExpression sbml_re("<\\s*sbml\\s*.*?>",
-        RegularExpression::RE_UNGREEDY);
+        RegularExpression::RE_UNGREEDY | RegularExpression::RE_DOTALL);
 
 /**
  * Check if the given sbml std::string uses the composite extension.
@@ -255,6 +257,25 @@ std::string SBMLReader::read(const std::string& str)
 
     if (is_sbml(str))
     {
+        if (has_comp(str))
+        {
+            return flatten_comp(str, "");
+        }
+        if (has_qual(str))
+        {
+            rrLog(Logger::LOG_ERROR) << "Qual model discovered, but not supported.";
+            throw std::domain_error("This SBML model contains information from the 'qual' package (for qualitative or 'logical' modeling).  These models are not supported by roadrunner or tellurium.  A good source of software that supports qual is the COLOMOTO consortium.  See http://www.colomoto.org/software/ for a list of other software packages that might work for you.");
+        }
+        if (has_spatial(str))
+        {
+            rrLog(Logger::LOG_ERROR) << "Spatial model discovered, but not supported.";
+            throw std::domain_error("This SBML model contains information from the 'spatial' package.  These models are not supported by roadrunner or tellurium.  A few software packages that do support spatial (as of 2023) include VCell (https://vcell.org/), XitoSBML (https://github.com/spatialsimulator/XitoSBML) and Spatial SBML (https://github.com/fbergmann/spatial-sbml).");
+        }
+        if (has_multi(str))
+        {
+            rrLog(Logger::LOG_ERROR) << "Multi model discovered, but not supported.";
+            throw std::domain_error("This SBML model contains information from the 'multi' package for multistate, multicomponent and multicompartment models, or 'rule-based' models.  These models are not supported by roadrunner or tellurium.  The best software package that supports multi (as of 2023) is Simmune (https://bioinformatics.niaid.nih.gov/simmune/).");
+        }
         return str;
     }
 

source/rrSBMLReader.h

@@ -36,13 +36,9 @@ class RR_DECLSPEC SBMLReader
 
     /**
      * read an SBML document from a local file path, a remote URI, or
-     * directlly from a sbml std::string. If the document is a local file,
-     * and has the comp extension, it is automatically flattened before
-     * begin returned.
-     *
-     * If the std::string is already a SBML std::string, it is passed through.
-     * If the std::string is a local file or URI, the document is read
-     * from the source and the contents returned (and flattened if comp).
+     * directly from a sbml std::string. If the document has the comp
+     * extension, it is automatically flattened before being returned,
+     * regardless of whether it came from a file, a URI, or a std::string.
      */
     static std::string read(const std::string& sbml_or_uri);
 

test/models/SBMLFeatures/comp_example.xml

@@ -0,0 +1,41 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<sbml xmlns="http://www.sbml.org/sbml/level3/version1/core"
+      xmlns:comp="http://www.sbml.org/sbml/level3/version1/comp/version1"
+      level="3" version="1" comp:required="true">
+  <model id="main_model">
+    <comp:listOfSubmodels>
+      <comp:submodel comp:id="sub1" comp:modelRef="submodel_def"/>
+    </comp:listOfSubmodels>
+  </model>
+  <comp:listOfModelDefinitions>
+    <comp:modelDefinition id="submodel_def">
+      <listOfCompartments>
+        <compartment id="C" size="1" constant="true"/>
+      </listOfCompartments>
+      <listOfSpecies>
+        <species id="S1" compartment="C" initialConcentration="1" hasOnlySubstanceUnits="false" boundaryCondition="false" constant="false"/>
+        <species id="S2" compartment="C" initialConcentration="0" hasOnlySubstanceUnits="false" boundaryCondition="false" constant="false"/>
+      </listOfSpecies>
+      <listOfReactions>
+        <reaction id="J0" reversible="false" fast="false">
+          <listOfReactants>
+            <speciesReference species="S1" stoichiometry="1" constant="true"/>
+          </listOfReactants>
+          <listOfProducts>
+            <speciesReference species="S2" stoichiometry="1" constant="true"/>
+          </listOfProducts>
+          <kineticLaw>
+            <math xmlns="http://www.w3.org/1998/Math/MathML">
+              <apply>
+                <times/>
+                <ci>C</ci>
+                <cn type="real">0.1</cn>
+                <ci>S1</ci>
+              </apply>
+            </math>
+          </kineticLaw>
+        </reaction>
+      </listOfReactions>
+    </comp:modelDefinition>
+  </comp:listOfModelDefinitions>
+</sbml>

test/sbml_features/CMakeLists.txt

@@ -3,6 +3,7 @@ add_test_executable(
         ${SharedTestFiles}
         named_stoich.cpp
         variable_stoich.cpp
+        unsupported_packages.cpp
 )
 
 # make test_targets list global to all tests

test/sbml_features/unsupported_packages.cpp

@@ -1,12 +1,15 @@
 #include "gtest/gtest.h"
 #include "rrRoadRunner.h"
+#include "rrSBMLReader.h"
 #include "rrException.h"
 #include "rrUtils.h"
 #include "rrTestSuiteModelSimulation.h"
 #include "sbml/SBMLTypes.h"
 #include "sbml/SBMLReader.h"
 #include "../test_util.h"
 #include <filesystem>
+#include <fstream>
+#include <sstream>
 #include "RoadRunnerTest.h"
 
 using namespace testing;
@@ -18,6 +21,13 @@ class SBMLFeatures : public RoadRunnerTest {
 public:
     path SBMLFeaturesDir = rrTestModelsDir_ / "SBMLFeatures";
     SBMLFeatures() = default;
+
+    static std::string readFileToString(const path& p) {
+        std::ifstream in(p);
+        std::stringstream ss;
+        ss << in.rdbuf();
+        return ss.str();
+    }
 };
 
 TEST_F(SBMLFeatures, SBML_qual)
@@ -38,6 +48,64 @@ TEST_F(SBMLFeatures, SBML_multi)
     EXPECT_THROW(rri.load((SBMLFeaturesDir / "simmune_Ecad.xml").string()), std::domain_error);
 }
 
+// A comp model read from a file path is flattened before being returned.
+TEST_F(SBMLFeatures, SBML_comp)
+{
+    std::string flattened = SBMLReader::read((SBMLFeaturesDir / "comp_example.xml").string());
+    EXPECT_EQ(flattened.find("level3/version1/comp"), std::string::npos);
+    EXPECT_NE(flattened.find("sub1__S1"), std::string::npos);
+}
+
+// A comp model read directly from a std::string (not a file path) is also flattened.
+TEST_F(SBMLFeatures, SBML_comp_fromString)
+{
+    std::string sbml = readFileToString(SBMLFeaturesDir / "comp_example.xml");
+    std::string flattened = SBMLReader::read(sbml);
+    EXPECT_EQ(flattened.find("level3/version1/comp"), std::string::npos);
+    EXPECT_NE(flattened.find("sub1__S1"), std::string::npos);
+}
+
+// qual/spatial/multi models read directly from a std::string are rejected,
+// just like when read from a file path (see SBML_qual/SBML_spatial/SBML_multi above).
+TEST_F(SBMLFeatures, SBML_qual_fromString)
+{
+    std::string sbml = readFileToString(SBMLFeaturesDir / "BIOMD0000000562_url.xml");
+    try {
+        SBMLReader::read(sbml);
+        FAIL() << "Expected std::domain_error to be thrown for a qual model read from a string";
+    } catch (const std::domain_error& e) {
+        std::string msg = e.what();
+        EXPECT_NE(msg.find("qual"), std::string::npos);
+        EXPECT_NE(msg.find("COLOMOTO"), std::string::npos);
+    }
+}
+
+TEST_F(SBMLFeatures, SBML_spatial_fromString)
+{
+    std::string sbml = readFileToString(SBMLFeaturesDir / "organelles.xml");
+    try {
+        SBMLReader::read(sbml);
+        FAIL() << "Expected std::domain_error to be thrown for a spatial model read from a string";
+    } catch (const std::domain_error& e) {
+        std::string msg = e.what();
+        EXPECT_NE(msg.find("spatial"), std::string::npos);
+        EXPECT_NE(msg.find("VCell"), std::string::npos);
+    }
+}
+
+TEST_F(SBMLFeatures, SBML_multi_fromString)
+{
+    std::string sbml = readFileToString(SBMLFeaturesDir / "simmune_Ecad.xml");
+    try {
+        SBMLReader::read(sbml);
+        FAIL() << "Expected std::domain_error to be thrown for a multi model read from a string";
+    } catch (const std::domain_error& e) {
+        std::string msg = e.what();
+        EXPECT_NE(msg.find("multi"), std::string::npos);
+        EXPECT_NE(msg.find("Simmune"), std::string::npos);
+    }
+}
+
 TEST_F(SBMLFeatures, SBML_fbc)
 {
     RoadRunner rri((SBMLFeaturesDir / "fbc_example.xml").string());