fix: flatten choice struct members within sequences

The PubChem PUG View schema[1] frequently uses the following pattern:

```xml
<!-- ... -->
    <xs:complexType>
      <xs:sequence>
        <xs:choice>
          <xs:element name="A" type="xs:string"/>
          <xs:element name="B" type="xs:int"/>
        </xs:choice>
        <xs:element name="..." type="xs:string" minOccurs="0"/>
      </xs:sequence>
    </xs:complexType>
  </xs:element>
<!-- ... -->
```

Prior to this change the `xs:choice` element was being treated as just
another xsd element rather than a choice. This commit correctly tags enum
struct fields so that they are properly flattened.

[1]: https://pubchem.ncbi.nlm.nih.gov/docs/pug-view#section=Formats

Author: adrienthebo

xsd-parser/src/parser/utils.rs

@@ -73,7 +73,7 @@ pub fn enum_to_field(en: Enum) -> StructField {
         name: en.name.clone(),
         type_name: en.name.clone(),
         subtypes: vec![RsEntity::Enum(en)],
-        source: StructFieldSource::Element,
+        source: StructFieldSource::Choice,
         ..Default::default()
     }
 }

xsd-parser/tests/mod.rs

@@ -18,3 +18,4 @@ mod tuple_with_vec;
 mod type_name_clash;
 mod union;
 mod xsd_string;
+mod sequence_choice;

xsd-parser/tests/sequence_choice/example.xml

@@ -0,0 +1,36 @@
+<Section
+    xmlns:xs="http://www.w3.org/2001/XMLSchema-instance"
+    xs:schemaLocation="http://pubchem.ncbi.nlm.nih.gov/pug_view https://pubchem.ncbi.nlm.nih.gov/pug_view/pug_view.xsd"
+>
+  <TOCHeading>Structures</TOCHeading>
+  <Description>Structure depictions of this compound, including computationally generated two-dimensional (2D) and three-dimensional (3D) structures, as well as experimentally determined 3D single-crystal structures.</Description>
+  <Section>
+    <TOCHeading>2D Structure</TOCHeading>
+    <Description>A two-dimensional (2D) structure representation of the compound.  Because this structure is processed through chemical structure standardization (Hähnke et al., J. Cheminform. 2018, 10, 36), it is not necessarily the same as the structures provided by individual data contributors.  </Description>
+    <URL>https://doi.org/10.1186/s13321-018-0293-8</URL>
+    <DisplayControls>
+      <MoveToTop>true</MoveToTop>
+    </DisplayControls>
+    <Information>
+      <ReferenceNumber>66</ReferenceNumber>
+      <Value>
+        <Boolean>true</Boolean>
+      </Value>
+    </Information>
+  </Section>
+  <Section>
+    <TOCHeading>3D Conformer</TOCHeading>
+    <Description>A three-dimensional (3D) structure representation of the compound.  This 3D structure is not experimentally determined, but computed by PubChem.  This structure may or may not be the same as the inherent structure of the compound you would expect to see in vacuum or in the gas phase, because the underlying computational algorithm aims to generate a protein-bound structure, which would be observed in a protein-ligand complex.  More detailed information on this conformer model can be found in Kim et al., J. Cheminform. 2013, 5, 1.</Description>
+    <URL>https://doi.org/10.1186/1758-2946-5-1</URL>
+    <DisplayControls>
+      <MoveToTop>true</MoveToTop>
+    </DisplayControls>
+    <Information>
+      <ReferenceNumber>66</ReferenceNumber>
+      <Description>2-Hydroxyethyl Methacrylate</Description>
+      <Value>
+        <Number>13360</Number>
+      </Value>
+    </Information>
+  </Section>
+</Section>

xsd-parser/tests/sequence_choice/expected.rs

@@ -0,0 +1,37 @@
+#[derive(Default, PartialEq, Debug, YaSerialize, YaDeserialize)]
+#[yaserde(namespace = "http://pubchem.ncbi.nlm.nih.gov/pug_view")]
+pub struct Section {
+    #[yaserde(flatten)]
+    pub section_choice: section::SectionChoice,
+
+    #[yaserde(rename = "Description")]
+    pub description: Option<String>,
+
+    #[yaserde(rename = "URL")]
+    pub url: Option<String>,
+}
+
+impl Validate for Section {}
+
+pub mod section {
+    use super::*;
+
+    #[derive(PartialEq, Debug, YaSerialize, YaDeserialize)]
+    #[yaserde(namespace = "http://pubchem.ncbi.nlm.nih.gov/pug_view")]
+
+    pub enum SectionChoice {
+        #[yaserde(rename = "TOCHeading")]
+        Tocheading(String),
+        #[yaserde(rename = "TOCID")]
+        Tocid(i32),
+        __Unknown__(String),
+    }
+
+    impl Default for SectionChoice {
+        fn default() -> SectionChoice {
+            Self::__Unknown__("No valid variants".into())
+        }
+    }
+
+    impl Validate for SectionChoice {}
+}

xsd-parser/tests/sequence_choice/input.xsd

@@ -0,0 +1,20 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<xs:schema
+  xmlns:xs="http://www.w3.org/2001/XMLSchema"
+  xmlns="http://pubchem.ncbi.nlm.nih.gov/pug_view"
+  targetNamespace="http://pubchem.ncbi.nlm.nih.gov/pug_view"
+  elementFormDefault="qualified"
+  attributeFormDefault="unqualified">
+  <xs:element name="Section">
+    <xs:complexType>
+      <xs:sequence>
+        <xs:choice>
+          <xs:element name="TOCHeading" type="xs:string"/>
+          <xs:element name="TOCID" type="xs:int"/>
+        </xs:choice>
+        <xs:element name="Description" type="xs:string" minOccurs="0"/> 
+        <xs:element name="URL" type="xs:string" minOccurs="0"/> 
+      </xs:sequence>
+    </xs:complexType>
+  </xs:element>
+</xs:schema>

xsd-parser/tests/sequence_choice/mod.rs

@@ -0,0 +1,31 @@
+use super::utils;
+
+#[test]
+fn deserialization_works() {
+    mod expected {
+        use xsd_parser::generator::validator::Validate;
+        use yaserde_derive::{YaDeserialize, YaSerialize};
+
+        include!("expected.rs");
+    }
+
+    let ser = include_str!("example.xml");
+
+    let de: expected::Section = yaserde::de::from_str(ser).unwrap();
+
+    assert_eq!(de, expected::Section { 
+        section_choice: expected::section::SectionChoice::Tocheading("2D Structure".into()),
+        description: Some("Structure depictions of this compound, including computationally generated two-dimensional (2D) and three-dimensional (3D) structures, as well as experimentally determined 3D single-crystal structures.".into()),
+        url: None,
+    });
+}
+
+#[test]
+fn generator_does_not_panic() {
+    println!("{}", utils::generate(include_str!("input.xsd")))
+}
+
+#[test]
+fn generator_output_has_correct_ast() {
+    utils::ast_test(include_str!("input.xsd"), include_str!("expected.rs"));
+}