oxml_core/
core_properties.rs

1//! Dublin Core metadata from `docProps/core.xml`.
2
3use quick_xml::Reader;
4use quick_xml::events::Event;
5
6use crate::error::Result;
7use crate::xml::local_name;
8
9/// Document metadata from `docProps/core.xml` (Dublin Core).
10#[derive(Debug, Clone, Default, PartialEq)]
11pub struct CoreProperties {
12    /// Document title (`dc:title`).
13    pub title: Option<String>,
14    /// Document creator/author (`dc:creator`).
15    pub creator: Option<String>,
16    /// Subject (`dc:subject`).
17    pub subject: Option<String>,
18    /// Description/comments (`dc:description`).
19    pub description: Option<String>,
20    /// Keywords (`cp:keywords`).
21    pub keywords: Option<String>,
22    /// Last modified by (`cp:lastModifiedBy`).
23    pub last_modified_by: Option<String>,
24    /// Date created (`dcterms:created`).
25    pub created: Option<String>,
26    /// Date modified (`dcterms:modified`).
27    pub modified: Option<String>,
28}
29
30impl CoreProperties {
31    /// Parse `docProps/core.xml` content.
32    pub fn from_xml(xml: &[u8]) -> Result<Self> {
33        let mut reader = Reader::from_reader(xml);
34        reader.config_mut().trim_text(true);
35
36        let mut props = CoreProperties::default();
37        let mut buf = Vec::new();
38
39        loop {
40            match reader.read_event_into(&mut buf) {
41                Ok(Event::Start(ref e)) => {
42                    let name = e.name();
43                    let field = match local_name(name.as_ref()) {
44                        b"title" => &mut props.title,
45                        b"creator" => &mut props.creator,
46                        b"subject" => &mut props.subject,
47                        b"description" => &mut props.description,
48                        b"keywords" => &mut props.keywords,
49                        b"lastModifiedBy" => &mut props.last_modified_by,
50                        b"created" => &mut props.created,
51                        b"modified" => &mut props.modified,
52                        _ => {
53                            buf.clear();
54                            continue;
55                        }
56                    };
57                    // Consume the whole element: a value containing an entity
58                    // arrives as several events, so a single Text event is not
59                    // enough to reconstruct it.
60                    let text = crate::xml_text::read_element_text(&mut reader, name);
61                    if !text.is_empty() {
62                        *field = Some(text);
63                    }
64                }
65                Ok(Event::Eof) => break,
66                Err(e) => return Err(e.into()),
67                _ => {}
68            }
69            buf.clear();
70        }
71
72        Ok(props)
73    }
74
75    /// Serialize to `docProps/core.xml` bytes.
76    pub fn to_xml(&self) -> Result<Vec<u8>> {
77        use quick_xml::Writer;
78        use quick_xml::events::{BytesDecl, BytesEnd, BytesStart, BytesText};
79
80        let mut writer = Writer::new_with_indent(Vec::new(), b' ', 2);
81
82        writer.write_event(Event::Decl(BytesDecl::new(
83            "1.0",
84            Some("UTF-8"),
85            Some("yes"),
86        )))?;
87
88        let mut root = BytesStart::new("cp:coreProperties");
89        root.push_attribute((
90            "xmlns:cp",
91            "http://schemas.openxmlformats.org/package/2006/metadata/core-properties",
92        ));
93        root.push_attribute(("xmlns:dc", "http://purl.org/dc/elements/1.1/"));
94        root.push_attribute(("xmlns:dcterms", "http://purl.org/dc/terms/"));
95        root.push_attribute(("xmlns:dcmitype", "http://purl.org/dc/dcmitype/"));
96        root.push_attribute(("xmlns:xsi", "http://www.w3.org/2001/XMLSchema-instance"));
97        writer.write_event(Event::Start(root))?;
98
99        fn write_element<W: std::io::Write>(
100            writer: &mut Writer<W>,
101            tag: &str,
102            value: &Option<String>,
103        ) -> Result<()> {
104            if let Some(val) = value {
105                writer.write_event(Event::Start(BytesStart::new(tag)))?;
106                writer.write_event(Event::Text(BytesText::new(val)))?;
107                writer.write_event(Event::End(BytesEnd::new(tag)))?;
108            }
109            Ok(())
110        }
111
112        fn write_date_element<W: std::io::Write>(
113            writer: &mut Writer<W>,
114            tag: &str,
115            value: &Option<String>,
116        ) -> Result<()> {
117            if let Some(val) = value {
118                let mut e = BytesStart::new(tag);
119                e.push_attribute(("xsi:type", "dcterms:W3CDTF"));
120                writer.write_event(Event::Start(e))?;
121                writer.write_event(Event::Text(BytesText::new(val)))?;
122                writer.write_event(Event::End(BytesEnd::new(tag)))?;
123            }
124            Ok(())
125        }
126
127        write_element(&mut writer, "dc:title", &self.title)?;
128        write_element(&mut writer, "dc:subject", &self.subject)?;
129        write_element(&mut writer, "dc:creator", &self.creator)?;
130        write_element(&mut writer, "cp:keywords", &self.keywords)?;
131        write_element(&mut writer, "dc:description", &self.description)?;
132        write_element(&mut writer, "cp:lastModifiedBy", &self.last_modified_by)?;
133        write_date_element(&mut writer, "dcterms:created", &self.created)?;
134        write_date_element(&mut writer, "dcterms:modified", &self.modified)?;
135
136        writer.write_event(Event::End(BytesEnd::new("cp:coreProperties")))?;
137
138        Ok(writer.into_inner())
139    }
140}
141
142#[cfg(test)]
143mod tests {
144    use super::*;
145
146    #[test]
147    fn parse_core_properties() {
148        let xml = br#"<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
149<cp:coreProperties xmlns:cp="http://schemas.openxmlformats.org/package/2006/metadata/core-properties"
150                   xmlns:dc="http://purl.org/dc/elements/1.1/"
151                   xmlns:dcterms="http://purl.org/dc/terms/"
152                   xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
153  <dc:title>Test Document</dc:title>
154  <dc:creator>John Doe</dc:creator>
155  <dc:subject>Testing</dc:subject>
156  <dc:description>A test document</dc:description>
157  <cp:keywords>test, document</cp:keywords>
158  <cp:lastModifiedBy>Jane Doe</cp:lastModifiedBy>
159  <dcterms:created xsi:type="dcterms:W3CDTF">2024-01-15T10:30:00Z</dcterms:created>
160  <dcterms:modified xsi:type="dcterms:W3CDTF">2024-06-20T14:00:00Z</dcterms:modified>
161</cp:coreProperties>"#;
162
163        let props = CoreProperties::from_xml(xml).unwrap();
164        assert_eq!(props.title, Some("Test Document".to_string()));
165        assert_eq!(props.creator, Some("John Doe".to_string()));
166        assert_eq!(props.subject, Some("Testing".to_string()));
167        assert_eq!(props.description, Some("A test document".to_string()));
168        assert_eq!(props.keywords, Some("test, document".to_string()));
169        assert_eq!(props.last_modified_by, Some("Jane Doe".to_string()));
170        assert_eq!(props.created, Some("2024-01-15T10:30:00Z".to_string()));
171        assert_eq!(props.modified, Some("2024-06-20T14:00:00Z".to_string()));
172    }
173
174    #[test]
175    fn round_trip_core_properties() {
176        let props = CoreProperties {
177            title: Some("My Title".to_string()),
178            creator: Some("Author".to_string()),
179            subject: None,
180            description: None,
181            keywords: Some("rust, docx".to_string()),
182            last_modified_by: None,
183            created: Some("2024-01-01T00:00:00Z".to_string()),
184            modified: Some("2024-06-01T00:00:00Z".to_string()),
185        };
186
187        let xml = props.to_xml().unwrap();
188        let parsed = CoreProperties::from_xml(&xml).unwrap();
189
190        assert_eq!(parsed.title, props.title);
191        assert_eq!(parsed.creator, props.creator);
192        assert_eq!(parsed.keywords, props.keywords);
193        assert_eq!(parsed.created, props.created);
194        assert_eq!(parsed.modified, props.modified);
195    }
196}
oxml_core/core_properties.rs

oxml_core/
core_properties.rs
