impl header reader with rust (#376)

* fix: document_rels reader to keep multi rels

* feat: Add header reader

* faet: assign header to sectionProperty

* spec: update snaps

* feat: read titlePg

* spec: update snaps

.github/workflows/ci.yml

@@ -88,7 +88,7 @@ jobs:
       - uses: actions-rs/toolchain@v1
         with:
           profile: minimal
-          toolchain: 1.45.1
+          toolchain: 1.56.1
           override: true
       - run: rustup component add clippy
       - uses: actions-rs/cargo@v1

docx-core/examples/reader.rs

@@ -4,7 +4,7 @@ use std::fs::File;
 use std::io::{Read, Write};
 
 pub fn main() {
-    let mut file = File::open("./spacing.docx").unwrap();
+    let mut file = File::open("./header.docx").unwrap();
     let mut buf = vec![];
     file.read_to_end(&mut buf).unwrap();
 

docx-core/src/documents/elements/delete.rs

@@ -97,8 +97,10 @@ impl Delete {
 impl HistoryId for Delete {}
 
 impl BuildXML for Delete {
+    #[allow(clippy::needless_borrow)]
     fn build(&self) -> Vec<u8> {
-        let mut b = XMLBuilder::new().open_delete(&self.generate(), &self.author, &self.date);
+        let id = self.generate();
+        let mut b = XMLBuilder::new().open_delete(&id, &self.author, &self.date);
         for c in &self.children {
             match c {
                 DeleteChild::Run(t) => b = b.add_child(t),

docx-core/src/documents/elements/header_reference.rs

@@ -6,8 +6,8 @@ use serde::Serialize;
 #[derive(Debug, Clone, PartialEq, Serialize)]
 #[serde(rename_all = "camelCase")]
 pub struct HeaderReference {
-    header_type: String,
-    id: String,
+    pub header_type: String,
+    pub id: String,
 }
 
 impl Default for HeaderReference {

docx-core/src/documents/elements/insert.rs

@@ -113,9 +113,10 @@ impl Insert {
     }
 
     pub fn add_comment_start(mut self, comment: Comment) -> Self {
-        self.children.push(InsertChild::CommentStart(Box::new(
-            CommentRangeStart::new(comment),
-        )));
+        self.children
+            .push(InsertChild::CommentStart(Box::new(CommentRangeStart::new(
+                comment,
+            ))));
         self
     }
 
@@ -139,6 +140,7 @@ impl Insert {
 impl HistoryId for Insert {}
 
 impl BuildXML for Insert {
+    #[allow(clippy::needless_borrow)]
     fn build(&self) -> Vec<u8> {
         XMLBuilder::new()
             .open_insert(&self.generate(), &self.author, &self.date)

docx-core/src/reader/document_rels.rs

@@ -1,4 +1,5 @@
 use std::collections::BTreeMap;
+use std::collections::HashSet;
 use std::io::{Cursor, Read};
 use std::path::*;
 use std::str::FromStr;
@@ -8,14 +9,18 @@ use xml::reader::{EventReader, XmlEvent};
 use super::errors::*;
 use super::*;
 
+pub type RId = String;
+
 #[derive(Debug, Clone, PartialEq)]
 pub struct ReadDocumentRels {
-    rels: BTreeMap<String, PathBuf>,
+    rels: BTreeMap<String, HashSet<(RId, PathBuf)>>,
 }
 
 impl ReadDocumentRels {
-    pub fn find_target_path(&self, target: &str) -> Option<PathBuf> {
-        self.rels.get(target).cloned()
+    pub fn find_target_path(&self, target: &str) -> Option<Vec<(RId, PathBuf)>> {
+        self.rels
+            .get(target)
+            .map(|s| s.clone().into_iter().collect())
     }
 }
 
@@ -51,6 +56,7 @@ fn read_rels_xml<R: Read>(
                 let e = XMLElement::from_str(&name.local_name).unwrap();
                 if let XMLElement::Relationship = e {
                     let mut rel_type = "".to_owned();
+                    let mut rid = "".to_owned();
                     let mut target = PathBuf::default();
                     for a in attributes {
                         let local_name = &a.name.local_name;
@@ -58,9 +64,18 @@ fn read_rels_xml<R: Read>(
                             rel_type = a.value.to_owned();
                         } else if local_name == "Target" {
                             target = Path::new(dir.as_ref()).join(a.value);
+                        } else if local_name == "Id" {
+                            rid = a.value.to_owned();
                         }
                     }
-                    rels.rels.insert(rel_type, target);
+                    let current = rels.rels.remove(&rel_type);
+                    if let Some(mut paths) = current {
+                        paths.insert((rid, target));
+                        rels.rels.insert(rel_type, paths);
+                    } else {
+                        let s: HashSet<(RId, PathBuf)> = vec![(rid, target)].into_iter().collect();
+                        rels.rels.insert(rel_type, s);
+                    }
                     continue;
                 }
             }

docx-core/src/reader/header.rs

@@ -0,0 +1,71 @@
+use std::io::Read;
+use std::str::FromStr;
+
+use crate::reader::*;
+use xml::reader::{EventReader, XmlEvent};
+
+use super::{Paragraph, Table};
+
+impl FromXML for Header {
+    fn from_xml<R: Read>(reader: R) -> Result<Self, ReaderError> {
+        let mut parser = EventReader::new(reader);
+        let mut header = Self::default();
+        loop {
+            let e = parser.next();
+            match e {
+                Ok(XmlEvent::StartElement {
+                    attributes, name, ..
+                }) => {
+                    let e = XMLElement::from_str(&name.local_name).unwrap();
+                    match e {
+                        XMLElement::Paragraph => {
+                            let p = Paragraph::read(&mut parser, &attributes)?;
+                            header = header.add_paragraph(p);
+                            continue;
+                        }
+                        XMLElement::Table => {
+                            let t = Table::read(&mut parser, &attributes)?;
+                            header = header.add_table(t);
+                            continue;
+                        }
+                        _ => {}
+                    }
+                }
+                Ok(XmlEvent::EndDocument) => break,
+                Err(_) => return Err(ReaderError::XMLReadError),
+                _ => {}
+            }
+        }
+        Ok(header)
+    }
+}
+
+#[test]
+fn test_header_from_xml() {
+    let xml = r#"<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
+<w:hdr xmlns:r="http://schemas.openxmlformats.org/officeDocument/2006/relationships"
+    xmlns:o="urn:schemas-microsoft-com:office:office"
+    xmlns:v="urn:schemas-microsoft-com:vml"
+    xmlns:w="http://schemas.openxmlformats.org/wordprocessingml/2006/main"
+    xmlns:w10="urn:schemas-microsoft-com:office:word"
+    xmlns:wp="http://schemas.openxmlformats.org/drawingml/2006/wordprocessingDrawing"
+    xmlns:wps="http://schemas.microsoft.com/office/word/2010/wordprocessingShape"
+    xmlns:wpg="http://schemas.microsoft.com/office/word/2010/wordprocessingGroup"
+    xmlns:mc="http://schemas.openxmlformats.org/markup-compatibility/2006"
+    xmlns:wp14="http://schemas.microsoft.com/office/word/2010/wordprocessingDrawing"
+    xmlns:w14="http://schemas.microsoft.com/office/word/2010/wordml" mc:Ignorable="w14 wp14">
+    <w:p w14:paraId="12345678">
+        <w:pPr>
+            <w:rPr />
+        </w:pPr>
+        <w:r>
+            <w:rPr />
+            <w:t xml:space="preserve">Hello Header</w:t>
+        </w:r>
+    </w:p>
+</w:hdr>"#;
+    let h = Header::from_xml(xml.as_bytes()).unwrap();
+    let expected =
+        Header::new().add_paragraph(Paragraph::new().add_run(Run::new().add_text("Hello Header")));
+    assert_eq!(h, expected)
+}

docx-core/src/reader/mod.rs

@@ -17,6 +17,7 @@ mod document_rels;
 mod drawing;
 mod errors;
 mod from_xml;
+mod header;
 mod ignore;
 mod insert;
 mod level;
@@ -50,7 +51,7 @@ mod wps_shape;
 mod wps_text_box;
 mod xml_element;
 
-use std::io::Cursor;
+use std::{collections::HashMap, io::Cursor};
 
 use crate::documents::*;
 
@@ -61,6 +62,7 @@ pub use from_xml::*;
 pub use mc_fallback::*;
 pub use read_zip::*;
 pub use xml_element::*;
+use zip::ZipArchive;
 
 // 2006
 const DOC_RELATIONSHIP_TYPE: &str =
@@ -77,10 +79,36 @@ const COMMENTS_TYPE: &str =
     "http://schemas.openxmlformats.org/officeDocument/2006/relationships/comments";
 const WEB_SETTINGS_TYPE: &str =
     "http://schemas.openxmlformats.org/officeDocument/2006/relationships/webSettings";
+const HEADER_TYPE: &str =
+    "http://schemas.openxmlformats.org/officeDocument/2006/relationships/header";
 // 2011
 const COMMENTS_EXTENDED_TYPE: &str =
     "http://schemas.microsoft.com/office/2011/relationships/commentsExtended";
 
+fn read_headers(
+    rels: &ReadDocumentRels,
+    archive: &mut ZipArchive<Cursor<&[u8]>>,
+) -> HashMap<RId, Header> {
+    let header_paths = rels.find_target_path(HEADER_TYPE);
+    let headers: HashMap<RId, Header> = header_paths
+        .unwrap_or_default()
+        .into_iter()
+        .filter_map(|(rid, path)| {
+            let data = read_zip(archive, path.to_str().expect("should have header path."));
+            if let Ok(d) = data {
+                if let Ok(h) = Header::from_xml(&d[..]) {
+                    Some((rid, h))
+                } else {
+                    None
+                }
+            } else {
+                None
+            }
+        })
+        .collect();
+    headers
+}
+
 pub fn read_docx(buf: &[u8]) -> Result<Docx, ReaderError> {
     let mut docx = Docx::new();
     let cur = Cursor::new(buf);
@@ -121,9 +149,12 @@ pub fn read_docx(buf: &[u8]) -> Result<Docx, ReaderError> {
 
     let rels = read_document_rels(&mut archive, &document_path)?;
 
+    let headers = read_headers(&rels, &mut archive);
+
     // Read commentsExtended
     let comments_extended_path = rels.find_target_path(COMMENTS_EXTENDED_TYPE);
     let comments_extended = if let Some(comments_extended_path) = comments_extended_path {
+        if let Some((_, comments_extended_path)) = comments_extended_path.get(0) {
             let data = read_zip(
                 &mut archive,
                 comments_extended_path
@@ -137,11 +168,15 @@ pub fn read_docx(buf: &[u8]) -> Result<Docx, ReaderError> {
             }
         } else {
             CommentsExtended::default()
+        }
+    } else {
+        CommentsExtended::default()
     };
 
     // Read comments
     let comments_path = rels.find_target_path(COMMENTS_TYPE);
-    let comments = if let Some(comments_path) = comments_path {
+    let comments = if let Some(paths) = comments_path {
+        if let Some((_, comments_path)) = paths.get(0) {
             let data = read_zip(
                 &mut archive,
                 comments_path.to_str().expect("should have comments."),
@@ -185,6 +220,9 @@ pub fn read_docx(buf: &[u8]) -> Result<Docx, ReaderError> {
             }
         } else {
             Comments::default()
+        }
+    } else {
+        Comments::default()
     };
 
     let document = {
@@ -193,6 +231,28 @@ pub fn read_docx(buf: &[u8]) -> Result<Docx, ReaderError> {
     };
     docx = docx.document(document);
 
+    // assign headers
+    if let Some(h) = docx.document.section_property.header_reference.clone() {
+        if let Some(header) = headers.get(&h.id) {
+            docx.document = docx.document.header(header.clone(), &h.id);
+        }
+    }
+    if let Some(ref h) = docx
+        .document
+        .section_property
+        .first_header_reference
+        .clone()
+    {
+        if let Some(header) = headers.get(&h.id) {
+            docx.document = docx.document.first_header(header.clone(), &h.id);
+        }
+    }
+    if let Some(ref h) = docx.document.section_property.even_header_reference.clone() {
+        if let Some(header) = headers.get(&h.id) {
+            docx.document = docx.document.even_header(header.clone(), &h.id);
+        }
+    }
+
     // store comments to paragraphs.
     if !comments.inner().is_empty() {
         docx.store_comments(comments.inner());
@@ -203,7 +263,8 @@ pub fn read_docx(buf: &[u8]) -> Result<Docx, ReaderError> {
     // Read document relationships
     // Read styles
     let style_path = rels.find_target_path(STYLE_RELATIONSHIP_TYPE);
-    if let Some(style_path) = style_path {
+    if let Some(paths) = style_path {
+        if let Some((_, style_path)) = paths.get(0) {
             let data = read_zip(
                 &mut archive,
                 style_path.to_str().expect("should have styles"),
@@ -211,10 +272,12 @@ pub fn read_docx(buf: &[u8]) -> Result<Docx, ReaderError> {
             let styles = Styles::from_xml(&data[..])?;
             docx = docx.styles(styles);
         }
+    }
 
     // Read numberings
     let num_path = rels.find_target_path(NUMBERING_RELATIONSHIP_TYPE);
-    if let Some(num_path) = num_path {
+    if let Some(paths) = num_path {
+        if let Some((_, num_path)) = paths.get(0) {
             let data = read_zip(
                 &mut archive,
                 num_path.to_str().expect("should have numberings"),
@@ -222,10 +285,12 @@ pub fn read_docx(buf: &[u8]) -> Result<Docx, ReaderError> {
             let nums = Numberings::from_xml(&data[..])?;
             docx = docx.numberings(nums);
         }
+    }
 
     // Read settings
     let settings_path = rels.find_target_path(SETTINGS_TYPE);
-    if let Some(settings_path) = settings_path {
+    if let Some(paths) = settings_path {
+        if let Some((_, settings_path)) = paths.get(0) {
             let data = read_zip(
                 &mut archive,
                 settings_path.to_str().expect("should have settings"),
@@ -233,11 +298,12 @@ pub fn read_docx(buf: &[u8]) -> Result<Docx, ReaderError> {
             let settings = Settings::from_xml(&data[..])?;
             docx = docx.settings(settings);
         }
+    }
 
     // Read web settings
     let web_settings_path = rels.find_target_path(WEB_SETTINGS_TYPE);
-    dbg!(&web_settings_path);
-    if let Some(web_settings_path) = web_settings_path {
+    if let Some(paths) = web_settings_path {
+        if let Some((_, web_settings_path)) = paths.get(0) {
             let data = read_zip(
                 &mut archive,
                 web_settings_path
@@ -247,6 +313,7 @@ pub fn read_docx(buf: &[u8]) -> Result<Docx, ReaderError> {
             let web_settings = WebSettings::from_xml(&data[..])?;
             docx = docx.web_settings(web_settings);
         }
+    }
 
     Ok(docx)
 }

docx-core/src/reader/rels.rs

@@ -52,13 +52,14 @@ mod tests {
   <Relationship Id="rId1" Type="http://schemas.openxmlformats.org/package/2006/relationships/metadata/core-properties" Target="docProps/core.xml" />
 </Relationships>"#;
         let c = Rels::from_xml(xml.as_bytes()).unwrap();
-        let mut rels = Vec::new();
-        rels.push((
+        let rels =
+            vec![
+        (
             "http://schemas.openxmlformats.org/package/2006/relationships/metadata/core-properties"
                 .to_owned(),
             "rId1".to_owned(),
             "docProps/core.xml".to_owned(),
-        ));
+        )];
         assert_eq!(Rels { rels }, c);
     }
 }

docx-core/src/reader/section_property.rs

@@ -57,6 +57,24 @@ fn read_page_margin(
     Ok(margin)
 }
 
+fn read_header_reference(attributes: &[OwnedAttribute]) -> Result<(String, String), ReaderError> {
+    let mut rid = "".to_owned();
+    let mut header_type = "default".to_owned();
+    for a in attributes {
+        let local_name = &a.name.local_name;
+        match local_name.as_str() {
+            "type" => {
+                header_type = a.value.to_owned();
+            }
+            "id" => {
+                rid = a.value.to_owned();
+            }
+            _ => {}
+        }
+    }
+    Ok((rid, header_type))
+}
+
 impl ElementReader for SectionProperty {
     fn read<R: Read>(
         r: &mut EventReader<R>,
@@ -84,6 +102,26 @@ impl ElementReader for SectionProperty {
                                 sp = sp.doc_grid(doc_grid);
                             }
                         }
+                        XMLElement::HeaderReference => {
+                            if let Ok((rid, header_type)) = read_header_reference(&attributes) {
+                                match header_type.as_str() {
+                                    "default" => {
+                                        sp.header_reference =
+                                            Some(HeaderReference::new(header_type, rid));
+                                    }
+                                    "first" => {
+                                        sp.first_header_reference =
+                                            Some(HeaderReference::new(header_type, rid));
+                                    }
+                                    "even" => {
+                                        sp.even_header_reference =
+                                            Some(HeaderReference::new(header_type, rid));
+                                    }
+                                    _ => {}
+                                }
+                            }
+                        }
+                        XMLElement::TitlePg => sp = sp.title_pg(),
                         _ => {}
                     }
                 }

docx-core/src/reader/xml_element.rs

@@ -139,6 +139,9 @@ pub enum XMLElement {
     PageSize,
     PageMargin,
     WebSettings,
+    HeaderReference,
+    TitlePg,
+    EvenAndOddHeaders,
     Unsupported,
 }
 
@@ -335,6 +338,9 @@ impl FromStr for XMLElement {
             "keepLines" => Ok(XMLElement::KeepLines),
             "pageBreakBefore" => Ok(XMLElement::PageBreakBefore),
             "windowControl" => Ok(XMLElement::WindowControl),
+            "headerReference" => Ok(XMLElement::HeaderReference),
+            "titlePg" => Ok(XMLElement::TitlePg),
+            "evenAndOddHeaders" => Ok(XMLElement::EvenAndOddHeaders),
             _ => Ok(XMLElement::Unsupported),
         }
     }

docx-wasm/test/__snapshots__/index.test.js.snap

@@ -1695,6 +1695,14 @@ Object {
         "gridType": "lines",
         "linePitch": 360,
       },
+      "header": Object {
+        "children": Array [],
+        "hasNumbering": false,
+      },
+      "headerReference": Object {
+        "headerType": "default",
+        "id": "rId4",
+      },
       "pageMargin": Object {
         "bottom": 1701,
         "footer": 992,
@@ -3815,6 +3823,14 @@ Object {
         "gridType": "lines",
         "linePitch": 360,
       },
+      "header": Object {
+        "children": Array [],
+        "hasNumbering": false,
+      },
+      "headerReference": Object {
+        "headerType": "default",
+        "id": "rId4",
+      },
       "pageMargin": Object {
         "bottom": 1701,
         "footer": 992,
@@ -7225,6 +7241,14 @@ Object {
         "gridType": "lines",
         "linePitch": 360,
       },
+      "header": Object {
+        "children": Array [],
+        "hasNumbering": false,
+      },
+      "headerReference": Object {
+        "headerType": "default",
+        "id": "rId4",
+      },
       "pageMargin": Object {
         "bottom": 1701,
         "footer": 992,
@@ -12491,6 +12515,79 @@ Object {
         "gridType": "lines",
         "linePitch": 360,
       },
+      "header": Object {
+        "children": Array [
+          Object {
+            "data": Object {
+              "children": Array [
+                Object {
+                  "data": Object {
+                    "children": Array [],
+                    "runProperty": Object {
+                      "bold": null,
+                      "boldCs": null,
+                      "characterSpacing": null,
+                      "color": null,
+                      "del": null,
+                      "fonts": null,
+                      "highlight": null,
+                      "ins": null,
+                      "italic": null,
+                      "italicCs": null,
+                      "sz": null,
+                      "szCs": null,
+                      "textBorder": null,
+                      "underline": null,
+                      "vanish": null,
+                      "vertAlign": null,
+                    },
+                  },
+                  "type": "run",
+                },
+              ],
+              "hasNumbering": false,
+              "id": "0C07C25B",
+              "property": Object {
+                "alignment": null,
+                "divId": null,
+                "indent": null,
+                "keepLines": false,
+                "keepNext": false,
+                "lineSpacing": null,
+                "numberingProperty": null,
+                "outlineLvl": null,
+                "pageBreakBefore": false,
+                "runProperty": Object {
+                  "bold": null,
+                  "boldCs": null,
+                  "characterSpacing": null,
+                  "color": null,
+                  "del": null,
+                  "fonts": null,
+                  "highlight": null,
+                  "ins": null,
+                  "italic": null,
+                  "italicCs": null,
+                  "sz": null,
+                  "szCs": null,
+                  "textBorder": null,
+                  "underline": null,
+                  "vanish": null,
+                  "vertAlign": null,
+                },
+                "style": null,
+                "windowControl": false,
+              },
+            },
+            "type": "paragraph",
+          },
+        ],
+        "hasNumbering": false,
+      },
+      "headerReference": Object {
+        "headerType": "default",
+        "id": "rId11",
+      },
       "pageMargin": Object {
         "bottom": 1701,
         "footer": 992,