Merge pull request #44 from messense/odf

Add support for Open Document formats

README.md

@@ -185,6 +185,9 @@ assert_eq!(kind.extension(), "foo");
 - **xlsx** - `application/vnd.openxmlformats-officedocument.spreadsheetml.sheet`
 - **ppt** - `application/vnd.ms-powerpoint`
 - **pptx** - `application/vnd.openxmlformats-officedocument.presentationml.presentation`
+- **odt** - `application/vnd.oasis.opendocument.text`
+- **ods** - `application/vnd.oasis.opendocument.spreadsheet`
+- **odp** - `application/vnd.oasis.opendocument.presentation`
 
 #### Font
 

src/map.rs

@@ -357,6 +357,25 @@ matcher_map!(
         "pptx",
         matchers::doc::is_pptx
     ),
+    // OpenDocument
+    (
+        MatcherType::Doc,
+        "application/vnd.oasis.opendocument.text",
+        "odt",
+        matchers::odf::is_odt
+    ),
+    (
+        MatcherType::Doc,
+        "application/vnd.oasis.opendocument.spreadsheet",
+        "ods",
+        matchers::odf::is_ods
+    ),
+    (
+        MatcherType::Doc,
+        "application/vnd.oasis.opendocument.presentation",
+        "odp",
+        matchers::odf::is_odp
+    ),
     // Archive
     (
         MatcherType::Archive,

src/matchers/doc.rs

@@ -1,5 +1,7 @@
 use core::convert::TryInto;
 
+use super::compare_bytes;
+
 #[allow(clippy::upper_case_acronyms)]
 #[derive(Debug, Eq, PartialEq)]
 enum DocType {
@@ -123,24 +125,6 @@ fn ole2(buf: &[u8]) -> Option<DocType> {
     Some(DocType::DOC)
 }
 
-fn compare_bytes(slice: &[u8], sub_slice: &[u8], start_offset: usize) -> bool {
-    let sl = sub_slice.len();
-
-    if start_offset + sl > slice.len() {
-        return false;
-    }
-
-    for (i, v) in slice.iter().skip(start_offset).take(sl).enumerate() {
-        let v2 = sub_slice[i];
-
-        if *v != v2 {
-            return false;
-        }
-    }
-
-    true
-}
-
 fn check_msooml(buf: &[u8], offset: usize) -> Option<DocType> {
     if compare_bytes(buf, &[b'w', b'o', b'r', b'd', b'/'], offset) {
         Some(DocType::DOCX)

src/matchers/mod.rs

@@ -5,5 +5,24 @@ pub mod book;
 pub mod doc;
 pub mod font;
 pub mod image;
+pub mod odf;
 pub mod text;
 pub mod video;
+
+pub(crate) fn compare_bytes(slice: &[u8], sub_slice: &[u8], start_offset: usize) -> bool {
+    let sl = sub_slice.len();
+
+    if start_offset + sl > slice.len() {
+        return false;
+    }
+
+    for (i, v) in slice.iter().skip(start_offset).take(sl).enumerate() {
+        let v2 = sub_slice[i];
+
+        if *v != v2 {
+            return false;
+        }
+    }
+
+    true
+}

src/matchers/odf.rs

@@ -0,0 +1,48 @@
+use super::compare_bytes;
+
+#[derive(Debug, Eq, PartialEq)]
+enum DocType {
+    Text,
+    Spreadsheet,
+    Presentation,
+}
+
+/// Returns whether a buffer is OpenDocument Text
+pub fn is_odt(buf: &[u8]) -> bool {
+    odf(buf) == Some(DocType::Text)
+}
+
+/// Returns whether a buffer is OpenDocument Spreadsheet
+pub fn is_ods(buf: &[u8]) -> bool {
+    odf(buf) == Some(DocType::Spreadsheet)
+}
+
+/// Returns whether a buffer is OpenDocument Presentation
+pub fn is_odp(buf: &[u8]) -> bool {
+    odf(buf) == Some(DocType::Presentation)
+}
+
+fn odf(buf: &[u8]) -> Option<DocType> {
+    let signature = [b'P', b'K', 0x03, 0x04];
+
+    // start by checking for ZIP local file header signature
+    if !compare_bytes(buf, &signature, 0) {
+        return None;
+    }
+
+    // Check mimetype
+    if !compare_bytes(buf, b"mimetype", 0x1E) {
+        return None;
+    }
+
+    if compare_bytes(buf, b"vnd.oasis.opendocument.text", 0x32) {
+        return Some(DocType::Text);
+    }
+    if compare_bytes(buf, b"vnd.oasis.opendocument.spreadsheet", 0x32) {
+        return Some(DocType::Spreadsheet);
+    }
+    if compare_bytes(buf, b"vnd.oasis.opendocument.presentation", 0x32) {
+        return Some(DocType::Presentation);
+    }
+    None
+}

tests/odf.rs

@@ -0,0 +1,25 @@
+mod common;
+
+test_format!(
+    Doc,
+    "application/vnd.oasis.opendocument.text",
+    "odt",
+    odt,
+    "sample.odt"
+);
+
+test_format!(
+    Doc,
+    "application/vnd.oasis.opendocument.spreadsheet",
+    "ods",
+    ods,
+    "sample.ods"
+);
+
+test_format!(
+    Doc,
+    "application/vnd.oasis.opendocument.presentation",
+    "odp",
+    odp,
+    "sample.odp"
+);