add extractors for docx, pptx, pdf, etc.

1 files changed, 9 insertions, 0 deletions

diff --git a/src/text/pptx.rs b/src/text/pptx.rs
new file mode 100644
index 0000000..7dac77e
--- /dev/null
+++ b/src/text/pptx.rs
@@ -0,0 +1,9 @@
+use std::io::Read;
+use dotext::{Pptx, MsDoc};
+
+pub fn get_text(path : &str) -> String {
+    let mut file = Pptx::open(path).unwrap();
+    let mut content = String::new();
+    let _ = file.read_to_string(&mut content);
+    content
+}