feat: preserve original filename with short hash suffix on upload

Author: bendersej

agents/src/routes.rs

@@ -138,8 +138,11 @@ async fn handle_post(
             .map_err(|_| {
                 AppError::BadRequest("Expected multipart/form-data with a 'file' field".into())
             })?;
-        let pdf_bytes = extract_multipart(multipart).await?;
-        let result = state.storage.upload(pdf_bytes).await?;
+        let upload = extract_multipart(multipart).await?;
+        let result = state
+            .storage
+            .upload(upload.bytes, upload.filename.as_deref())
+            .await?;
 
         Ok(Json(AgentResponse::new(
             &result.presigned_url,
@@ -179,18 +182,27 @@ async fn handle_post(
     }
 }
 
-async fn extract_multipart(mut multipart: Multipart) -> Result<Vec<u8>, AppError> {
+struct MultipartUpload {
+    bytes: Vec<u8>,
+    filename: Option<String>,
+}
+
+async fn extract_multipart(mut multipart: Multipart) -> Result<MultipartUpload, AppError> {
     while let Some(field) = multipart
         .next_field()
         .await
         .map_err(|_| AppError::BadRequest("Failed to read multipart field".into()))?
     {
         if field.name() == Some("file") {
+            let filename = field.file_name().map(String::from);
             let bytes = field
                 .bytes()
                 .await
                 .map_err(|_| AppError::BadRequest("Failed to read file".into()))?;
-            return Ok(bytes.to_vec());
+            return Ok(MultipartUpload {
+                bytes: bytes.to_vec(),
+                filename,
+            });
         }
     }
 

agents/src/storage.rs

@@ -42,13 +42,17 @@ impl Storage {
         }
     }
 
-    pub async fn upload(&self, bytes: Vec<u8>) -> Result<UploadResult, AppError> {
+    pub async fn upload(
+        &self,
+        bytes: Vec<u8>,
+        original_filename: Option<&str>,
+    ) -> Result<UploadResult, AppError> {
         if bytes.len() < 5 || &bytes[..5] != b"%PDF-" {
             return Err(AppError::BadRequest("Not a valid PDF file".into()));
         }
 
-        let id = Uuid::new_v4().to_string();
-        let key = format!("uploads/{id}.pdf");
+        let hash = &Uuid::new_v4().to_string()[..8];
+        let key = build_key(original_filename, hash);
 
         self.client
             .put_object()
@@ -78,3 +82,66 @@ impl Storage {
         Ok(UploadResult { presigned_url })
     }
 }
+
+fn build_key(original_filename: Option<&str>, hash: &str) -> String {
+    let stem_and_ext = original_filename
+        .filter(|name| !name.is_empty())
+        .map(|name| {
+            let name = name
+                .rsplit('/')
+                .next()
+                .unwrap_or(name)
+                .rsplit('\\')
+                .next()
+                .unwrap_or(name);
+
+            match name.rsplit_once('.') {
+                Some((stem, ext)) => (stem.to_string(), format!(".{ext}")),
+                None => (name.to_string(), String::new()),
+            }
+        });
+
+    match stem_and_ext {
+        Some((stem, ext)) => format!("uploads/{stem}-{hash}{ext}"),
+        None => format!("uploads/{hash}.pdf"),
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_build_key_with_filename() {
+        assert_eq!(
+            build_key(Some("invoice.pdf"), "a1b2c3d4"),
+            "uploads/invoice-a1b2c3d4.pdf"
+        );
+    }
+
+    #[test]
+    fn test_build_key_with_path() {
+        assert_eq!(
+            build_key(Some("path/to/report.pdf"), "a1b2c3d4"),
+            "uploads/report-a1b2c3d4.pdf"
+        );
+    }
+
+    #[test]
+    fn test_build_key_without_extension() {
+        assert_eq!(
+            build_key(Some("document"), "a1b2c3d4"),
+            "uploads/document-a1b2c3d4"
+        );
+    }
+
+    #[test]
+    fn test_build_key_none() {
+        assert_eq!(build_key(None, "a1b2c3d4"), "uploads/a1b2c3d4.pdf");
+    }
+
+    #[test]
+    fn test_build_key_empty() {
+        assert_eq!(build_key(Some(""), "a1b2c3d4"), "uploads/a1b2c3d4.pdf");
+    }
+}