feat(ocr): tree-like structure with text levels

Cargo.toml

@@ -49,7 +49,7 @@ nvisy-rig = { path = "./crates/nvisy-rig", version = "0.1.0" }
 nvisy-server = { path = "./crates/nvisy-server", version = "0.1.0" }
 
 # LLM framework
-rig-core = { version = "0.31", features = [] }
+rig-core = { version = "0.32", features = [] }
 
 # HTTP client and middleware
 reqwest = { version = "0.12", default-features = false, features = ["json", "rustls-tls", "multipart"] }

crates/nvisy-codec/src/document/any.rs

@@ -170,14 +170,16 @@ impl From<Document<DocxHandler>> for AnyDocument {
 
 #[cfg(test)]
 mod tests {
+    use nvisy_core::fs::{AudioFormat, ImageFormat, TextFormat};
+
     use super::*;
 
     #[test]
     fn from_txt_handler() {
         let doc = Document::new(TxtHandler::new(vec!["hello".into()], false));
         let any: AnyDocument = doc.into();
         assert!(any.as_text().is_some());
-        assert_eq!(any.document_type(), DocumentType::Txt);
+        assert_eq!(any.document_type(), DocumentType::Text(TextFormat::Txt));
     }
 
     #[test]
@@ -186,15 +188,15 @@ mod tests {
         let doc = Document::new(PngHandler::new(img));
         let any: AnyDocument = doc.into();
         assert!(any.as_image().is_some());
-        assert_eq!(any.document_type(), DocumentType::Png);
+        assert_eq!(any.document_type(), DocumentType::Image(ImageFormat::Png));
     }
 
     #[test]
     fn from_wav_handler() {
         let doc = Document::new(WavHandler::new(bytes::Bytes::from_static(b"wav")));
         let any: AnyDocument = doc.into();
         assert!(any.as_audio().is_some());
-        assert_eq!(any.document_type(), DocumentType::Wav);
+        assert_eq!(any.document_type(), DocumentType::Audio(AudioFormat::Wav));
     }
 
     #[test]

crates/nvisy-codec/src/document/mod.rs

@@ -178,7 +178,10 @@ mod tests {
     fn document_type_delegates() {
         let handler = TxtHandler::new(vec![], false);
         let doc = Document::new(handler);
-        assert_eq!(doc.document_type(), DocumentType::Txt);
+        assert_eq!(
+            doc.document_type(),
+            DocumentType::Text(nvisy_core::fs::TextFormat::Txt),
+        );
     }
 
     #[test]

crates/nvisy-codec/src/handler/audio/audio_handler.rs

@@ -104,7 +104,10 @@ mod tests {
     #[tokio::test]
     async fn wav_variant_delegates() {
         let h = AnyAudio::Wav(WavHandler::new(bytes::Bytes::from_static(b"wav-data")));
-        assert_eq!(h.document_type(), DocumentType::Wav);
+        assert_eq!(
+            h.document_type(),
+            DocumentType::Audio(nvisy_core::fs::AudioFormat::Wav),
+        );
         let spans: Vec<_> = h.audio_spans().await.collect().await;
         assert_eq!(spans.len(), 1);
         assert_eq!(spans[0].data.as_bytes().as_ref(), b"wav-data");
@@ -113,7 +116,10 @@ mod tests {
     #[tokio::test]
     async fn mp3_variant_delegates() {
         let h = AnyAudio::Mp3(Mp3Handler::new(bytes::Bytes::from_static(b"mp3-data")));
-        assert_eq!(h.document_type(), DocumentType::Mp3);
+        assert_eq!(
+            h.document_type(),
+            DocumentType::Audio(nvisy_core::fs::AudioFormat::Mp3),
+        );
         assert_eq!(h.encode().unwrap().as_bytes(), b"mp3-data");
     }
 

crates/nvisy-codec/src/handler/audio/mp3_handler.rs

@@ -10,7 +10,7 @@
 use bytes::Bytes;
 use futures::StreamExt;
 use nvisy_core::Error;
-use nvisy_core::fs::DocumentType;
+use nvisy_core::fs::{AudioFormat, DocumentType};
 use nvisy_core::io::ContentData;
 use nvisy_core::path::ContentSource;
 
@@ -45,7 +45,7 @@ impl Mp3Handler {
 
 impl Handler for Mp3Handler {
     fn document_type(&self) -> DocumentType {
-        DocumentType::Mp3
+        DocumentType::Audio(AudioFormat::Mp3)
     }
 
     #[tracing::instrument(name = "mp3.encode", skip_all, fields(output_bytes))]

crates/nvisy-codec/src/handler/audio/wav_handler.rs

@@ -10,7 +10,7 @@
 use bytes::Bytes;
 use futures::StreamExt;
 use nvisy_core::Error;
-use nvisy_core::fs::DocumentType;
+use nvisy_core::fs::{AudioFormat, DocumentType};
 use nvisy_core::io::ContentData;
 use nvisy_core::path::ContentSource;
 
@@ -45,7 +45,7 @@ impl WavHandler {
 
 impl Handler for WavHandler {
     fn document_type(&self) -> DocumentType {
-        DocumentType::Wav
+        DocumentType::Audio(AudioFormat::Wav)
     }
 
     #[tracing::instrument(name = "wav.encode", skip_all, fields(output_bytes))]

crates/nvisy-codec/src/handler/image/image_handler.rs

@@ -113,13 +113,19 @@ mod tests {
     #[test]
     fn png_variant_document_type() {
         let h = AnyImage::Png(make_png());
-        assert_eq!(h.document_type(), DocumentType::Png);
+        assert_eq!(
+            h.document_type(),
+            DocumentType::Image(nvisy_core::fs::ImageFormat::Png),
+        );
     }
 
     #[test]
     fn jpeg_variant_document_type() {
         let h = AnyImage::Jpeg(make_jpeg());
-        assert_eq!(h.document_type(), DocumentType::Jpeg);
+        assert_eq!(
+            h.document_type(),
+            DocumentType::Image(nvisy_core::fs::ImageFormat::Jpeg),
+        );
     }
 
     #[tokio::test]

crates/nvisy-codec/src/handler/image/jpeg_handler.rs

@@ -22,7 +22,7 @@ pub struct JpegHandler {
 
 impl_image_handler!(
     JpegHandler,
-    nvisy_core::fs::DocumentType::Jpeg,
+    nvisy_core::fs::DocumentType::Image(nvisy_core::fs::ImageFormat::Jpeg),
     image::ImageFormat::Jpeg,
     "jpeg-handler",
     "jpeg.encode"

crates/nvisy-codec/src/handler/image/png_handler.rs

@@ -22,7 +22,7 @@ pub struct PngHandler {
 
 impl_image_handler!(
     PngHandler,
-    nvisy_core::fs::DocumentType::Png,
+    nvisy_core::fs::DocumentType::Image(nvisy_core::fs::ImageFormat::Png),
     image::ImageFormat::Png,
     "png-handler",
     "png.encode"

crates/nvisy-codec/src/handler/rich/docx_handler.rs

@@ -1,7 +1,7 @@
 //! DOCX handler (stub: awaiting migration to full Loader/Handler pattern).
 
 use nvisy_core::Error;
-use nvisy_core::fs::DocumentType;
+use nvisy_core::fs::{DocumentType, WordFormat};
 use nvisy_core::io::ContentData;
 
 use crate::document::{SpanEditStream, SpanStream};
@@ -14,7 +14,7 @@ pub struct DocxHandler;
 
 impl Handler for DocxHandler {
     fn document_type(&self) -> DocumentType {
-        DocumentType::Docx
+        DocumentType::Word(WordFormat::Docx)
     }
 
     #[tracing::instrument(name = "docx.encode", skip_all)]

crates/nvisy-codec/src/handler/text/csv_handler.rs

@@ -17,7 +17,7 @@
 
 use futures::StreamExt;
 use nvisy_core::Error;
-use nvisy_core::fs::DocumentType;
+use nvisy_core::fs::{DocumentType, SpreadsheetFormat};
 use nvisy_core::io::ContentData;
 use nvisy_core::path::ContentSource;
 
@@ -84,7 +84,7 @@ pub struct CsvHandler {
 
 impl Handler for CsvHandler {
     fn document_type(&self) -> DocumentType {
-        DocumentType::Csv
+        DocumentType::Spreadsheet(SpreadsheetFormat::Csv)
     }
 
     #[tracing::instrument(name = "csv.encode", skip_all, fields(output_bytes))]

crates/nvisy-codec/src/handler/text/csv_loader.rs

@@ -137,7 +137,7 @@ mod tests {
     use bytes::Bytes;
     use futures::StreamExt;
     use nvisy_core::Error;
-    use nvisy_core::fs::DocumentType;
+    use nvisy_core::fs::{DocumentType, SpreadsheetFormat};
     use nvisy_core::path::ContentSource;
 
     use super::*;
@@ -151,7 +151,7 @@ mod tests {
         let content = content_from_str("name,age\nAlice,30\nBob,25\n");
         let doc = CsvLoader.decode(&content, &CsvParams::default()).await?;
 
-        assert_eq!(doc.document_type(), DocumentType::Csv);
+        assert_eq!(doc.document_type(), DocumentType::Spreadsheet(SpreadsheetFormat::Csv));
         assert_eq!(
             doc.headers(),
             Some(["name", "age"].map(String::from).as_slice())

crates/nvisy-codec/src/handler/text/json_handler.rs

@@ -23,7 +23,7 @@ use std::num::NonZeroU32;
 
 use futures::StreamExt;
 use nvisy_core::Error;
-use nvisy_core::fs::DocumentType;
+use nvisy_core::fs::{DocumentType, TextFormat};
 use nvisy_core::io::ContentData;
 use nvisy_core::path::ContentSource;
 use serde::{Deserialize, Serialize};
@@ -127,7 +127,7 @@ pub struct JsonHandler {
 
 impl Handler for JsonHandler {
     fn document_type(&self) -> DocumentType {
-        DocumentType::Json
+        DocumentType::Text(TextFormat::Json)
     }
 
     #[tracing::instrument(name = "json.encode", skip_all, fields(output_bytes))]

crates/nvisy-codec/src/handler/text/json_loader.rs

@@ -94,7 +94,7 @@ fn detect_formatting(source: &str) -> (JsonIndent, bool) {
 mod tests {
     use bytes::Bytes;
     use nvisy_core::Error;
-    use nvisy_core::fs::DocumentType;
+    use nvisy_core::fs::{DocumentType, TextFormat};
     use nvisy_core::path::ContentSource;
     use serde_json::json;
 
@@ -109,7 +109,7 @@ mod tests {
         let content = content_from_str(r#"{"name": "Alice", "age": 30}"#);
         let doc = JsonLoader.decode(&content, &JsonParams::default()).await?;
 
-        assert_eq!(doc.document_type(), DocumentType::Json);
+        assert_eq!(doc.document_type(), DocumentType::Text(TextFormat::Json));
         assert_eq!(doc.value(), &json!({"name": "Alice", "age": 30}));
         Ok(())
     }