Add component to read PDF files into assistants (#439)

runtime/Cargo.lock

@@ -188,6 +188,12 @@ dependencies = [
  "system-deps 6.2.2",
 ]
 
+[[package]]
+name = "atoi_simd"
+version = "0.16.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "4790f9e8961209112beb783d85449b508673cf4a6a419c8449b210743ac4dbe9"
+
 [[package]]
 name = "atomic"
 version = "0.5.3"
@@ -430,15 +436,17 @@ dependencies = [
 
 [[package]]
 name = "calamine"
-version = "0.26.1"
+version = "0.27.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "138646b9af2c5d7f1804ea4bf93afc597737d2bd4f7341d67c48b03316976eb1"
+checksum = "6d80f81ba5c68206b9027e62346d49dc26fb32ffc4fe6ef7022a8ae21d348ccb"
 dependencies = [
+ "atoi_simd",
  "byteorder",
  "codepage",
  "encoding_rs",
+ "fast-float2",
  "log",
- "quick-xml 0.31.0",
+ "quick-xml 0.37.5",
  "serde",
  "zip 2.5.0",
 ]
@@ -1064,6 +1072,12 @@ dependencies = [
  "zune-inflate",
 ]
 
+[[package]]
+name = "fast-float2"
+version = "0.2.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f8eb564c5c7423d25c886fb561d1e4ee69f72354d16918afa32c08811f6b6a55"
+
 [[package]]
 name = "fastrand"
 version = "2.1.0"
@@ -3200,9 +3214,9 @@ dependencies = [
 
 [[package]]
 name = "pdfium-render"
-version = "0.8.30"
+version = "0.8.31"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2773a939ec2c736640f5f2e62a325c2e1a997d694961c50f17cadfb4c8682e84"
+checksum = "a473e1f4c2630f7240b127eb9efebd1d9e55053015b624bfe9dbeae0299947e1"
 dependencies = [
  "bitflags 2.6.0",
  "bytemuck",
@@ -3563,20 +3577,20 @@ checksum = "a993555f31e5a609f617c12db6250dedcac1b0a85076912c436e6fc9b2c8e6a3"
 
 [[package]]
 name = "quick-xml"
-version = "0.31.0"
+version = "0.32.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1004a344b30a54e2ee58d66a71b32d2db2feb0a31f9a2d302bf0536f15de2a33"
+checksum = "1d3a6e5838b60e0e8fa7a43f22ade549a37d61f8bdbe636d0d7816191de969c2"
 dependencies = [
- "encoding_rs",
  "memchr",
 ]
 
 [[package]]
 name = "quick-xml"
-version = "0.32.0"
+version = "0.37.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1d3a6e5838b60e0e8fa7a43f22ade549a37d61f8bdbe636d0d7816191de969c2"
+checksum = "331e97a1af0bf59823e6eadffe373d7b27f485be8748f71471c662c1f269b7fb"
 dependencies = [
+ "encoding_rs",
  "memchr",
 ]
 

runtime/Cargo.toml

@@ -34,8 +34,8 @@ hmac = "0.12.1"
 sha2 = "0.10.8"
 rcgen = { version = "0.13.2", features = ["pem"] }
 file-format = "0.26.0"
-calamine = "0.26.1"
-pdfium-render = "0.8.30"
+calamine = "0.27.0"
+pdfium-render = "0.8.31"
 sys-locale = "0.3.2"
 
 # Fixes security vulnerability downstream, where the upstream is not fixed yet:

runtime/src/app_window.rs

@@ -7,13 +7,14 @@ use rocket::serde::json::Json;
 use rocket::serde::Serialize;
 use serde::Deserialize;
 use tauri::updater::UpdateResponse;
-use tauri::{Manager, Window};
+use tauri::{Manager, PathResolver, Window};
 use tauri::api::dialog::blocking::FileDialogBuilder;
 use tokio::time;
 use crate::api_token::APIToken;
 use crate::dotnet::stop_dotnet_server;
 use crate::environment::{is_prod, CONFIG_DIRECTORY, DATA_DIRECTORY};
 use crate::log::switch_to_file_logging;
+use crate::metadata::META_DATA;
 
 /// The Tauri main window.
 static MAIN_WINDOW: Lazy<Mutex<Option<Window>>> = Lazy::new(|| Mutex::new(None));
@@ -39,6 +40,7 @@ pub fn start_tauri() {
             info!(Source = "Bootloader Tauri"; "Reconfigure the file logger to use the app data directory {data_path:?}");
             switch_to_file_logging(data_path).map_err(|e| error!("Failed to switch logging to file: {e}")).unwrap();
 
+            deploy_pdfium(app.path_resolver());
             Ok(())
         })
         .plugin(tauri_plugin_window_state::Builder::default().build())
@@ -319,4 +321,50 @@ pub struct PreviousFile {
 pub struct FileSelectionResponse {
     user_cancelled: bool,
     selected_file_path: String,
+}
+
+fn deploy_pdfium(path_resolver: PathResolver) {
+    info!(Source = "Bootloader Tauri"; "Deploy PDFium from the resources...");
+    let working_directory = std::env::current_dir().unwrap();
+    let pdfium_target_path = working_directory;
+    let metadata = &META_DATA;
+    let metadata = metadata.lock().unwrap();
+    let arch = metadata.clone().unwrap().architecture;
+    let pdfium_filename = match arch.as_str() {
+        "linux-x64" => Some("libpdfium.so"),
+        "linux-arm64" => Some("libpdfium.so"),
+
+        "win-x64" => Some("pdfium.dll"),
+        "win-arm64" => Some("pdfium.dll"),
+
+        "osx-x64" => Some("libpdfium.dylib"),
+        "osx-arm64" => Some("libpdfium.dylib"),
+
+        _ => None,
+    };
+
+    if pdfium_filename.is_none() {
+        error!(Source = "Bootloader Tauri"; "Failed to find the PDFium library for the current platform.");
+        return;
+    }
+
+    let pdfium_filename = pdfium_filename.unwrap();
+    let pdfium_relative_source_path = String::from("resources/libraries/") + pdfium_filename;
+    let pdfium_source_path = path_resolver.resolve_resource(pdfium_relative_source_path);
+    if pdfium_source_path.is_none() {
+        error!(Source = "Bootloader Tauri"; "Failed to find the PDFium library for the current platform.");
+        return;
+    }
+
+    let pdfium_source_path = pdfium_source_path.unwrap();
+    let pdfium_target_path = pdfium_target_path.join(pdfium_filename);
+
+    info!(Source = "Bootloader Tauri"; "Detected platform: {arch:?}, expected PDFium filename: {pdfium_filename:?}, source path: {pdfium_source_path:?}, target path: {pdfium_target_path:?}");
+
+    if let Err(e) = std::fs::copy(pdfium_source_path, pdfium_target_path) {
+        error!(Source = "Bootloader Tauri"; "Failed to copy the PDFium library for the current platform: {e}");
+        return;
+    }
+
+    info!(Source = "Bootloader Tauri"; "Successfully deployed PDFium.");
 }

runtime/src/file_data.rs

@@ -15,6 +15,7 @@ use rocket::response::stream::{EventStream, Event};
 use rocket::tokio::select;
 use rocket::serde::Serialize;
 use rocket::get;
+use crate::api_token::APIToken;
 
 #[derive(Debug, Serialize)]
 pub struct Chunk {
@@ -39,7 +40,7 @@ type Result<T> = std::result::Result<T, Box<dyn std::error::Error + Send + Sync>
 type ChunkStream = Pin<Box<dyn Stream<Item = Result<Chunk>> + Send>>;
 
 #[get("/retrieval/fs/extract?<path>")]
-pub async fn extract_data(path: String, mut end: Shutdown) -> EventStream![] {
+pub async fn extract_data(_token: APIToken, path: String, mut end: Shutdown) -> EventStream![] {
     EventStream! {
         let stream_result = stream_data(&path).await;
         match stream_result {
@@ -78,7 +79,7 @@ async fn stream_data(file_path: &str) -> Result<ChunkStream> {
         FileFormat::from_file(&file_path_clone)
     }).await??;
 
-    let ext = file_path.split('.').last().unwrap_or("");
+    let ext = file_path.split('.').next_back().unwrap_or("");
     let stream = match ext {
         DOCX | ODT => {
             let from = if ext == DOCX { "docx" } else { "odt" };
@@ -91,7 +92,7 @@ async fn stream_data(file_path: &str) -> Result<ChunkStream> {
         
         _ => match fmt.kind() {
             Kind::Document => match fmt {
-                FileFormat::PortableDocumentFormat => read_pdf(file_path).await?,
+                FileFormat::PortableDocumentFormat => stream_pdf(file_path).await?,
                 FileFormat::MicrosoftWordDocument => {
                     convert_with_pandoc(file_path, "docx", TO_MARKDOWN).await?
                 }
@@ -133,7 +134,7 @@ async fn stream_text_file(file_path: &str) -> Result<ChunkStream> {
     let mut line_number = 0;
 
     let stream = stream! {
-        while let Ok(Some(line)) = lines.next_line().await { // Korrektur hier
+        while let Ok(Some(line)) = lines.next_line().await {
             line_number += 1;
             yield Ok(Chunk {
                 content: line,
@@ -145,23 +146,47 @@ async fn stream_text_file(file_path: &str) -> Result<ChunkStream> {
     Ok(Box::pin(stream))
 }
 
-async fn read_pdf(file_path: &str) -> Result<ChunkStream> {
+#[get("/retrieval/fs/read/pdf?<file_path>")]
+pub fn read_pdf(_token: APIToken, file_path: String) -> String {
+    let pdfium = Pdfium::default();
+    let doc = match pdfium.load_pdf_from_file(&file_path, None) {
+        Ok(document) => document,
+        Err(e) => return e.to_string(),
+    };
+
+    let mut pdf_content = String::new();
+    for page in doc.pages().iter() {
+        let content = match page.text().map(|text_content| text_content.all()) {
+            Ok(content) => content,
+            Err(_) => {
+                continue
+            }
+        };
+
+        pdf_content.push_str(&content);
+        pdf_content.push_str("\n\n");
+    }
+
+    pdf_content
+}
+
+async fn stream_pdf(file_path: &str) -> Result<ChunkStream> {
     let path = file_path.to_owned();
     let (tx, rx) = mpsc::channel(10);
 
     tokio::task::spawn_blocking(move || {
         let pdfium = Pdfium::default();
         let doc = match pdfium.load_pdf_from_file(&path, None) {
-            Ok(d) => d,
+            Ok(document) => document,
             Err(e) => {
                 let _ = tx.blocking_send(Err(e.into()));
                 return;
             }
         };
 
-        for (i, page) in doc.pages().iter().enumerate() {
+        for (num_page, page) in doc.pages().iter().enumerate() {
             let content = match page.text().map(|t| t.all()) {
-                Ok(c) => c,
+                Ok(text_content) => text_content,
                 Err(e) => {
                     let _ = tx.blocking_send(Err(e.into()));
                     continue;
@@ -170,7 +195,7 @@ async fn read_pdf(file_path: &str) -> Result<ChunkStream> {
 
             if tx.blocking_send(Ok(Chunk {
                 content,
-                metadata: Metadata::Pdf { page_number: i + 1 },
+                metadata: Metadata::Pdf { page_number: num_page + 1 },
             })).is_err() {
                 break;
             }

runtime/src/lib.rs

@@ -9,4 +9,5 @@ pub mod secret;
 pub mod clipboard;
 pub mod runtime_api;
 pub mod certificate;
-pub mod file_data;
+pub mod file_data;
+pub mod metadata;

runtime/src/main.rs

@@ -10,6 +10,7 @@ use mindwork_ai_studio::certificate::{generate_certificate};
 use mindwork_ai_studio::dotnet::start_dotnet_server;
 use mindwork_ai_studio::environment::is_dev;
 use mindwork_ai_studio::log::init_logging;
+use mindwork_ai_studio::metadata::{MetaData, META_DATA};
 use mindwork_ai_studio::runtime_api::start_runtime_api;
 
 #[tokio::main]
@@ -26,6 +27,22 @@ async fn main() {
     let tauri_version = metadata_lines.next().unwrap();
     let app_commit_hash = metadata_lines.next().unwrap();
     let architecture = metadata_lines.next().unwrap();
+    let pdfium_version = metadata_lines.next().unwrap();
+
+    let metadata = MetaData {
+        architecture: architecture.to_string(),
+        app_commit_hash: app_commit_hash.to_string(),
+        app_version: app_version.to_string(),
+        build_number: build_number.to_string(),
+        build_time: build_time.to_string(),
+        dotnet_sdk_version: dotnet_sdk_version.to_string(),
+        dotnet_version: dotnet_version.to_string(),
+        mud_blazor_version: mud_blazor_version.to_string(),
+        rust_version: rust_version.to_string(),
+        tauri_version: tauri_version.to_string(),
+    };
+
+    *META_DATA.lock().unwrap() = Some(metadata);
 
     init_logging();
     info!("Starting MindWork AI Studio:");
@@ -40,6 +57,7 @@ async fn main() {
     info!(".. Rust: v{rust_version}");
     info!(".. MudBlazor: v{mud_blazor_version}");
     info!(".. Tauri: v{tauri_version}");
+    info!(".. PDFium: v{pdfium_version}");
 
     if is_dev() {
         warn!("Running in development mode.");

runtime/src/metadata.rs

@@ -0,0 +1,18 @@
+use std::sync::Mutex;
+use once_cell::sync::Lazy;
+
+pub static META_DATA: Lazy<Mutex<Option<MetaData>>> = Lazy::new(|| Mutex::new(None));
+
+#[derive(Clone)]
+pub struct MetaData {
+    pub app_version: String,
+    pub build_time: String,
+    pub build_number: String,
+    pub dotnet_sdk_version: String,
+    pub dotnet_version: String,
+    pub rust_version: String,
+    pub mud_blazor_version: String,
+    pub tauri_version: String,
+    pub app_commit_hash: String,
+    pub architecture: String,
+}

runtime/src/runtime_api.rs

@@ -79,6 +79,7 @@ pub fn start_runtime_api() {
                 crate::environment::get_config_directory,
                 crate::environment::read_user_language,
                 crate::file_data::extract_data,
+                crate::file_data::read_pdf,
                 crate::log::get_log_paths,
             ])
             .ignite().await.unwrap()

runtime/tauri.conf.json

@@ -29,6 +29,9 @@
         "scope": [
           "http://localhost"
         ]
+      },
+      "fs": {
+        "scope": ["$RESOURCE/resources/*"]
       }
     },
     "windows": [
@@ -57,6 +60,9 @@
       "externalBin": [
         "../app/MindWork AI Studio/bin/dist/mindworkAIStudioServer"
       ],
+      "resources": [
+        "resources/*"
+      ],
       "macOS": {
         "exceptionDomain": "localhost"
       },