WIP Rework markdown parsing

Cargo.lock

@@ -1730,6 +1730,15 @@ dependencies = [
  "windows-targets 0.48.5",
 ]
 
+[[package]]
+name = "getopts"
+version = "0.2.21"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "14dbbfd5c71d70241ecf9e6f13737f7b5ce823821063188d7e46c41d371eebd5"
+dependencies = [
+ "unicode-width",
+]
+
 [[package]]
 name = "getrandom"
 version = "0.1.16"
@@ -2869,7 +2878,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "fc2f4eb4bc735547cfed7c0a4922cbd04a4655978c09b54f1f7b228750664c34"
 dependencies = [
  "cfg-if",
- "windows-targets 0.48.5",
+ "windows-targets 0.52.6",
 ]
 
 [[package]]
@@ -2955,6 +2964,7 @@ dependencies = [
  "mime_guess",
  "once_cell",
  "pin-project-lite",
+ "pulldown-cmark",
  "rand 0.8.5",
  "regex",
  "reqwest",
@@ -4039,6 +4049,26 @@ dependencies = [
  "psl-types",
 ]
 
+[[package]]
+name = "pulldown-cmark"
+version = "0.12.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f86ba2052aebccc42cbbb3ed234b8b13ce76f75c3551a303cb2bcffcff12bb14"
+dependencies = [
+ "bitflags 2.8.0",
+ "getopts",
+ "memchr",
+ "pulldown-cmark-escape",
+ "serde",
+ "unicase",
+]
+
+[[package]]
+name = "pulldown-cmark-escape"
+version = "0.11.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "007d8adb5ddab6f8e3f491ac63566a7d5002cc7ed73901f72057943fa71ae1ae"
+
 [[package]]
 name = "quick-xml"
 version = "0.37.2"
@@ -6166,6 +6196,12 @@ version = "1.12.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "f6ccf251212114b54433ec949fd6a7841275f9ada20dddd2f29e9ceea4501493"
 
+[[package]]
+name = "unicode-width"
+version = "0.1.14"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7dd6e30e90baa6f72411720665d41d89b9a3d039dc45b8faea1ddd07f617f6af"
+
 [[package]]
 name = "unicode-xid"
 version = "0.2.6"

llama_forge_rs/Cargo.toml

@@ -83,6 +83,7 @@ tracing-test = "0.2.4"
 sysinfo = { version = "0.30.11", optional = true }
 derive_more = { version = "0.99.17", features = ["nightly"] }
 sqlx-macros = { version = "0.7.4", optional = true, features = ["chrono", "json", "migrate", "sqlite", "uuid"] }
+pulldown-cmark = { version = "0.12.2", features = ["serde"] }
 # qdrant-client = "1.11.2"
 # swiftide = "0.9.1"
 

llama_forge_rs/src/api/backend_process.rs

@@ -48,11 +48,15 @@ pub async fn run_starter_task(pool: sqlx::SqlitePool) {
     use crate::server::backends::BackendService;
     let _ = tracing::debug_span!("starter_task");
 
+    tracing::debug!("AAAAAAAAAAAAAAA");
+    return; // TODO ????
     tracing::debug!("Starter task started");
 
     let service_handle = BackendService::new();
     let mut stream = IntervalStream::new(time::interval(Duration::from_millis(1000)));
     while let Some(instant) = stream.next().await {
+        break; // TODO integrate proxy man ?
+
         tracing::debug!("fire; instant={:?}", instant);
 
         let waiting_to_start: Vec<BackendProcess> = sqlx::query_as(

llama_forge_rs/src/app/mod.rs

@@ -41,7 +41,8 @@ pub fn App() -> impl IntoView {
                 <Routes>
                     <Route path="" view=MainPage>
                         <Route path="/chat" view=ChatPage/>
-                        <SettingsRoutes/>
+                        // TODO make settings page for proxy-man
+                        // <SettingsRoutes/>
                     </Route>
                 </Routes>
             </main>

llama_forge_rs/src/app/pages/chat.rs

@@ -12,7 +12,7 @@ use crate::{
     app::components::{svgs::*, Card},
 };
 
-#[component]
+#[island]
 fn ChatMessageBubble(
     msg: RwSignal<ChatMessage>,
     history: RwSignal<Vec<RwSignal<ChatMessage>>>,
@@ -33,12 +33,37 @@ fn ChatMessageBubble(
 
     let textarea_ref = NodeRef::<html::P>::new();
 
+    use pulldown_cmark;
     let editable_p = move || {
+        // TODO Convert back to raw str when editable
         let mode = if edit_mode.get() { "true" } else { "false" };
+        let msg_str = move || msg.get().content.clone();
+        let md_str = move || {
+            let owned_str = msg_str();
+            let parser = pulldown_cmark::Parser::new(&owned_str);
+            let mut md_output = String::new();
+            pulldown_cmark::html::push_html(&mut md_output, parser);
+
+            md_output
+        };
+
+        let inner_p = move || {
+            if edit_mode.get() {
+                view! {
+
+                    <p inner_html=move || { msg_str() }></p>
+                }
+            } else {
+                view! {
+                    <p inner_html=move || { md_str() }></p>
+                }
+            }
+        };
+
         view! {
             <p node_ref=textarea_ref contenteditable=move || { mode }>
 
-                <span inner_html=msg.get().content.clone()></span>
+                {inner_p}
             </p>
         }
     };

llama_forge_rs/src/server/backends/llama_chat.rs

@@ -11,16 +11,16 @@ use crate::api::{ChannelMessage, Chat, ChatMessage};
 
 #[derive(Serialize, Debug)]
 struct LlamaChatCompletionRequest {
-    stream:   bool,
+    stream: bool,
-    model:    String,
+    model: String,
     messages: Vec<LlamaChatMessage>,
 }
 
 impl From<Chat> for LlamaChatCompletionRequest {
     fn from(value: Chat) -> Self {
         Self {
-            stream:   true,
+            stream: true,
-            model:    "default".to_string(),
+            model: "default".to_string(),
             messages: value.history.into_iter().map(|e| e.into()).collect(),
         }
     }
@@ -28,14 +28,14 @@ impl From<Chat> for LlamaChatCompletionRequest {
 
 #[derive(Serialize, Debug)]
 struct LlamaChatMessage {
-    role:    String,
+    role: String,
     content: String,
 }
 
 impl From<ChatMessage> for LlamaChatMessage {
     fn from(chat_message: ChatMessage) -> Self {
         Self {
-            role:    chat_message.role.into(),
+            role: chat_message.role.into(),
             content: chat_message.content,
         }
     }
@@ -68,9 +68,7 @@ pub struct LlamaService {
 
 impl LlamaService {
     pub fn new(id: Uuid) -> Self {
-        Self {
+        Self { id }
-            id,
-        }
     }
 }
 
@@ -80,7 +78,8 @@ async fn do_chat_request(chat: Chat, sender: mpsc::Sender<ChannelMessage>) -> an
     let request_body: LlamaChatCompletionRequest = chat.into();
 
     let request_builder = client
-        .post("http://localhost:8080/v1/chat/completions")
+        // # .post("http://localhost:8080/v1/chat/completions")
+        .post("http://100.64.0.3:18080/v1/chat/completions")
         .header("Content-Type", "application/json")
         .json(&request_body);
 
@@ -92,7 +91,19 @@ async fn do_chat_request(chat: Chat, sender: mpsc::Sender<ChannelMessage>) -> an
             Ok(Event::Message(event)) => match event.event.as_str() {
                 "message" => {
                     let data = event.data;
-                    let response: LlamaChatResponse = serde_json::from_str(&data).unwrap();
+                    tracing::debug!(?data);
+                    if data == "[DONE]" {
+                        sender
+                            .send(ChannelMessage::Stop)
+                            .await
+                            .expect("channel fail");
+
+                        es.close();
+
+                        break;
+                    }
+
+                    let response: LlamaChatResponse = serde_json::from_str(&data).expect("no json");
 
                     for choice in response.choices.into_iter() {
                         if let Some(delta) = choice.delta {

llama_forge_rs/src/server/backends/llama_completion.rs

@@ -112,7 +112,7 @@ pub async fn do_completion_request() -> Result<(), Box<dyn std::error::Error>> {
     let request_body = CompletionRequest::default();
 
     let request_builder = client
-        .post("http://localhost:8080/completion")
+        .post("http://100.64.0.3:18080/completion")
         .header("Accept", "text/event-stream")
         .header("Content-Type", "application/json")
         .header("User-Agent", "llama_forge_rs")

llama_forge_rs/src/server/backends/mod.rs

@@ -177,10 +177,7 @@ mod tests {
     use crate::{
         api::{ChannelMessage, ChatMessage, ChatRole},
         server::backends::{
-            llama_chat::LlamaService,
+            llama_chat::LlamaService, BackendService, BackendServiceStatus, ChatService,
-            BackendService,
-            BackendServiceStatus,
-            ChatService,
         },
     };
 

llama_forge_rs/src/server/backends/runner.rs

@@ -4,7 +4,7 @@ use serde::{Deserialize, Serialize};
 use tokio::process::Command;
 
 pub struct RunnerArgs {
-    ctx_size:   i64,
+    ctx_size: i64,
     gpu_layers: i64,
     model_path: String,
 }
@@ -36,8 +36,8 @@ impl From<RunnerArgs> for Vec<String> {
 
 #[derive(Debug, Serialize, Deserialize, Clone)]
 pub struct Runner {
-    pwd:  Option<String>,
+    pwd: Option<String>,
-    cmd:  String,
+    cmd: String,
     args: Vec<String>,
 }
 
@@ -45,8 +45,8 @@ impl Runner {
     // FIXME does not exit properly when it is killed
     pub fn new_llamafile_bin(runner_args: RunnerArgs) -> Self {
         Self {
-            pwd:  None,
+            pwd: None,
-            cmd:  "bash".to_string(),
+            cmd: "bash".to_string(),
             args: vec![
                 format!(
                     "{}/llamafile",
@@ -64,8 +64,8 @@ impl Runner {
 
     pub fn new_llama_server_bin(runner_args: RunnerArgs) -> Self {
         Self {
-            pwd:  None,
+            pwd: None,
-            cmd:  "llama-server".to_string(),
+            cmd: "llama-server".to_string(),
             args: runner_args.into(),
         }
     }

llama_forge_rs/src/server/middleware.rs

@@ -18,9 +18,7 @@ impl<S> Layer<S> for LoggingLayer {
     type Service = LoggingService<S>;
 
     fn layer(&self, inner: S) -> Self::Service {
-        LoggingService {
+        LoggingService { inner }
-            inner,
-        }
     }
 }
 
@@ -50,8 +48,8 @@ where
 
         LoggingServiceFuture {
             inner: self.inner.call(req),
-            uuid:  Arc::new(request_uuid), // Store UUID in an Arc for shared ownership
+            uuid: Arc::new(request_uuid), // Store UUID in an Arc for shared ownership
-            span:  Arc::new(span),
+            span: Arc::new(span),
         }
     }
 }

llama_forge_rs/src/server/mod.rs

@@ -6,26 +6,20 @@ use axum::{
     http::Request,
     response::IntoResponse,
     routing::get,
-    Extension,
+    Extension, Router,
-    Router,
 };
 use leptos::*;
 use leptos_axum::{generate_route_list, handle_server_fns_with_context, LeptosRoutes};
 use leptos_router::RouteListing;
 use sqlx::{
     sqlite::{SqliteConnectOptions, SqliteJournalMode, SqlitePoolOptions, SqliteSynchronous},
-    ConnectOptions,
+    ConnectOptions, SqlitePool,
-    SqlitePool,
 };
 use tower::Layer;
 use tower_http::{
     compression::CompressionLayer,
     trace::{
-        DefaultMakeSpan,
+        DefaultMakeSpan, DefaultOnEos, DefaultOnFailure, DefaultOnRequest, DefaultOnResponse,
-        DefaultOnEos,
-        DefaultOnFailure,
-        DefaultOnRequest,
-        DefaultOnResponse,
         TraceLayer,
     },
     CompressionLevel,
@@ -122,7 +116,7 @@ pub async fn app(leptos_options: LeptosOptions) -> Router {
 
     let pool = new_pool().await.expect("pool err");
 
-    // TODO move this out of server(pool has to be moved out too)
+    // // TODO move this out of server(pool has to be moved out too)
     let task = run_starter_task(pool.clone());
     tokio::task::spawn(task);