+ prev coommit

game/.gitignore

@@ -1,4 +1,5 @@
 /target
 .cargo
 .vscode
-/data
+/data
+/Microsoft.WebView2.FixedVersionRuntime.107.0.1418.24.x64

game/Cargo.toml

@@ -21,7 +21,7 @@ serde_json = "1.0"
 serde-pickle = "1.1.1"
 serde = { version = "1.0.144", features = ["derive"] }
 tauri = { version = "1.1.1", features = ["api-all"] }
-pyo3 = { version = "0.17.1", features = [] }
+tempfile = "3.3.0"
 num_cpus = "1.13.1"
 crossbeam = "0.8.2"
 flume = "0.10.14"
@@ -31,6 +31,8 @@ rand = "0.8.5"
 rand_distr = "0.4.3"
 probability = "0.18.0"
 
+pyo3 = { version = "0.17.1", features = [],  optional = true }
+
 [dependencies.uuid]
 version = "1.1.2"
 features = [
@@ -41,16 +43,9 @@ features = [
 
 
 [build-dependencies]
-tauri-build = { version = "1.1.1" }
+tauri-build = { version = "1.1.1", features = [] }
 copy_to_output = "2.0.0"
 
-[features]
-# by default Tauri runs in production mode
-# when `tauri dev` runs it is executed with `cargo run --no-default-features` if `devPath` is an URL
-default = ["custom-protocol"]
-# this feature is used used for production builds where `devPath` points to the filesystem
-# DO NOT remove this
-custom-protocol = ["tauri/custom-protocol"]
 
 
 # BLAS source choices:
@@ -70,8 +65,19 @@ blas-src = { version = "0.8.0", default-features = false, features = [
 ndarray = { version = "0.15.6" }
 
 
-#[profile.release]
-#strip = true      # Automatically strip symbols from the binary.
-#opt-level = "z"   # Optimize for size.
-#lto = true
-#codegen-units = 1
+[profile.release]
+strip = true      # Automatically strip symbols from the binary.
+opt-level = "z"   # Optimize for size.
+lto = true
+codegen-units = 1
+
+
+
+[features]
+# by default Tauri runs in production mode
+# when `tauri dev` runs it is executed with `cargo run --no-default-features` if `devPath` is an URL
+default = ["custom-protocol"]
+# this feature is used used for production builds where `devPath` points to the filesystem
+# DO NOT remove this
+custom-protocol = ["tauri/custom-protocol"]
+train = [ "dep:pyo3" ]

game/build_windows.bat

@@ -1,3 +1,5 @@
+REM cargo install tauri-cli
+
 CALL "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvars64.bat"
 SET LIBCLANG_PATH=C:\Program Files\Microsoft^ Visual^ Studio\2022\Enterprise\VC\Tools\Llvm\x64\bin
 
@@ -7,3 +9,4 @@ SET TFLITEC_PREBUILT_PATH_X86_64_PC_WINDOWS_GNU=%CD%\tensorflowlite_c.dll
 SET RUSTFLAGS=-C target-feature=+crt-static
 cargo build --release
 
+cargo tauri build 

game/src/game.rs

@@ -48,7 +48,7 @@ impl From<u8> for Color {
 }
 
 impl From<f32> for Color {
-    fn from(v: f32) -> Self {
+    fn from(_v: f32) -> Self {
         unimplemented!()
     }
 }

game/src/human.rs

@@ -21,12 +21,15 @@ use crate::model::TfLiteModel;
 use crossbeam::atomic::AtomicCell;
 
 use std::sync::atomic::Ordering;
-
+use std::include_bytes;
 use rand::seq::SliceRandom;
 use std::cell::RefCell;
 use std::sync::{atomic::AtomicBool, Arc};
 use tokio::sync::oneshot::{self, Receiver};
 use tokio::task::JoinHandle;
+use tempfile::{tempdir, TempDir};
+use std::fs::File;
+use std::io::prelude::*;
 
 lazy_static! {
     static ref SINGLETON_CHANNEL: AtomicCell<Receiver<HumanVsMachineMatch>> = {
@@ -37,6 +40,19 @@ lazy_static! {
         }
         AtomicCell::new(rx)
     };
+
+    static ref MODEL_FILE : (String, TempDir) = {
+        let mut dir = tempdir().expect("Unable to create template file");
+        let filename = dir.path().join("model.tflite").display().to_string();
+        println!("{}", &filename);
+
+        let mut file = File::create(&filename).expect("Unable to create temp file");
+
+        let bytes = include_bytes!("../best.tflite");
+        file.write_all(bytes).expect("Unable to save model to file");
+        file.flush().expect("Unable to flush model to file");
+        ( filename, dir )
+    };
 }
 
 pub async fn access<Fut, F>(mut f: F) -> MatchState
@@ -96,7 +112,7 @@ impl BoardInfo {
 impl HumanVsMachineMatch {
     fn new(human_play_black: bool) -> Self {
         let ai_player = AiPlayer::new();
-        let max_threads = 2; //(num_cpus::get() - 0).max(1);
+        let max_threads = 1; //(num_cpus::get() - 0).max(1);
         Self {
             ai_player: Arc::new(ai_player),
             board: RenjuBoard::default(),
@@ -268,27 +284,26 @@ impl AiPlayer {
         choices: &Vec<(usize, usize)>,
     ) {
         self.tree
-            .rollout(
-                breadth_first,
-                board,
-                choices,
-                |state_tensor: StateTensor| {
-                    MODEL.with(|ref_cell| {
-                        let mut model = ref_cell.borrow_mut();
-                        if model.is_none() {
-                            *model = Some(
-                                TfLiteModel::load("best.tflite")
-                                    .expect("Unable to load saved model"),
-                            )
-                        }
-                        model
-                            .as_ref()
-                            .unwrap()
-                            .predict_one(state_tensor)
-                            .expect("Unable to predict_one")
-                    })
-                },
-            )
+            .rollout(breadth_first, board, choices, |state_tensor: StateTensor| {
+                MODEL.with(|ref_cell| {
+                    let mut model = ref_cell.borrow_mut();
+                    if model.is_none() {
+                        #[cfg(feature="train")]
+                        let filepath = "best.tflite";
+
+                        #[cfg(not(feature="train"))]
+                        let filepath = &MODEL_FILE.0;
+                        *model = Some(
+                            TfLiteModel::load(filepath).expect("Unable to load model"),
+                        )
+                    }
+                    model
+                        .as_ref()
+                        .unwrap()
+                        .predict_one(state_tensor)
+                        .expect("Unable to predict_one")
+                })
+            })
             .await
             .expect("rollout failed")
     }

game/src/lib.rs

@@ -20,4 +20,5 @@ pub mod model;
 
 pub use game::{RenjuBoard, SquareMatrix, SquaredMatrixExtension, StateTensor, TerminalState};
 pub use mcts::{MonteCarloTree, TreeNode};
+#[cfg(feature="train")]
 pub use model::PolicyValueModel;

game/src/main.rs

@@ -31,11 +31,13 @@ mod game;
 mod human;
 mod mcts;
 mod model;
+#[cfg(feature="train")]
 mod selfplay;
 use human::MatchState;
+#[cfg(feature="train")]
 use selfplay::Trainer;
 
-static ABOUT_TEXT: &str = "Renju game ";
+static ABOUT_TEXT: &str = "Renju Game";
 
 static SELF_PLAY_MATCH_HELP_TEXT: &str = "
 Produce matches by self-play
@@ -65,28 +67,20 @@ enum Verb {
         export_dir: String,
     },
 
+    #[cfg(feature="train")]
     /// Self play and train
     #[clap(after_help=TRAIN_HELP_TEXT)]
     Train {},
 
-    /// contest between two model
-    #[clap()]
-    Contest {
-        /// old model name
-        #[clap(required = true)]
-        old_model: String,
 
-        /// new model name
-        #[clap(required = true)]
-        new_model: String,
-    },
 }
 
 #[tokio::main(flavor = "multi_thread")]
 async fn main() {
     let args = Arguments::parse();
 
     match args.verb {
+        #[cfg(feature="train")]
         Some(Verb::SelfPlay {
             model_file,
             export_dir,
@@ -98,6 +92,7 @@ async fn main() {
                 .await;
         }
 
+        #[cfg(feature="train")]
         Some(Verb::Train {}) => {
             let mut trainer = Trainer::new();
             trainer.run().await;

game/src/model.rs

@@ -15,21 +15,26 @@
  */
 use bytes::Bytes;
 use num_cpus;
+#[cfg(feature="train")]
 use pyo3::prelude::*;
+#[cfg(feature="train")]
 use pyo3::types::{PyBytes, PyList, PyTuple};
+#[cfg(feature="train")]
 use std::fs::OpenOptions;
+#[cfg(feature="train")]
 use std::io::prelude::*;
 
 use tflitec::interpreter::{Interpreter, Options};
 use tflitec::tensor;
 
 use crate::game::*;
 
+#[cfg(feature="train")]
 pub struct PolicyValueModel {
     module: Py<PyModule>,
 }
 
-//
+#[cfg(feature="train")]
 impl PolicyValueModel {
     pub fn new(filename: &str) -> PyResult<Self> {
         let mut source_code = String::new();
@@ -185,7 +190,7 @@ impl TfLiteModel {
     pub fn load(tflite_model_path: &str) -> Result<Self, tflitec::Error> {
         // Create interpreter options
         let mut options = Options::default();
-        options.thread_count = num_cpus::get() as i32 / 2;
+        options.thread_count = (num_cpus::get() as i32 - 2).max(1);
         options.is_xnnpack_enabled = true;
         println!("is_xnnpack_enabled={}", options.is_xnnpack_enabled);
 

game/tauri.conf.json

@@ -1,12 +1,12 @@
 {
   "build": {
-    "beforeBuildCommand": "npm run build --prefix=ui",
+    "beforeBuildCommand": "npm run build",
     "beforeDevCommand": "npm run build --prefix=ui",
     "devPath": "http://localhost:8080",
     "distDir": "ui/dist"
   },
   "package": {
-    "productName": "renju_board",
+    "productName": "renju",
     "version": "0.1.0"
   },
   "tauri": {
@@ -15,7 +15,7 @@
     },
     "bundle": {
       "active": true,
-      "category": "DeveloperTool",
+      "category": "games",
       "copyright": "",
       "deb": {
         "depends": []
@@ -28,8 +28,8 @@
         "icons/icon.icns",
         "icons/icon.ico"
       ],
-      "identifier": "com.tauri.dev",
-      "longDescription": "",
+      "identifier": "com.github.wangjia184.renju",
+      "longDescription": "Renju Game https://github.com/wangjia184/renju",
       "macOS": {
         "entitlements": null,
         "exceptionDomain": "",
@@ -38,12 +38,16 @@
         "signingIdentity": null
       },
       "resources": [],
-      "shortDescription": "",
+      "shortDescription": "https://github.com/wangjia184/renju",
       "targets": "all",
       "windows": {
         "certificateThumbprint": null,
         "digestAlgorithm": "sha256",
-        "timestampUrl": ""
+        "timestampUrl": "",
+        "webviewInstallMode": {
+          "type": "fixedRuntime",
+          "path": "./Microsoft.WebView2.FixedVersionRuntime.107.0.1418.24.x64/"
+        }
       }
     },
     "security": {