diff --git a/.vscode/settings.json b/.vscode/settings.json
new file mode 100755
index 00000000..c7cd746d
--- /dev/null
+++ b/.vscode/settings.json
@@ -0,0 +1,91 @@
+{
+    "files.associations": {
+        "*.ipp": "cpp",
+        "memory": "cpp",
+        "regex": "cpp",
+        "utility": "cpp",
+        "__bit_reference": "cpp",
+        "__config": "cpp",
+        "__debug": "cpp",
+        "__errc": "cpp",
+        "__functional_base": "cpp",
+        "__hash_table": "cpp",
+        "__locale": "cpp",
+        "__mutex_base": "cpp",
+        "__node_handle": "cpp",
+        "__nullptr": "cpp",
+        "__split_buffer": "cpp",
+        "__string": "cpp",
+        "__threading_support": "cpp",
+        "__tree": "cpp",
+        "__tuple": "cpp",
+        "algorithm": "cpp",
+        "array": "cpp",
+        "atomic": "cpp",
+        "bit": "cpp",
+        "bitset": "cpp",
+        "cctype": "cpp",
+        "chrono": "cpp",
+        "clocale": "cpp",
+        "cmath": "cpp",
+        "codecvt": "cpp",
+        "complex": "cpp",
+        "condition_variable": "cpp",
+        "csignal": "cpp",
+        "cstdarg": "cpp",
+        "cstddef": "cpp",
+        "cstdint": "cpp",
+        "cstdio": "cpp",
+        "cstdlib": "cpp",
+        "cstring": "cpp",
+        "ctime": "cpp",
+        "cwchar": "cpp",
+        "cwctype": "cpp",
+        "deque": "cpp",
+        "exception": "cpp",
+        "forward_list": "cpp",
+        "fstream": "cpp",
+        "functional": "cpp",
+        "future": "cpp",
+        "initializer_list": "cpp",
+        "iomanip": "cpp",
+        "ios": "cpp",
+        "iosfwd": "cpp",
+        "iostream": "cpp",
+        "istream": "cpp",
+        "iterator": "cpp",
+        "limits": "cpp",
+        "list": "cpp",
+        "locale": "cpp",
+        "map": "cpp",
+        "mutex": "cpp",
+        "new": "cpp",
+        "numeric": "cpp",
+        "optional": "cpp",
+        "ostream": "cpp",
+        "queue": "cpp",
+        "random": "cpp",
+        "ratio": "cpp",
+        "set": "cpp",
+        "sstream": "cpp",
+        "stack": "cpp",
+        "stdexcept": "cpp",
+        "streambuf": "cpp",
+        "string": "cpp",
+        "string_view": "cpp",
+        "system_error": "cpp",
+        "thread": "cpp",
+        "tuple": "cpp",
+        "type_traits": "cpp",
+        "typeindex": "cpp",
+        "typeinfo": "cpp",
+        "unordered_map": "cpp",
+        "unordered_set": "cpp",
+        "valarray": "cpp",
+        "variant": "cpp",
+        "vector": "cpp",
+        "__functional_03": "cpp",
+        "filesystem": "cpp"
+    },
+    "python.pythonPath": "/usr/local/Caskroom/miniconda/base/envs/moRL/bin/python"
+}
\ No newline at end of file
diff --git a/README.md b/README.md
old mode 100644
new mode 100755
diff --git a/elf/ai.h b/elf/ai.h
old mode 100644
new mode 100755
index a2658460..8d8b37d2
--- a/elf/ai.h
+++ b/elf/ai.h
@@ -17,7 +17,7 @@ namespace elf {
 
 using namespace std;
 
-template <typename S, typename A>
+template <typename S, typename A>  // RTSState RTSMCAction
 class AI_T {
 public:
     using Action = A;
@@ -43,6 +43,8 @@ class AI_T {
 
     virtual ~AI_T() { }
 
+    void Print(){std::cout<<"AIName: _name: "<<_name<<" id: "<<_id<<std::endl;}
+
 private:
     const std::string _name;
     int _id;
@@ -51,7 +53,7 @@ class AI_T {
     virtual void on_set_id() { }
 };
 
-template <typename S, typename A, typename AIComm>
+template <typename S, typename A, typename AIComm>   //带有AICommT的AI
 class AIWithCommT : public AI_T<S, A> {
 public:
     using AI = AI_T<S, A>;
diff --git a/elf/comm_template.h b/elf/comm_template.h
old mode 100644
new mode 100755
index 26a2a110..3224898f
--- a/elf/comm_template.h
+++ b/elf/comm_template.h
@@ -200,14 +200,16 @@ class CommT {
             _signal->use_queue_per_group(_groups.size());
         }
 
+        //std::cout<<"-------CommT CollectorsReady---------------"<<std::endl;
         for (auto &p : _map) {
             p.second.InitCond(_exclusive_groups.size());
         }
-
+        // std::cout<<"pool_size 1: "<<_pool.size()<<std::endl;  // 初始为 0
         _pool.resize(_groups.size());
+        //std::cout<<"pool_size 2: "<<_pool.size()<<std::endl;  // 设为16
         for (auto &g : _groups) {
           CollectorGroup *p = g.get();
-          _pool.push([p, this](int) { p->MainLoop(); });
+          _pool.push([p, this](int) { p->MainLoop(); });  // 16 个 Batch Collector
         }
     }
 
@@ -343,13 +345,15 @@ class ContextT {
     const Options &options() const { return _options; }
 
     void Start(GameStartFunc game_start_func) {
-        _comm.CollectorsReady();
-
+        std::cout<<"--------ContextT Start---------------"<<std::endl;
+        _comm.CollectorsReady();  //设置 BatchCollector
+        
+        std::cout<<"_pool.size"<<_pool.size()<<std::endl;  // 1024 - num_games
         // Now we start all jobs.
         for (int i = 0; i < _pool.size(); ++i) {
             _pool.push([i, this, &game_start_func](int){
                 elf::Signal signal(_done.flag(), _prepare_stop);
-                game_start_func(i, _context_options, _options, signal, &_comm);
+                game_start_func(i, _context_options, _options, signal, &_comm);  //每个线程调用一次GameStartFunc
                 // std::cout << "G[" << i << "] is ending" << std::endl;
                 _done.notify();
             });
diff --git a/elf/context_utils.py b/elf/context_utils.py
index 85117f52..c319269b 100644
--- a/elf/context_utils.py
+++ b/elf/context_utils.py
@@ -15,22 +15,22 @@ def __init__(self):
             call_from = self,
             define_args = [
                 ("num_games", 1024),
-                ("batchsize", 128),
-                ("game_multi", dict(type=int, default=None)),
-                ("T", 6),
-                ("eval", dict(action="store_true")),
-                ("wait_per_group", dict(action="store_true")),
-                ("num_collectors", 0),
-                ("verbose_comm", dict(action="store_true")),
-                ("verbose_collector", dict(action="store_true")),
-                ("mcts_threads", 0),
+                ("batchsize", 128),  # 64
+                ("game_multi", dict(type=int, default=None)),  
+                ("T", 6),  # 20
+                ("eval", dict(action="store_true")),  # False
+                ("wait_per_group", dict(action="store_true")), # False
+                ("num_collectors", 0), 
+                ("verbose_comm", dict(action="store_true")), # False
+                ("verbose_collector", dict(action="store_true")), # False
+                ("mcts_threads", 0),  
                 ("mcts_rollout_per_thread", 1),
-                ("mcts_verbose", dict(action="store_true")),
-                ("mcts_save_tree_filename", ""),
-                ("mcts_verbose_time", dict(action="store_true")),
+                ("mcts_verbose", dict(action="store_true")), # False
+                ("mcts_save_tree_filename", ""), 
+                ("mcts_verbose_time", dict(action="store_true")), # False
 
-                ("mcts_use_prior", dict(action="store_true")),
-                ("mcts_pseudo_games", 0),
+                ("mcts_use_prior", dict(action="store_true")), # False
+                ("mcts_pseudo_games", 0), 
                 ("mcts_pick_method", "most_visited"),
             ],
             on_get_args = self._on_get_args
diff --git a/elf/game_base.h b/elf/game_base.h
old mode 100644
new mode 100755
index 6c397efe..b7cd36fd
--- a/elf/game_base.h
+++ b/elf/game_base.h
@@ -78,8 +78,14 @@ class GameBaseT {
         return res;
     }
 
-    void MainLoop(const std::atomic_bool *done = nullptr) {
-        _state->Init();
+    void MainLoop(const std::atomic_bool *done = nullptr,bool isPrint = false) {
+        if(isPrint)
+         std::cout<<"-------MainLoop----------"<<std::endl;
+        _state->Init(isPrint);  // 初始化游戏
+        // if(isPrint){
+        //     std::cout<<"--------Start PleyerInfo"<<std::endl;
+        //     std::cout<<_state->env().PrintPlayerInfo()<<std::endl;
+        // }
         while (true) {
             if (Step(done) != GAME_NORMAL) break;
             if (done != nullptr && done->load()) break;
@@ -87,6 +93,10 @@ class GameBaseT {
         // Send message to AIs.
         _act(false, done);
         _game_end();
+        // if(isPrint){
+        //     std::cout<<"--------End PleyerInfo"<<std::endl;
+        //     std::cout<<_state->env().PrintPlayerInfo()<<std::endl;
+        // }
         _state->Finalize();
     }
 
@@ -122,6 +132,7 @@ class GameBaseT {
             bot.ai->GameEnd();
         }
         if (_spectator != nullptr) {
+            std::cout<<"_spectator"<<std::endl;
             _spectator->GameEnd();
         }
     }
diff --git a/elf/python_options_utils_cpp.h b/elf/python_options_utils_cpp.h
index 7063182a..20cfd76e 100644
--- a/elf/python_options_utils_cpp.h
+++ b/elf/python_options_utils_cpp.h
@@ -19,13 +19,13 @@
 
 struct ContextOptions {
     // How many simulation threads we are running.
-    int num_games = 1;
+    int num_games = 1;  // 1024
 
     // The maximum number of threads per game
-    int max_num_threads = 0;
+    int max_num_threads = 0;  //0
 
     // History length. How long we should keep the history.
-    int T = 1;
+    int T = 1; // 20
 
     // verbose options.
     bool verbose_comm = false;
@@ -34,7 +34,7 @@ struct ContextOptions {
     // Whether we wait for each group or we wait jointly.
     bool wait_per_group = false;
 
-    int num_collectors = 1;
+    int num_collectors = 1;  // 0
 
     mcts::TSOptions mcts_options;
 
diff --git a/elf/tree_search_options.h b/elf/tree_search_options.h
index 4813786e..a1bf8b0f 100644
--- a/elf/tree_search_options.h
+++ b/elf/tree_search_options.h
@@ -20,12 +20,12 @@ using namespace std;
 
 struct TSOptions {
     int max_num_moves = 0;
-    int num_threads = 16;
-    int num_rollout_per_thread = 100;
-    bool verbose = false;
+    int num_threads = 16;   // 0
+    int num_rollout_per_thread = 100; // 1
+    bool verbose = false; 
     bool verbose_time = false;
 
-    string save_tree_filename;
+    string save_tree_filename;  // ""
 
     bool persistent_tree = false;
     // [TODO] Not a good design.
diff --git a/elf/utils_elf.py b/elf/utils_elf.py
old mode 100644
new mode 100755
index d3e30027..8ed9200c
--- a/elf/utils_elf.py
+++ b/elf/utils_elf.py
@@ -223,12 +223,14 @@ def __init__(self, GC, co, descriptions, use_numpy=False, gpu=None, params=dict(
             gpu(int): gpu to use.
             params(dict): additional parameters
         '''
-
+ 
+        #self.isPrint = False
         self._init_collectors(GC, co, descriptions, use_gpu=gpu is not None, use_numpy=use_numpy)
         self.gpu = gpu
         self.inputs_gpu = [ self.inputs[gids[0]].cpu2gpu(gpu=gpu) for gids in self.gpu2gid ] if gpu is not None else None
         self.params = params
         self._cb = { }
+        
 
     def _init_collectors(self, GC, co, descriptions, use_gpu=True, use_numpy=False):
         num_games = co.num_games
@@ -236,7 +238,7 @@ def _init_collectors(self, GC, co, descriptions, use_gpu=True, use_numpy=False):
         total_batchsize = 0
         for key, v in descriptions.items():
             total_batchsize += v["batchsize"]
-
+        
         if co.num_collectors > 0:
             num_recv_thread = co.num_collectors
         else:
@@ -269,11 +271,11 @@ def _init_collectors(self, GC, co, descriptions, use_gpu=True, use_numpy=False):
             for i in range(num_recv_thread):
                 group_id = GC.AddCollectors(batchsize, len(gpu2gid) - 1, timeout_usec, gstat)
 
-                input_batch = Batch.load(GC, "input", input, group_id, use_gpu=use_gpu, use_numpy=use_numpy)
+                input_batch = Batch.load(GC, "input", input, group_id, use_gpu=use_gpu, use_numpy=use_numpy) # 加载输入Batch
                 input_batch.batchsize = batchsize
                 inputs.append(input_batch)
                 if reply is not None:
-                    reply_batch = Batch.load(GC, "reply", reply, group_id, use_gpu=use_gpu, use_numpy=use_numpy)
+                    reply_batch = Batch.load(GC, "reply", reply, group_id, use_gpu=use_gpu, use_numpy=use_numpy) # 加载回复Batch
                     reply_batch.batchsize= batchsize
                     replies.append(reply_batch)
                 else:
@@ -298,6 +300,14 @@ def _init_collectors(self, GC, co, descriptions, use_gpu=True, use_numpy=False):
         self.name2idx = name2idx
         self.gid2gpu = gid2gpu
         self.gpu2gid = gpu2gid
+        # if not self.isPrint:
+            # print("idx2name",self.idx2name)
+            # print("name2idx",self.name2idx)
+            # print("gid2gpu",self.gid2gpu)
+            # print("gpu2gid",self.gpu2gid)
+            # print("num_collectors: ",co.num_collectors)
+            # self.isPrint = True
+
 
     def reg_has_callback(self, key):
         return key in self.name2idx
@@ -311,6 +321,7 @@ def reg_callback_if_exists(self, key, cb):
 
     def reg_callback(self, key, cb):
         '''Set callback function for key
+        注册回调函数，有符合要求和数量的Batch到来时，调用对应的函数
 
         Parameters:
             key(str): the key used to register the callback function.
@@ -332,7 +343,7 @@ def _call(self, infos):
             raise ValueError("info.gid[%d] is not in callback functions" % infos.gid)
 
         if self._cb[infos.gid] is None:
-            return;
+            return
 
         batchsize = len(infos.s)
 
diff --git a/rlpytorch/sampler/sample_methods.py b/rlpytorch/sampler/sample_methods.py
old mode 100644
new mode 100755
index ec657de7..b5b98f38
--- a/rlpytorch/sampler/sample_methods.py
+++ b/rlpytorch/sampler/sample_methods.py
@@ -33,10 +33,10 @@ def sample_with_check(probs, greedy=False):
     '''
     num_action = probs.size(1)
     if greedy:
-        _, actions = probs.max(1)
+        _, actions = probs.max(1)  # 贪婪算法，每次取概率最大的动作
         return actions
     while True:
-        actions = probs.multinomial(1)[:,0]
+        actions = probs.multinomial(1)[:,0]  # 按照概率选择一个动作
         cond1 = (actions < 0).sum()
         cond2 = (actions >= num_action).sum()
         if cond1 == 0 and cond2 == 0:
@@ -74,8 +74,9 @@ def sample_eps_with_check(probs, epsilon, greedy=False):
         rej_p = probs.new().resize_(2)
         rej_p[0] = 1 - epsilon
         rej_p[1] = epsilon
+        # rej 按照概率取 0 或 1（batchsize次），取到1时(epsilon)表示此次不选择该动作并随机取样
         rej = rej_p.multinomial(batchsize, replacement=True).byte()
-
+        # 随机取样
         uniform_p = probs.new().resize_(num_action).fill_(1.0 / num_action)
         uniform_sampling = uniform_p.multinomial(batchsize, replacement=True)
         actions[rej] = uniform_sampling[rej]
@@ -110,7 +111,7 @@ def sample_multinomial(state_curr, args, node="pi", greedy=False):
         return actions
     else:
         probs = state_curr[node].data
-        return sample_eps_with_check(probs, args.epsilon, greedy=greedy)
+        return sample_eps_with_check(probs, args.epsilon, greedy=greedy)  # probs 0 False
 
 def epsilon_greedy(state_curr, args, node="pi"):
     ''' epsilon greedy sampling
diff --git a/rlpytorch/sampler/sampler.py b/rlpytorch/sampler/sampler.py
old mode 100644
new mode 100755
diff --git a/rlpytorch/trainer/trainer.py b/rlpytorch/trainer/trainer.py
old mode 100644
new mode 100755
index 8f2b4241..66bbbe15
--- a/rlpytorch/trainer/trainer.py
+++ b/rlpytorch/trainer/trainer.py
@@ -40,6 +40,7 @@ def __init__(self, name="eval", stats=True, verbose=False, actor_name="actor"):
             on_get_args = self._on_get_args,
             child_providers = child_providers
         )
+        self.isPrint = False
 
     def _on_get_args(self, _):
         if self.stats is not None and not self.stats.is_valid():
@@ -75,6 +76,8 @@ def actor(self, batch):
 
         if self.sampler is not None:
             reply_msg = self.sampler.sample(state_curr)
+            # if not self.isPrint:
+            #     print("sampler reply: ",reply_msg)
         else:
             reply_msg = dict(pi=state_curr["pi"].data)
 
@@ -88,6 +91,11 @@ def actor(self, batch):
             reply_msg["V"] = state_curr["V"].data
 
         self.actor_count += 1
+        # if not self.isPrint:
+        #     print("batch: ",batch)
+        #     print("state_curr",state_curr)
+        #     print("reply_msg",reply_msg)
+        #     self.isPrint = True
         return reply_msg
 
     def episode_summary(self, i):
diff --git a/rts/backend/main_loop.cc b/rts/backend/main_loop.cc
old mode 100644
new mode 100755
index 139c8402..ef62ac15
--- a/rts/backend/main_loop.cc
+++ b/rts/backend/main_loop.cc
@@ -435,7 +435,9 @@ int main(int argc, char *argv[]) {
         p.stop(true);
         std::cout << gstats.PrintInfo() << std::endl;
     } else {
+        std::cout<<"======RTSStateExtend===="<<std::endl;
         RTSStateExtend state(options);
+        std::cout<<"=====RTSGame"<<std::endl;
         RTSGame game(&state);
         cout << "Players: " << players << endl;
         add_players(players, frame_skip, &game);
@@ -443,6 +445,7 @@ int main(int argc, char *argv[]) {
 
         chrono::duration<double> duration = chrono::system_clock::now() - time_start;
         cout << "Total time spent = " << duration.count() << "s" << endl;
+        std::cout<<"======MainLoop======="<<std::endl;
         game.MainLoop();
         /*
             // Load replay etc.
diff --git a/rts/engine/cmd_receiver.h b/rts/engine/cmd_receiver.h
old mode 100644
new mode 100755
index f1f07b53..20a28edd
--- a/rts/engine/cmd_receiver.h
+++ b/rts/engine/cmd_receiver.h
@@ -36,8 +36,8 @@ typedef int VerboseChoice;
 #define CR_NODERIVED 8
 #define CR_ALL ( CR_DURATIVE | CR_IMMEDIATE | CR_DERIVED | CR_NODERIVED )
 
-// receive command and record them in the history.
-// The cmds are ordered so that the lowest level is executed first.
+// receive command and record them in the history.  接收命令
+// The cmds are ordered so that the lowest level is executed first. 
 class CmdReceiver {
 private:
     Tick _tick;
diff --git a/rts/engine/common_loader.py b/rts/engine/common_loader.py
index 25cde403..09968060 100644
--- a/rts/engine/common_loader.py
+++ b/rts/engine/common_loader.py
@@ -17,17 +17,17 @@ def __init__(self, module):
         self.module = module
 
         basic_define_args = [
-            ("handicap_level", 0),
+            ("handicap_level", 0),  
             ("players", dict(type=str, help=";-separated player infos. For example: type=AI_NN,fs=50,args=backup/AI_SIMPLE|decay/0.99|start/1000,fow=True;type=AI_SIMPLE,fs=50")),
             ("max_tick", dict(type=int, default=30000, help="Maximal tick")),
-            ("shuffle_player", dict(action="store_true")),
-            ("num_frames_in_state", 1),
+            ("shuffle_player", dict(action="store_true")),    # false
+            ("num_frames_in_state", 1),    
             ("max_unit_cmd", 1),
             ("seed", 0),
-            ("actor_only", dict(action="store_true")),
-            ("model_no_spatial", dict(action="store_true")), # TODO, put it to model
-            ("save_replay_prefix", dict(type=str, default=None)),
-            ("output_file", dict(type=str, default=None)),
+            ("actor_only", dict(action="store_true")),   # false
+            ("model_no_spatial", dict(action="store_true")), # TODO, put it to model false
+            ("save_replay_prefix", dict(type=str, default=None)),  
+            ("output_file", dict(type=str, default=None)),         
             ("cmd_dumper_prefix", dict(type=str, default=None)),
             ("gpu", dict(type=int, help="gpu to use", default=None)),
         ]
@@ -85,7 +85,7 @@ def _parse_players(self, opt, player_names):
     def _init_gc(self, player_names=None):
         args = self.args
 
-        co = self.module.ContextOptions()
+        co = self.module.ContextOptions()   #/elf/python_options_utils_cpp.h
         self.context_args.initialize(co)
 
         opt = self.module.PythonOptions()
@@ -96,7 +96,7 @@ def _init_gc(self, player_names=None):
         # [TODO] Put it to TD.
         opt.handicap_level = args.handicap_level
 
-        player_infos = self._parse_players(opt, player_names)
+        player_infos = self._parse_players(opt, player_names) # 处理Players并设置为AI
 
         # opt.output_filename = b"simulators.txt"
         # opt.output_filename = b"cout"
@@ -153,7 +153,7 @@ def initialize(self):
             # For training, we want input, action (filled by actor models), value (filled by actor models) and reward.
             desc["train"] = self._get_train_spec()
 
-        self.more_labels.add_labels(desc)
+        self.more_labels.add_labels(desc)  # ?
 
         params.update(dict(
             num_group = 1 if args.actor_only else 2,
diff --git a/rts/engine/game.cc b/rts/engine/game.cc
old mode 100644
new mode 100755
index 09d0ad08..407c1062
--- a/rts/engine/game.cc
+++ b/rts/engine/game.cc
@@ -20,6 +20,7 @@ using namespace std::chrono;
 ////////////////////////// RTSStateExtend ////////////////////////////////////
 RTSStateExtend::RTSStateExtend(const RTSGameOptions &options)
     : _options(options), _snapshot_to_load(-1), _paused(false), _output_stream_owned(false), _output_stream(nullptr) {
+        //std::cout<<"----------RTSStateExtend Constructor------------"<<std::endl;
     if (_options.output_file.empty()) {
         _output_stream = _options.output_stream;
         _output_stream_owned = false;
@@ -69,8 +70,10 @@ CmdReturn RTSStateExtend::dispatch_cmds(const UICmd& cmd) {
     return CMD_FAILED;
 }
 
-bool RTSStateExtend::Init() {
-    if (! RTSState::Prepare(_options, _output_stream)) return false;
+bool RTSStateExtend::Init(bool isPrint = false) {
+    if(isPrint)
+      std::cout<<"------初始化游戏-------"<<std::endl;
+    if (! RTSState::Prepare(_options, _output_stream,isPrint)) return false;
 
     if (_output_stream) *_output_stream << "In the main loop " << endl << flush;
 
diff --git a/rts/engine/game.h b/rts/engine/game.h
old mode 100644
new mode 100755
index 83a2a0a8..16739111
--- a/rts/engine/game.h
+++ b/rts/engine/game.h
@@ -28,7 +28,7 @@ class RTSStateExtend : public RTSState {
     ~RTSStateExtend();
 
     // Function used in GameLoop
-    bool Init() override;
+    bool Init(bool isPrint) override;
     void PreAct() override;
     void IncTick() override;
 
diff --git a/rts/engine/game_env.cc b/rts/engine/game_env.cc
old mode 100644
new mode 100755
index 3387ecc5..b2e5d21b
--- a/rts/engine/game_env.cc
+++ b/rts/engine/game_env.cc
@@ -11,7 +11,9 @@
 
 GameEnv::GameEnv() {
     // Load the map.
+   // std::cout<<"-------GameEnv-----------"<<std::endl;
     _map = unique_ptr<RTSMap>(new RTSMap());
+    // std::cout<<"-------GameEnv-----------"<<std::endl;
     _game_counter = -1;
     Reset();
 }
@@ -35,6 +37,7 @@ void GameEnv::ClearAllPlayers() {
 }
 
 void GameEnv::Reset() {
+    //std::cout<<"-------Reset-----------"<<std::endl;
     _map->ClearMap();
     _next_unit_id = 0;
     _winner_id = INVALID;
@@ -121,6 +124,7 @@ bool GameEnv::RemoveUnit(const UnitId &id) {
     return true;
 }
 
+// 找到最近的基地
 UnitId GameEnv::FindClosestBase(PlayerId player_id) const {
     // Find closest base. [TODO]: Not efficient here.
     for (auto it = _units.begin(); it != _units.end(); ++it) {
@@ -324,3 +328,15 @@ string GameEnv::PrintDebugInfo() const {
     ss << _map->PrintDebugInfo() << endl;
     return ss.str();
 }
+
+string GameEnv::PrintPlayerInfo() const {
+    stringstream ss;
+    for (const auto& player : _players) {
+        ss << "Player " << player.GetId() << endl;
+        ss << player.PrintInfo()<< endl;
+    }
+
+    ss << _map->Draw() << endl;
+    ss << _map->PrintDebugInfo() << endl;
+    return ss.str();
+}
diff --git a/rts/engine/game_env.h b/rts/engine/game_env.h
old mode 100644
new mode 100755
index cfdb9584..60e8bc5a
--- a/rts/engine/game_env.h
+++ b/rts/engine/game_env.h
@@ -89,6 +89,7 @@ class GameEnv {
 
     // Initialize different units for this game.
     void InitGameDef() {
+        //std::cout<<"----------InitGameDef-------"<<std::endl;
         _gamedef.Init();
     }
     const GameDef &GetGameDef() const { return _gamedef; }
@@ -174,6 +175,8 @@ class GameEnv {
     uint64_t CurrentHashCode() const;
 
     string PrintDebugInfo() const;
+
+    string PrintPlayerInfo() const;
     ~GameEnv() { }
 
     // Some utility function to pick first from units in a group.
diff --git a/rts/engine/game_options.h b/rts/engine/game_options.h
old mode 100644
new mode 100755
index 1f6c6b3c..832fd870
--- a/rts/engine/game_options.h
+++ b/rts/engine/game_options.h
@@ -26,7 +26,7 @@ struct RTSGameOptions {
     // Snapshots to load from.
     vector<Tick> snapshots;
 
-    // Initial seed. If seed = 0, then we use time(NULL)
+    // Initial seed. If seed = 0, then we use time(NULL)  初始化种子
     // When seed != 0, the game should be deterministic (If not then there is a bug somewhere).
     int seed = 0;
 
diff --git a/rts/engine/game_state.cc b/rts/engine/game_state.cc
old mode 100644
new mode 100755
index 52c7c45c..a680448e
--- a/rts/engine/game_state.cc
+++ b/rts/engine/game_state.cc
@@ -12,6 +12,7 @@ using namespace std;
 using namespace std::chrono;
 
 RTSState::RTSState() {
+   //std::cout<<"-----------RTSState Constructor --------"<<std::endl;
     _env.InitGameDef();
     // TODO Need to add players accordingly.
     _env.ClearAllPlayers();
@@ -19,13 +20,16 @@ RTSState::RTSState() {
 
 void RTSState::AppendPlayer(const string &name) {
     _env.AddPlayer(name, PV_KNOW_ALL);
+    //_env.AddPlayer(name, PV_NORMAL);
 }
 
 void RTSState::RemoveLastPlayer() {
     _env.RemovePlayer();
 }
 
-bool RTSState::Prepare(const RTSGameOptions &options, ostream *output) {
+bool RTSState::Prepare(const RTSGameOptions &options, ostream *output,bool isPrint) {
+    if(isPrint)
+       std::cout<<"------------Prepare-------------"<<std::endl;
     _cmd_receiver.SetVerbose(options.cmd_verbose, 0);
 
     const unsigned int game_counter = _env.GetGameCounter();
@@ -74,13 +78,26 @@ bool RTSState::Prepare(const RTSGameOptions &options, ostream *output) {
         }
 
         if (options.map_filename.empty()) {
+            //CmdRandomSeed *cmd_Random = new CmdRandomSeed(INVALID, seed);
+            // if(isPrint){
+            //     std::cout<<"SendCmdWithTick(CmdBPtr(new CmdRandomSeed(INVALID, seed))  --seed "<<seed<<std::endl;
+            //     std::cout<<cmd_Random->PrintInfo()<<std::endl;
+            // }
+            
             if (output) *output << "Generate from scratch, seed = " << seed << endl << flush;
-            _cmd_receiver.SendCmdWithTick(CmdBPtr(new CmdRandomSeed(INVALID, seed)), 0);
+             _cmd_receiver.SendCmdWithTick(CmdBPtr(new CmdRandomSeed(INVALID, seed)), 0);
+           
         } else {
             // _cmd_receiver.SendCmdWithTick(CmdBPtr(new CmdLoadMap(INVALID, options.map_filename)));
         }
-
+         if(isPrint){
+                std::cout<<"-----cmd Info------"<<std::endl;
+            }
         for (auto&& cmd_pair : _env.GetGameDef().GetInitCmds(options)) {
+            if(isPrint){
+                std::cout<<"---send InitCmd---"<<std::endl;
+                std::cout<<cmd_pair.first->PrintInfo()<<"  "<<cmd_pair.second<<std::endl;
+            }
             _cmd_receiver.SendCmdWithTick(std::move(cmd_pair.first), cmd_pair.second);
         }
     }
diff --git a/rts/engine/game_state.h b/rts/engine/game_state.h
old mode 100644
new mode 100755
index 48e548f0..17eb8bee
--- a/rts/engine/game_state.h
+++ b/rts/engine/game_state.h
@@ -19,7 +19,7 @@ class RTSState {
 public:
     RTSState();
 
-    bool Prepare(const RTSGameOptions &options, ostream *output = nullptr);
+    bool Prepare(const RTSGameOptions &options, ostream *output = nullptr, bool isPrint = false);
 
     void Save(string *s) const {
         serializer::saver saver(true);
@@ -79,7 +79,7 @@ class RTSState {
     void SetReplayPrefix(const std::string &prefix) { _save_replay_prefix = prefix; }
 
     // Function used in GameLoop
-    virtual bool Init() { return true; }
+    virtual bool Init(bool isPrint = false) { return true; }
     virtual void PreAct() { }
     virtual void IncTick() { _cmd_receiver.IncTick(); }
 
diff --git a/rts/engine/gamedef.cc b/rts/engine/gamedef.cc
old mode 100644
new mode 100755
index c219c3bc..f5933839
--- a/rts/engine/gamedef.cc
+++ b/rts/engine/gamedef.cc
@@ -7,7 +7,16 @@
 */
 
 #include "gamedef.h"
-
+/**
+ *  定义单位属性
+ *  cost   造价(不需要)
+ *  hp     血量
+ *  defence防御力 
+ *  speed  速度
+ *  att    攻击力
+ *  att_r  攻击距离
+ *  vis_r  可视距离
+ * */
 UnitTemplate _C(int cost, int hp, int defense, float speed, int att, int att_r, int vis_r,
         const vector<int> &cds, const vector<CmdType> &l, UnitAttr attr) {
 
@@ -21,7 +30,7 @@ UnitTemplate _C(int cost, int hp, int defense, float speed, int att, int att_r,
     p._attr = attr;
     p._att_r = att_r;
     p._vis_r = vis_r;
-    for (int i = 0; i < NUM_COOLDOWN; ++i) {
+    for (int i = 0; i < NUM_COOLDOWN; ++i) {   //设置CD
         p._cds[i].Set(cds[i]);
     }
 
diff --git a/rts/engine/gamedef.h b/rts/engine/gamedef.h
old mode 100644
new mode 100755
index a2140cd7..6cbc88e7
--- a/rts/engine/gamedef.h
+++ b/rts/engine/gamedef.h
@@ -21,8 +21,8 @@ struct UnitProperty {
     int _att_r;
     float _speed;
 
-    // Visualization range.
-    int _vis_r;
+    // Visualization range. 可视距离
+    int _vis_r;  
 
     int _changed_hp;
     UnitId _damage_from;
diff --git a/rts/engine/map.cc b/rts/engine/map.cc
old mode 100644
new mode 100755
index e11659f6..4826e217
--- a/rts/engine/map.cc
+++ b/rts/engine/map.cc
@@ -11,11 +11,14 @@
 
 // Constructor
 RTSMap::RTSMap() {
+    //std::cout<<"RTSMAP"<<std::endl;
     load_default_map();
     reset_intermediates();
 }
 
 bool RTSMap::find_two_nearby_empty_slots(const std::function<uint16_t(int)>& f, int *x1, int *y1, int *x2, int *y2, int i) const {
+   // std::cout<<"find_two_nearby_empty_slots"<<std::endl;
+    
     const int kDist = 4;
     int kMaxTrial = 100;
     for (int j = 0; j < kMaxTrial; ++j) {
@@ -31,8 +34,10 @@ bool RTSMap::find_two_nearby_empty_slots(const std::function<uint16_t(int)>& f,
 }
 
 bool RTSMap::GenerateImpassable(const std::function<uint16_t(int)>& f, int nImpassable) {
-    _map.assign(_m * _n * _level, MapSlot());
-    for (int i = 0; i < nImpassable; ++i) {
+    _map.assign(_m * _n * _level, MapSlot());   //初始化地图格子 类型为NORMAL
+
+    
+    for (int i = 0; i < nImpassable; ++i) {  //随机选一些格子，设为IMPOSSIBLE
         const int x = f(_m);
         const int y = f(_n);
         _map[GetLoc(Coord(x, y))].type = IMPASSABLE;
@@ -115,9 +120,10 @@ bool RTSMap::GenerateTDMaze(const std::function<uint16_t(int)>& f) {
 bool RTSMap::GenerateMap(const std::function<uint16_t(int)>& f, int nImpassable, int num_player, int init_resource) {
     // load a map for now simple format.
     bool success;
+  // std::cout<<"-------GenerateMap nImpassable = "<<nImpassable<<std::endl; 
     do {
         success = true;
-        GenerateImpassable(f, nImpassable);
+        GenerateImpassable(f, nImpassable);  //初始化地图格子，并随机设置一些点为IMPOSSIBLE
         int x1 = -1, y1 = -1, x2 = -1, y2 = -1;
         _infos.clear();
         for (PlayerId i = 0; i < num_player; ++i) {
@@ -149,10 +155,17 @@ void RTSMap::reset_intermediates() {
 }
 
 void RTSMap::load_default_map() {
-    _m = 20;
-    _n = 20;
+    // _m = 20;
+    // _n = 20;
+    // _level = 1;
+    // _map.assign(_m * _n * _level, MapSlot());
+
+    _m = 35;
+    _n = 35;
     _level = 1;
     _map.assign(_m * _n * _level, MapSlot());
+
+
 }
 
 void RTSMap::precompute_all_pair_distances() {
@@ -162,8 +175,9 @@ void RTSMap::precompute_all_pair_distances() {
     // once for each map, we could just use it.
 }
 
+// 增加单位 
 bool RTSMap::AddUnit(const UnitId &id, const PointF& new_p) {
-    if (_locality.Exists(id)) return false;
+    if (_locality.Exists(id)) return false;  // 位置已有单位，不可添加
     if (! _locality.IsEmpty(new_p, kUnitRadius, INVALID)) return false;
 
     _locality.Add(id, new_p, kUnitRadius);
@@ -204,7 +218,9 @@ vector<Loc> RTSMap::GetSight(const Loc& loc, int range) const {
     const int xmax = std::min(_m - 1, c.x + range);
 
     for (int x = xmin; x <= xmax; ++x) {
-        const int yrange = range - std::abs(c.x  - x);
+        //const int yrange = range - std::abs(c.x  - x);
+        const int x_1 = std::abs(c.x - x);
+        const int yrange = std::sqrt(range*range - x_1*x_1);
         const int ymin = std::max(0, c.y - yrange);
         const int ymax = std::min(_n - 1, c.y + yrange);
         for (int y = ymin; y <= ymax; ++y) {
diff --git a/rts/engine/map.h b/rts/engine/map.h
old mode 100644
new mode 100755
index 9ea78932..5d2d4f2a
--- a/rts/engine/map.h
+++ b/rts/engine/map.h
@@ -16,6 +16,7 @@
 
 struct MapSlot {
   // three layers, terrian, ground and air.
+  // 只有两层 ground air
   Terrain type;
   int height;
 
diff --git a/rts/engine/player.cc b/rts/engine/player.cc
old mode 100644
new mode 100755
index 06316b30..8cac3772
--- a/rts/engine/player.cc
+++ b/rts/engine/player.cc
@@ -58,6 +58,7 @@ void Player::ComputeFOW(const Units &units) {
     // or loop on the units.
     // [TODO]: We could do better with LocalitySearch.
     // Clear fogs.
+    //std::cout<<"------ComputFOW--------"<<std::endl;
     for (Fog &f : _fogs) {
         f.MakeInvisible();
     }
@@ -84,9 +85,9 @@ void Player::ComputeFOW(const Units &units) {
     for (auto it = units.begin(); it != units.end(); ++it) {
         const Unit *u = it->second.get();
         if (ExtractPlayerId(u->GetId()) != _player_id) {
-            Loc l = _filter_with_fow(*u);
+            Loc l = _filter_with_fow(*u);  //判断该点的位置是否在视野范围内
             // Add the unit info to the loc.
-            if (l != -1) _fogs[l].SaveUnit(*u);
+            if (l != -1) _fogs[l].SaveUnit(*u);  //在迷雾格中存储该单位
         }
     }
 }
@@ -106,7 +107,9 @@ string Player::PrintInfo() const {
     stringstream ss;
     ss << "Map ptr = " << _map << endl;
     ss << "Player id = " << _player_id << endl;
+    ss << "Player name = "<< _name << endl;
     ss << "Resource = " << _resource << endl;
+    ss << "PlayerPrivilege = "<<_privilege<<endl;
     ss << "Fog[" << _fogs.size() << "] = ";
     for (size_t i = 0; i < _fogs.size(); ++i) ss << _fogs[i]._fog << " ";
     ss << endl;
diff --git a/rts/engine/player.h b/rts/engine/player.h
old mode 100644
new mode 100755
index 3aad727e..e2fb737b
--- a/rts/engine/player.h
+++ b/rts/engine/player.h
@@ -19,9 +19,9 @@ class Unit;
 struct Fog {
     // Fog level: 0 no fog, 100 completely invisible.
     int _fog = 100;
-    vector<Unit> _prev_seen_units;
+    vector<Unit> _prev_seen_units;  // 该点可见单位
 
-    void MakeInvisible() {  _fog = 100; }
+    void MakeInvisible() {  _fog = 100; }  // 让该点不可见
     void SetClear() { _fog = 0; _prev_seen_units.clear(); }
     bool CanSeeTerrain() const { return _fog < 50; }
     bool CanSeeUnit() const { return _fog < 30; }
diff --git a/rts/engine/python_common_options.h b/rts/engine/python_common_options.h
old mode 100644
new mode 100755
index e548c7e0..0f0da064
--- a/rts/engine/python_common_options.h
+++ b/rts/engine/python_common_options.h
@@ -71,21 +71,21 @@ struct PythonOptions {
     int map_size_x, map_size_y;
 
     // Maximum unit command you could send per action.
-    int max_unit_cmd;
+    int max_unit_cmd;  // 1
 
     // Max tick.
-    int max_tick;
+    int max_tick;  // 30000
 
     // Random seed to use. seed = 0 mean uses time(NULL).
     // If seed != 0, then each simulation thread will use a seed which is a deterministic function
     // of PythonOption.seed and the thread id.
-    int seed;
+    int seed;  // 0
 
-    bool shuffle_player;
+    bool shuffle_player;  // false 
 
     int game_name;
     // [TODO] put handicap to TD.
-    int handicap_level;
+    int handicap_level;  // 0
 
     PythonOptions()
       : simulation_type(ST_NORMAL), map_size_x(20), map_size_y(20), max_unit_cmd(10), max_tick(30000), seed(0), shuffle_player(false), game_name(0), handicap_level(0) {
diff --git a/rts/engine/rule_actor.h b/rts/engine/rule_actor.h
index 474563cc..eed23247 100644
--- a/rts/engine/rule_actor.h
+++ b/rts/engine/rule_actor.h
@@ -16,7 +16,7 @@
 
 custom_enum(AIState, STATE_START = 0, STATE_BUILD_WORKER, STATE_BUILD_BARRACK,
   STATE_BUILD_MELEE_TROOP, STATE_BUILD_RANGE_TROOP, STATE_ATTACK,
-  STATE_ATTACK_IN_RANGE, STATE_HIT_AND_RUN, STATE_DEFEND, NUM_AISTATE);
+  STATE_ATTACK_IN_RANGE, STATE_HIT_AND_RUN, STATE_DEFEND, NUM_AISTATE); // Action Space 
 
 custom_enum(FlagState, FLAGSTATE_START = 0, FLAGSTATE_GET_FLAG, FLAGSTATE_ATTACK_FLAG,
     FLAGSTATE_ESCORT_FLAG, FLAGSTATE_PROTECT_FLAG, //FLAGSTATE_ATTACK, FLAGSTATE_MOVE,
diff --git a/rts/engine/wrapper_template.h b/rts/engine/wrapper_template.h
old mode 100644
new mode 100755
index eca6d382..9a5726ee
--- a/rts/engine/wrapper_template.h
+++ b/rts/engine/wrapper_template.h
@@ -30,7 +30,11 @@ class WrapperT {
             const PythonOptions &options, const elf::Signal &signal,
             const std::map<std::string, int> *more_params, Comm *comm) {
         const string& replay_prefix = options.save_replay_prefix;
-
+        
+        bool isPrint = false; 
+        if (game_idx==1)
+            isPrint = true;
+        
         // Create a game.
         RTSGameOptions op;
         op.seed = (options.seed == 0 ? 0 : options.seed + game_idx);
@@ -41,30 +45,47 @@ class WrapperT {
         op.output_file = options.output_filename;
         op.cmd_dumper_prefix = options.cmd_dumper_prefix;
 
-        // std::cout << "before running wrapper" << std::endl;
-
-        WrapperCB wrapper(game_idx, context_options, options, comm);
-        wrapper.OnGameOptions(&op);
-
-        // std::cout << "before initializing the game" << std::endl;
+        if(isPrint){
+            string ss = op.PrintInfo();
+            std::cout<<"----------RTSGameOptions----"<<std::endl;
+            std::cout<<ss<<std::endl;
+        }
 
-        // Note that all the bots created here will be owned by game.
-        // Note that AddBot() will set its receiver. So there is no need to specify it here.
-        RTSStateExtend s(op);
-        RTSGame game(&s);
-        wrapper.OnGameInit(&game, more_params);
+        //std::cout << "before running wrapper" << std::endl;
 
-        s.SetGlobalStats(&_gstats);
+        WrapperCB wrapper(game_idx, context_options, options, comm);
+        wrapper.OnGameOptions(&op);  //设置了handicap_level用于gameTD?
+  
+        //std::cout << "before initializing the game" << std::endl;
+
+        // Note that all the bots created here will be owned by game. 
+        // Note that AddBot() will set its receiver. So there is no need to specify it here. 
+        RTSStateExtend s(op);    // 初始化 RTSStateExtend 
+        RTSGame game(&s);   // 初始化 RTSGame, s = RTSStateExtend
+
+        if(isPrint){
+            std::cout<<"----------python options-------"<<endl;
+            options.Print();
+            
+        }        
+        //wrapper.OnGameInit(&game, more_params);
+        wrapper.OnGameInit(&game, more_params,isPrint); //根据PythonOption初始化AI和玩家并添加到游戏中
+
+        s.SetGlobalStats(&_gstats);  // 设置游戏状态
+        // if(isPrint){
+        //     std::cout<<"GlobalStats: "<<_gstats.PrintInfo()<<endl;
+        // }
 
         unsigned long int seed = (op.seed == 0 ? time(NULL) : op.seed);
         std::mt19937 rng;
         rng.seed(seed);
 
         int iter = 0;
-        // std::cout << "Start the main loop" << std::endl;
+        if(isPrint)
+          std::cout << "Start the main loop" << std::endl;
         while (! signal.IsDone()) {
-            wrapper.OnEpisodeStart(iter, &rng, &game);
-            game.MainLoop(&signal.done());
+            wrapper.OnEpisodeStart(iter, &rng, &game);   //将这三个参数转为void类型
+            game.MainLoop(&signal.done(),isPrint);
             game.Reset();
             ++ iter;
         }
diff --git a/rts/frontend/._minirts.html b/rts/frontend/._minirts.html
new file mode 100755
index 00000000..727a2ed8
Binary files /dev/null and b/rts/frontend/._minirts.html differ
diff --git a/rts/frontend/game.js b/rts/frontend/game.js
old mode 100644
new mode 100755
index 1e6b5604..d3e81c8e
--- a/rts/frontend/game.js
+++ b/rts/frontend/game.js
@@ -1,24 +1,26 @@
 /**
 * Copyright (c) 2017-present, Facebook, Inc.
 * All rights reserved.
-
 * This source code is licensed under the BSD-style license found in the
 * LICENSE file in the root directory of this source tree.
 */
 
-// Create the canvas
+// 创建画布，仅仅是容器，需要调用脚本来执行绘图
 var canvas = document.createElement("canvas");
+// 获得绘图环境，在该函数下绘图
 var ctx = canvas.getContext("2d");
-canvas.width = 1400;
-canvas.height = 1000;
+canvas.width = 30000;           
+canvas.height = 28000;       
 var left_frame_width = 1000;
-var cell_size = 50;
-var rect_size = 50;
-var unit_size = 32;
+var cell_size = 40;         // 50
+var rect_size = 40;       // 50
+var unit_size = 32;     // 32
 var cell_colors = ['#404040', 'blue', 'black'];
-var player_colors = ['blue', 'red', 'yellow']
+var player_colors = ['blue', 'red', 'yellow'];
 
-var unit_names_minirts = ["RESOURCE", "WORKER", "MELEE_ATTACKER", "RANGE_ATTACKER", "BARRACKS", "BASE"];
+// 不同游戏的单位名称
+// var unit_names_minirts = ["RESOURCE", "WORKER", "MELEE_ATTACKER", "RANGE_ATTACKER", "BARRACKS", "BASE"];     RANGE  雷达 范围大
+var unit_names_minirts = [["RESOURCE", "WORKER", "MELEE_ATTACKER2", "RANGE_ATTACKER2", "BARRACKS", "BASE"],["RESOURCE", "WORKER", "MELEE_ATTACKER1", "RANGE_ATTACKER1", "BARRACKS", "BASE"]];
 var unit_names_flag = ["FLAG_BASE", "FLAG_ATHLETE", "FLAG"];
 var unit_names_td = ["BASE", "WORKER", "RANGE_ATTACKER"];
 var x_down = null;
@@ -33,6 +35,7 @@ var speed = 0;
 var min_speed = -10;
 var max_speed = 5;
 
+// FPS 进度条
 var range2 = document.createElement("INPUT");
 range2.type = "range";
 range2.min = min_speed;
@@ -60,8 +63,9 @@ range2.oninput = function(){
     }
 }
 
-document.body.appendChild(range2);
+// document.body.appendChild(range2);
 
+// 按钮
 var addButton = function(text, cmd) {
     var button = document.createElement("button");
     button.innerHTML = text;
@@ -94,10 +98,12 @@ var addButton = function(text, cmd) {
     });
 };
 
-addButton("Faster", "F");
-addButton("Slower", "W");
-addButton("Cycle", "C");
-addButton("Pause", "P");
+// 添加几个按键
+// addButton("Faster", "F");
+// addButton("Slower", "W");
+// addButton("Cycle", "C");
+// addButton("Pause", "P");
+
 
 var range1 = document.createElement("INPUT");
 range1.type = "range";
@@ -114,34 +120,40 @@ range1.style.height = "30px";
 range1.oninput = function(){
     send_cmd(tick + ' S ' + this.value);
 }
-document.body.appendChild(range1);
+// document.body.appendChild(range1);
 
 document.body.appendChild(canvas);
 
+
 var send_cmd = function(s) {
   dealer.send(s);
 };
 
+
 canvas.oncontextmenu = function (e) {
     e.preventDefault();
 };
 
+
 document.addEventListener("keydown", function (e) {
     send_cmd(tick + ' ' + e.key);
 }, false);
 
+
 canvas.addEventListener("mousedown", function (e) {
     if (e.button === 0) {
         var xy0 = convert_xy_back(e.pageX, e.pageY);
-        if (xy0[0] > 20 || xy0[1] > 20) return;
+        // console.log(xy0);
+        if (xy0[0] >700 || xy0[1] > 700) return;
         x_down = e.pageX;
         y_down = e.pageY;
     }
 }, false);
 
+
 canvas.addEventListener("mouseup", function (e) {
     var xy0 = convert_xy_back(e.pageX, e.pageY);
-    if (xy0[0] > 20 || xy0[1] > 20) return;
+    if (xy0[0] > 700 || xy0[1] > 700) return;     
     if (e.button === 0) {
         var xy = convert_xy_back(x_down, y_down);
         if (dragging && x_down && y_down) {
@@ -158,6 +170,7 @@ canvas.addEventListener("mouseup", function (e) {
     }
 }, false);
 
+
 canvas.addEventListener("mousemove", function (e) {
     if (x_down && y_down) {
         x_curr = e.pageX;
@@ -168,21 +181,21 @@ canvas.addEventListener("mousemove", function (e) {
     }
 }, false);
 
-
+// game.rts_map
 var onMap = function(m) {
     var counter = 0;
     for (y = 0; y < m.height; y++) {
     	for (x = 0; x < m.width; x++){
-    		var color = cell_colors[m.slots[counter]];
-            var x1 = x * cell_size;
-            var y1 = y * cell_size;
+            var color = cell_colors[m.slots[counter]];                    // counter = 0   m.slots = #404040
+            var x1 = x * cell_size;                                                           // 一个格子的长是 50
+            var y1 = y * cell_size;                                                             // 宽  50
             ctx.beginPath();
             ctx.fillStyle = color;
-            ctx.lineWidth = 1;
-		    ctx.rect(x1, y1, rect_size, rect_size);
-		    ctx.strokeStyle = 'black';
-		    ctx.stroke();
-		    ctx.fillRect(x1, y1, rect_size, rect_size);
+            ctx.lineWidth = 1;                                              // 格子之间的线宽  1
+		    ctx.rect(x1, y1, rect_size, rect_size);     
+		    ctx.strokeStyle = 'black';                                                // 矩形 用 black 填充 
+		    ctx.stroke();                                                           
+		    ctx.fillRect(x1, y1, rect_size, rect_size);             
 		    ctx.closePath();
             counter += 1;
     	}
@@ -200,7 +213,9 @@ var draw_hp = function(bbox, states, font_color, player_color){
     var margin = 2;
     ctx.fillStyle = 'black';
     ctx.lineWidth = margin;
+     // 开始一条路径
     ctx.beginPath();
+        // 绘制一个矩阵
     ctx.rect(x1, y1, x2 - x1, y2 - y1);
     ctx.fillRect(x1, y1, x2 - x1, y2 - y1);
     ctx.strokeStyle = player_color;
@@ -211,6 +226,8 @@ var draw_hp = function(bbox, states, font_color, player_color){
     if (hp_ratio <= 0.2) color = 'red';
     ctx.fillStyle = color;
     ctx.fillRect(x1, y1, Math.floor((x2 - x1) * hp_ratio + 0.5), y2 - y1);
+
+    // 如果单位有名字，则显示名字
     if (state_str){
     	ctx.beginPath();
     	ctx.fillStyle = font_color;
@@ -220,14 +237,15 @@ var draw_hp = function(bbox, states, font_color, player_color){
     }
 }
 
+// 单位的绘制实现      
 var onUnit = function(u, isSelected) {
-    var player_color = player_colors[u.player_id];
-    var p =  u.p;
-    var last_p = u.last_p;
-    var diffx = p.x - last_p.x;
+    var player_color = player_colors[u.player_id];          // unit 内部的 player_id，用于区分单位是哪一方的
+    var p =  u.p;                                                                            //  单位此刻的位置                                  
+    var last_p = u.last_p;                                                              // 上一刻的位置
+    var diffx = p.x - last_p.x;                                                        
     var diffy = p.y - last_p.y;
-    var ori = "down";
-    if (Math.abs(diffx) > Math.abs(diffy)) {
+    var ori = "down";                                                                       
+    if (Math.abs(diffx) > Math.abs(diffy)) {                                
         if (diffx >= 0) {
             ori = "right";
         } else {
@@ -242,12 +260,13 @@ var onUnit = function(u, isSelected) {
     }
     var xy = convert_xy(p.x, p.y);
 
-    draw_sprites(sprites[unit_names_minirts[u.unit_type]], xy[0], xy[1], ori);
+    draw_sprites(sprites[unit_names_minirts[u.player_id][u.unit_type]], xy[0], xy[1], ori);     
 
-    var hp_ratio = u.hp / u.max_hp;
+    var hp_ratio = u.hp / u.max_hp;         // 掉血的比例
     var state_str;
     if ("cmd" in u) {
         if (u.cmd.cmd[0] != 'I') {
+            // 农民的名字   G1  G0 
             state_str = u.cmd.cmd[0] + u.cmd.state;
         }
     }
@@ -255,6 +274,7 @@ var onUnit = function(u, isSelected) {
     var y1 = xy[1] - 27;
     var x2 = x1 + unit_size;
     var y2 = y1 + 5;
+        // 绘制血条。后两个参数：字体颜色、血条边框颜色
     draw_hp([x1, y1, x2, y2], [hp_ratio, state_str], 'white', player_color);
     if (isSelected) {
         ctx.lineWidth = 2;
@@ -266,40 +286,44 @@ var onUnit = function(u, isSelected) {
     }
 };
 
+// 绘制子弹
 var onBullet = function(bullet) {
     var xy = convert_xy(bullet.p.x, bullet.p.y);
+        // 加载子弹图片
     draw_sprites(bullets, xy[0], xy[1], bullet.state);
 }
 
-var onPlayerStats = function(player) {
-    if (player.player_id == 2) {
-        unit_names_minirts = unit_names_flag;
-    }
-    var x1 = left_frame_width + 10;
-    var y1 = (player.player_id + 1) * 50;
-    var label = ["PlayerId", player.player_id, "Resource", player.resource].join(" ");
-    ctx.beginPath()
-	ctx.fillStyle = "Black";
-	ctx.font = "15px Arial";
-	ctx.fillText(label, x1, y1);
-    ctx.closePath();
-}
+// 显示 玩家资源的 文本信息
+// var onPlayerStats = function(player) {
+//     if (player.player_id == 2) {
+//         unit_names_minirts = unit_names_flag;
+//     }
+//     var x1 = left_frame_width + 10;
+//     var y1 = (player.player_id + 1) * 50;
+//     var label = ["PlayerId", player.player_id, "Resource", player.resource].join(" ");
+//     ctx.beginPath()
+// 	ctx.fillStyle = "Black";
+// 	ctx.font = "15px Arial";
+// 	ctx.fillText(label, x1, y1);
+//     ctx.closePath();
+// }
+
 
-// Draw units that have been seen.
 var onPlayerSeenUnits = function(m) {
     if ("units" in m) {
+        // 未执行
         var oldAlpha = ctx.globalAlpha;
         ctx.globalAlpha = 0.3;
 
         for (var i in m.units) {
             onUnit(m.units[i], false);
         }
-        // console.log(m.units.length)
 
         ctx.globalAlpha = oldAlpha;
     }
 }
 
+// 空白出显示游戏单位状态
 var draw_state = function(u) {
     var x1 = left_frame_width + 10;
     var y1 = 150;
@@ -339,7 +363,7 @@ var convert_xy_back = function(x, y){
 };
 
 var load_sprites = function(spec) {
-    // Default behavior.
+    // 默认行为
     var specReady = false;
     var specImage = new Image();
     specImage.onload = function () {
@@ -350,22 +374,25 @@ var load_sprites = function(spec) {
     return spec;
 };
 
-var draw_sprites = function(spec, px, py, ori) {
+// 绘制图片函数
+var draw_sprites = function(spec, px, py, ori) {    // 图片 、
     var image = spec["image"]
     var width = image.width;
     var height = image.height;
-    if (!("_sizes" in spec)) {
-        ctx.drawImage(image, px - width / 2, py - height / 2);
+    if (!("_sizes" in spec)) {          
+        ctx.drawImage(image, px - width / 2, py - height / 2);      // 图片，在画布上放的位置的x，y坐标
     } else {
-        var sw = spec["_sizes"][0];
-        var sh = spec["_sizes"][1];
+
+        var sw = spec["_sizes"][0];                                                     
+        var sh = spec["_sizes"][1];                                                     
         var nw = Math.floor(width / sw);
         var nh = Math.floor(height/ sh);
         var xidx = spec[ori][0];
-        var yidx = spec[ori][1];
-        var cx = xidx[Math.floor(tick / 3) % xidx.length] * sw;
-        var cy = yidx[Math.floor(tick / 3) % yidx.length] * sh;
-        ctx.drawImage(image, cx, cy, sw, sh, px - sw / 2, py - sh / 2, sw, sh);
+        var yidx = spec[ori][1];                                                                // 
+        var cx = xidx[Math.floor(tick / 3) % xidx.length] * sw;     
+        var cy = yidx[Math.floor(tick / 3) % yidx.length] * sh;      
+        //  px - sw / 2, py - sh / 2        代表图像在画布上的 x   y 坐标
+        ctx.drawImage(image, cx, cy, sw, sh, px - sw / 2, py - sh / 2, sw, sh);     // 剪切图像，并在画布上定位被剪切的部分
     }
 };
 
@@ -374,10 +401,44 @@ var myrange = function (j, k){
 	return Array.from(new Array(n), (x,i) => i + j);
 };
 
-// load pics
+
 var sprites = {};
 
-sprites["RANGE_ATTACKER"] = load_sprites({
+// sprites["RANGE_ATTACKER"] = load_sprites({
+//     "up" : [myrange(15, 22), [0]],
+//     "down": [myrange(15, 22), [1]],
+//     "left": [[16], myrange(2, 9)],
+//     "right": [[15], myrange(2, 9)],
+//     "_file" : "imgs/tiles.png",
+//     "_sizes" : [32, 32]
+// });
+
+// sprites["MELEE_ATTACKER"] = load_sprites({
+//     "up" : [myrange(15, 22), [9]],
+//     "down": [myrange(15, 22), [10]],
+//     "left": [[20], myrange(2, 9)],
+//     "right": [[21], myrange(2, 9)],
+//     "_file" : "imgs/tiles.png",
+//     "_sizes" : [32, 32]
+// });
+sprites["RANGE_ATTACKER1"] = load_sprites({
+    "up" : [myrange(15, 22), [0]],
+    "down": [myrange(15, 22), [1]],
+    "left": [[16], myrange(2, 9)],
+    "right": [[15], myrange(2, 9)],
+    "_file" : "imgs/tile11.png",
+    "_sizes" : [32, 32]
+});
+
+sprites["MELEE_ATTACKER1"] = load_sprites({
+    "up" : [myrange(15, 22), [9]],
+    "down": [myrange(15, 22), [10]],
+    "left": [[20], myrange(2, 9)],
+    "right": [[21], myrange(2, 9)],
+    "_file" : "imgs/tile11.png",
+    "_sizes" : [32, 32]
+});
+sprites["RANGE_ATTACKER2"] = load_sprites({
     "up" : [myrange(15, 22), [0]],
     "down": [myrange(15, 22), [1]],
     "left": [[16], myrange(2, 9)],
@@ -386,7 +447,7 @@ sprites["RANGE_ATTACKER"] = load_sprites({
     "_sizes" : [32, 32]
 });
 
-sprites["MELEE_ATTACKER"] = load_sprites({
+sprites["MELEE_ATTACKER2"] = load_sprites({
     "up" : [myrange(15, 22), [9]],
     "down": [myrange(15, 22), [10]],
     "left": [[20], myrange(2, 9)],
@@ -395,18 +456,20 @@ sprites["MELEE_ATTACKER"] = load_sprites({
     "_sizes" : [32, 32]
 });
 
+// 资源
 sprites["RESOURCE"] = load_sprites({
     "_file" : "imgs/mineral1.png",
 });
-
+// 基地
 sprites["BASE"] = load_sprites({
     "_file" : "imgs/base.png"
 });
-
+// 兵营
 sprites["BARRACKS"] = load_sprites({
     "_file" : "imgs/barracks.png",
 });
 
+// 未执行
 var targets = load_sprites({
     "attack" : [[11], [6]],
     "move" : [[14], [6]],
@@ -423,9 +486,10 @@ sprites["WORKER"] = load_sprites({
     "_sizes" : [32, 32]
 });
 
+// 子弹
 var bullets = load_sprites({
-    "BULLET_READY" : [[7], [0]],
-    "BULLET_EXPLODE1" : [[0], [0]],
+    "BULLET_READY" : [[7], [0]],        // 子弹类型
+    "BULLET_EXPLODE1" : [[0], [0]],     // 三种状态
     "BULLET_EXPLODE2" : [[1], [0]],
     "BULLET_EXPLODE3": [[2], [0]],
     "_file" : "imgs/tiles.png",
@@ -451,36 +515,45 @@ sprites["FLAG_BASE"] = load_sprites({
 });
 
 
+
 var render = function (game) {
     tick = game.tick;
-    ctx.beginPath()
-	ctx.fillStyle = "Black";
-	ctx.font = "15px Arial";
-    var label = "Tick: " + tick;
-	ctx.fillText(label, left_frame_width + 10, 20);
-    ctx.closePath();
+    // Tick
+    // ctx.beginPath()
+	// ctx.fillStyle = "Black";
+	// ctx.font = "15px Arial";
+    // var label = "Tick: " + tick;
+    //    // tick 位置
+	// ctx.fillText(label, left_frame_width + 10, 20);
+    // ctx.closePath();
+
+    // 加载地图
     onMap(game.rts_map);
-    if (! game.spectator) {
+    if (! game.spectator) {         // game.spectator = true   
+        // 切换成玩家视角
+        // rts_map 信息变为某个玩家传来的，下一帧画面生效
        onPlayerSeenUnits(game.rts_map);
     }
 
     var all_units = {};
     var selected = {};
-    for (var i in game.players) {
-        onPlayerStats(game.players[i]);
-    }
-    for (var i in game.units) {
+    // for (var i in game.players) {       // 两个字典，{player_id: 0; resource: 0}
+    //     onPlayerStats(game.players[i]);         // {player_id: 1; resource: 0}    {player_id: 0; resource: 0}
+    // }
+    for (var i in game.units) {             
         var unit = game.units[i];
         all_units[unit.id] = unit;
 
-        var s_units = game.selected_units;
-        var isSelected = (s_units && s_units.indexOf(unit.id) >= 0);
+        var s_units = game.selected_units;                      // 选中后执行
+        var isSelected = (s_units && s_units.indexOf(unit.id) >= 0);           
         if (isSelected) {
             selected[unit.id] = unit;
         }
+        
 
         onUnit(unit, isSelected);
     }
+
     if (dragging && x_down && y_down) {
         ctx.lineWidth = 2;
         ctx.beginPath();
@@ -489,46 +562,61 @@ var render = function (game) {
         ctx.stroke();
         ctx.closePath();
     }
+
+    // 显示子弹
     for (var i in game.bullets) {
+        // console.log(game.bullets[i]);
         onBullet(game.bullets[i]);
     }
-    var len = Object.keys(selected).length;
-    if (len == 1) {
-        var idx = Object.keys(selected)[0];
-        var unit = selected[idx];
-        draw_state(unit);
-    }
-    ctx.beginPath();
-	ctx.fillStyle = "Black";
-	ctx.font = "15px Arial";
-    if (len > 1) {
-        var label = len + " units";
-    	ctx.fillText(label ,left_frame_width + 50, 200);
-    }
-    var label = "Current FPS is " + Math.floor(50 * Math.pow(1.3, speed));
-    ctx.fillText(label, left_frame_width + 50, 570);
-    if (game.replay_length) {
-        range1.value = 100 * game.tick / game.replay_length;
-    }
-
-    var label = "Current progress_percent is " + range1.value;
-    ctx.fillText(label, left_frame_width + 50, 670);
-    ctx.closePath();
+    var len = Object.keys(selected).length;     // 选中几个单位
+    // 选中某单位时，空白处显示当前选择单位的状态
+    // if (len == 1) {
+    //     var idx = Object.keys(selected)[0];     // 选中的单位的编号
+    //     var unit = selected[idx];
+    //     // console.log(unit);      // 获取选中的单位的所有属性
+    //     draw_state(unit);       // 在右边空白出显示具体信息
+    // }
+    // ctx.beginPath();
+	// ctx.fillStyle = "Black";
+    // ctx.font = "15px Arial";
+    // // 选中多个单位时，空白处显示  len + " units"
+    // if (len > 1) {
+    //     var label = len + " units";
+    // 	ctx.fillText(label ,left_frame_width + 50, 200);
+    // }
+    // var label = "Current FPS is " + Math.floor(50 * Math.pow(1.3, speed));
+    // ctx.fillText(label, left_frame_width + 50, 570);
+    // if (game.replay_length) {
+    //     range1.value = 100 * game.tick / game.replay_length;
+    // }
+
+    // var label = "Current progress_percent is " + range1.value;
+    // ctx.fillText(label, left_frame_width + 50, 670);
+    // ctx.closePath();
 };
 
+
 var main = function () {
+        // 建立连接
   dealer = new WebSocket('ws://localhost:8000');
+  // 连接建立成功调用的方法
   dealer.onopen = function(event) {
+            // 连接成功后输出
     console.log("WS Opened.");
   }
 
+
   dealer.onmessage = function (message) {
-    var s = message.data;
+
+    var s = message.data;   
+    // 将 s 转换为 json  存储在 game 中
     var game = JSON.parse(s);
+    //console.log(game);
     ctx.clearRect(0, 0, canvas.width, canvas.height);
     render(game);
   };
 };
 
 var then = Date.now();
+// 开始
 main();
diff --git a/rts/frontend/imgs/tile.png b/rts/frontend/imgs/tile.png
new file mode 100644
index 00000000..d00885f6
Binary files /dev/null and b/rts/frontend/imgs/tile.png differ
diff --git a/rts/frontend/imgs/tile11.png b/rts/frontend/imgs/tile11.png
new file mode 100755
index 00000000..58e23116
Binary files /dev/null and b/rts/frontend/imgs/tile11.png differ
diff --git a/rts/frontend/imgs/tiles.png b/rts/frontend/imgs/tiles.png
index d00885f6..48540e6f 100644
Binary files a/rts/frontend/imgs/tiles.png and b/rts/frontend/imgs/tiles.png differ
diff --git a/rts/game_MC/cmd_specific.cc b/rts/game_MC/cmd_specific.cc
old mode 100644
new mode 100755
index 406480bc..3a7b8382
--- a/rts/game_MC/cmd_specific.cc
+++ b/rts/game_MC/cmd_specific.cc
@@ -16,6 +16,7 @@
 #include "cmd_specific.gen.h"
 
 bool CmdGenerateMap::run(GameEnv *env, CmdReceiver*) {
+   // std::cout<<"CmdGenerateMap"<<std::endl;
     return env->GenerateMap(_num_obstacles, _init_resource) ? true : false;
 }
 
@@ -24,6 +25,7 @@ bool CmdGenerateMap::run(GameEnv *env, CmdReceiver*) {
 
 
 bool CmdGameStartSpecific::run(GameEnv*, CmdReceiver* receiver) {
+   // std::cout<<"CmdGameStartSpecific"<<std::endl;
     const PlayerId player_id = 0;
     const PlayerId enemy_id = 1;
     _CREATE(RESOURCE, PointF(2, 1), player_id);
@@ -46,56 +48,82 @@ bool CmdGameStartSpecific::run(GameEnv*, CmdReceiver* receiver) {
     return true;
 }
 
+//-----------------Test--------------------
 bool CmdGenerateUnit::run(GameEnv *env, CmdReceiver *receiver) {
-    auto f = env->GetRandomFunc();
-    int lr_seed = f(2);
-    int ud_seed = f(2);
-    bool shuffle_lr = (lr_seed == 0);
-    bool shuffle_ud = (ud_seed == 0);
-    auto shuffle_loc = [&] (PointF p, bool b1, bool b2) -> PointF {
-        int x = b1 ? 19 - p.x : p.x;
-        int y = b2 ? 19 - p.y : p.y;
-        return PointF(x, y);
-    };
+    //std::cout<<"CmdGenerateTDUnit"<<std::endl;
+    const PlayerId player_id = 0;
+    const PlayerId enemy_id = 1;
+    // 保护目标
+    _CREATE(BASE, PointF(17, 17), player_id);
+    // 雷达
+    //_CREATE(RANGE_ATTACKER,PointF(4,17),player_id);
+    //_CREATE(RANGE_ATTACKER,PointF(15,17),player_id);
+    //_CREATE(RANGE_ATTACKER,PointF(17,4),player_id);
+    _CREATE(RANGE_ATTACKER,PointF(15,17),player_id);
+    // 炮台
+    //_CREATE(MELEE_ATTACKER,PointF(6,9),player_id);
+   // _CREATE(MELEE_ATTACKER,PointF(13,9),player_id);
+  //  _CREATE(MELEE_ATTACKER,PointF(9,6),player_id);
+     _CREATE(MELEE_ATTACKER,PointF(19,17),player_id);
 
-    receiver->GetGameStats().PickBase(lr_seed * 2 + ud_seed);
-    for (const auto &info : env->GetMap().GetPlayerMapInfo()) {
-        PlayerId id = info.player_id;
-        _CREATE(BASE, shuffle_loc(PointF(info.base_coord), shuffle_lr, shuffle_ud), id);
-        _CREATE(RESOURCE, shuffle_loc(PointF(info.resource_coord), shuffle_lr, shuffle_ud), id);
-        _CHANGE_RES(id, info.initial_resource);
-        //base_locs[id] = PointF(info.base_coord);
-    }
-    // for (size_t i = 0; i < base_locs.size(); ++i) {
-    //    std::cout << "[" << i << "] Baseloc: " << base_locs[i].x << ", " << base_locs[i].y << std::endl;
-    //}
-    auto gen_loc = [&] (int player_id) -> PointF {
-        // Note that we could not write
-        //    PointF( f(8) + ...,  f(8) + ...)
-        // since the result will depend on which f is evaluated first, and will yield different results on
-        // different platform/compiler (e.g., clang and gcc yields different results).
-        // The following implementation is uniquely determined.
-        int x = f(6) + player_id * 10 + 2;
-        int y = f(6) + player_id * 10 + 2;
-        return PointF(x, y);
-    };
-    for (PlayerId player_id = 0; player_id < 2; player_id++) {
-        PlayerId id = player_id;
-        // Generate workers (up to three).
-        for (int i = 0; i < 3; i++) {
-            if (f(10) >= 5) {
-                _CREATE(WORKER, shuffle_loc(gen_loc(player_id), shuffle_lr, shuffle_ud), id);
-            }
-        }
-        if (f(10) >= 8)
-            _CREATE(BARRACKS, shuffle_loc(gen_loc(player_id), shuffle_lr, shuffle_ud), id);
-        if (f(10) >= 5)
-            _CREATE(MELEE_ATTACKER, shuffle_loc(gen_loc(player_id), shuffle_lr, shuffle_ud), id);
-        if (f(10) >= 5)
-            _CREATE(RANGE_ATTACKER, shuffle_loc(gen_loc(player_id), shuffle_lr, shuffle_ud), id);
-    }
+    // enemy
+    _CREATE(BASE,PointF(1, 1),enemy_id);
     return true;
+
 }
 
+//-----------------Test--------------------
+// bool CmdGenerateUnit::run(GameEnv *env, CmdReceiver *receiver) {
+//    // std::cout<<"CmdGenerateUnit"<<std::endl;
+//     auto f = env->GetRandomFunc();
+//     int lr_seed = f(2);
+//     int ud_seed = f(2);
+//     bool shuffle_lr = (lr_seed == 0);
+//     bool shuffle_ud = (ud_seed == 0);
+//     auto shuffle_loc = [&] (PointF p, bool b1, bool b2) -> PointF {
+//         int x = b1 ? 19 - p.x : p.x;
+//         int y = b2 ? 19 - p.y : p.y;
+//         return PointF(x, y);
+//     };
+
+//     receiver->GetGameStats().PickBase(lr_seed * 2 + ud_seed);
+//     for (const auto &info : env->GetMap().GetPlayerMapInfo()) {
+//         PlayerId id = info.player_id;
+//         _CREATE(BASE, shuffle_loc(PointF(info.base_coord), shuffle_lr, shuffle_ud), id);
+//         _CREATE(RESOURCE, shuffle_loc(PointF(info.resource_coord), shuffle_lr, shuffle_ud), id);
+//         _CHANGE_RES(id, info.initial_resource);
+//         //base_locs[id] = PointF(info.base_coord);
+//     }
+//     // for (size_t i = 0; i < base_locs.size(); ++i) {
+//     //    std::cout << "[" << i << "] Baseloc: " << base_locs[i].x << ", " << base_locs[i].y << std::endl;
+//     //}
+//     auto gen_loc = [&] (int player_id) -> PointF {
+//         // Note that we could not write
+//         //    PointF( f(8) + ...,  f(8) + ...)
+//         // since the result will depend on which f is evaluated first, and will yield different results on
+//         // different platform/compiler (e.g., clang and gcc yields different results).
+//         // The following implementation is uniquely determined.
+//         int x = f(6) + player_id * 10 + 2;
+//         int y = f(6) + player_id * 10 + 2;
+//         return PointF(x, y);
+//     };
+//     for (PlayerId player_id = 0; player_id < 2; player_id++) {
+//         PlayerId id = player_id;
+//         // Generate workers (up to three).
+//         for (int i = 0; i < 3; i++) {
+//             if (f(10) >= 5) {
+//                 _CREATE(WORKER, shuffle_loc(gen_loc(player_id), shuffle_lr, shuffle_ud), id);
+//             }
+//         }
+//         if (f(10) >= 8)
+//             _CREATE(BARRACKS, shuffle_loc(gen_loc(player_id), shuffle_lr, shuffle_ud), id);
+//         if (f(10) >= 5)
+//             _CREATE(MELEE_ATTACKER, shuffle_loc(gen_loc(player_id), shuffle_lr, shuffle_ud), id);
+//         if (f(10) >= 5)
+//             _CREATE(RANGE_ATTACKER, shuffle_loc(gen_loc(player_id), shuffle_lr, shuffle_ud), id);
+//     }
+//     return true;
+// }
+
 #undef _CHANGE_RES
 #undef _CREATE
diff --git a/rts/game_MC/game.py b/rts/game_MC/game.py
old mode 100644
new mode 100755
index 40c36437..d2cf20fb
--- a/rts/game_MC/game.py
+++ b/rts/game_MC/game.py
@@ -20,11 +20,11 @@ def __init__(self):
 
     def _define_args(self):
         return [
-            ("use_unit_action", dict(action="store_true")),
-            ("disable_time_decay", dict(action="store_true")),
-            ("use_prev_units", dict(action="store_true")),
-            ("attach_complete_info", dict(action="store_true")),
-            ("feature_type", "ORIGINAL")
+            ("use_unit_action", dict(action="store_true")),     # false
+            ("disable_time_decay", dict(action="store_true")),  # false
+            ("use_prev_units", dict(action="store_true")),      # false 
+            ("attach_complete_info", dict(action="store_true")),# false
+            ("feature_type", "ORIGINAL")                        # "ORIGINAL"
         ]
 
     def _on_gc(self, GC):
@@ -32,11 +32,11 @@ def _on_gc(self, GC):
         opt.use_time_decay = not self.args.disable_time_decay
         opt.save_prev_seen_units = self.args.use_prev_units
         opt.attach_complete_info = self.args.attach_complete_info
-        GC.ApplyExtractorParams(opt)
+        GC.ApplyExtractorParams(opt)  # 设置 MCExtractor
 
         usage = minirts.MCExtractorUsageOptions()
         usage.Set(self.args.feature_type)
-        GC.ApplyExtractorUsage(usage)
+        GC.ApplyExtractorUsage(usage) # 设置 ExtractorUsage
 
     def _unit_action_keys(self):
         if self.args.use_unit_action:
@@ -44,12 +44,12 @@ def _unit_action_keys(self):
         else:
             return []
 
-    def _get_actor_spec(self):
+    def _get_actor_spec(self): #d=定义用于actor的batch字典
         reply_keys = ["V", "pi", "a"]
 
         return dict(
             batchsize=self.args.batchsize,
-            input=dict(T=1, keys=set(["s", "last_r", "terminal"])),
+            input=dict(T=1, keys=set(["s", "last_r", "terminal"])),  # 期待收到 s last_r terminal
             reply=dict(T=1, keys=set(reply_keys + self._unit_action_keys())),
         )
 
@@ -58,7 +58,7 @@ def _get_train_spec(self):
         return dict(
             batchsize=self.args.batchsize,
             input=dict(T=self.args.T, keys=set(keys + self._unit_action_keys())),
-            reply=None
+            reply=None  # train 不需要回复
         )
 
     def _get_reduced_predict(self):
diff --git a/rts/game_MC/game_action.h b/rts/game_MC/game_action.h
old mode 100644
new mode 100755
diff --git a/rts/game_MC/gamedef.cc b/rts/game_MC/gamedef.cc
old mode 100644
new mode 100755
index 1bbde036..cb199805
--- a/rts/game_MC/gamedef.cc
+++ b/rts/game_MC/gamedef.cc
@@ -35,7 +35,7 @@ bool GameDef::CheckAddUnit(RTSMap *_map, UnitType, const PointF& p) const{
 }
 
 void GameDef::GlobalInit() {
-    reg_engine();
+    reg_engine();   //
     reg_engine_specific();
     reg_minirts_specific();
 
@@ -55,21 +55,47 @@ void GameDef::GlobalInit() {
 
 void GameDef::Init() {
     _units.assign(GetNumUnitType(), UnitTemplate());
+    /**
+     * int cost, int hp, int defense, float speed, int att, int att_r, int vis_r,
+        const vector<int> &cds, const vector<CmdType> &l, UnitAttr attr**/
+
+    // _units[RESOURCE] = _C(0, 1000, 1000, 0, 0, 0, 0, vector<int>{0, 0, 0, 0}, vector<CmdType>{}, ATTR_INVULNERABLE);
+    // _units[WORKER] = _C(50, 50, 0, 0.1, 2, 1, 3, vector<int>{0, 10, 40, 40}, vector<CmdType>{MOVE, ATTACK, BUILD, GATHER});
+    // _units[MELEE_ATTACKER] = _C(100, 100, 1, 0.1, 15, 1, 3, vector<int>{0, 15, 0, 0}, vector<CmdType>{MOVE, ATTACK});
+    // _units[RANGE_ATTACKER] = _C(100, 50, 0, 0.2, 10, 5, 5, vector<int>{0, 10, 0, 0}, vector<CmdType>{MOVE, ATTACK});
+    // _units[BARRACKS] = _C(200, 200, 1, 0.0, 0, 0, 5, vector<int>{0, 0, 0, 50}, vector<CmdType>{BUILD});
+    // _units[BASE] = _C(500, 500, 2, 0.0, 0, 0, 5, {0, 0, 0, 50}, vector<CmdType>{BUILD});
+
+ /**   
+ *  定义单位属性
+ *  cost   造价(不需要)
+ *  hp     血量
+ *  defence防御力 
+ *  speed  速度
+ *  att    攻击力
+ *  att_r  攻击距离
+ *  vis_r  可视距离
+ * *                     cost hp    def  sp att att_r vis_r
+ * */
+     
     _units[RESOURCE] = _C(0, 1000, 1000, 0, 0, 0, 0, vector<int>{0, 0, 0, 0}, vector<CmdType>{}, ATTR_INVULNERABLE);
-    _units[WORKER] = _C(50, 50, 0, 0.1, 2, 1, 3, vector<int>{0, 10, 40, 40}, vector<CmdType>{MOVE, ATTACK, BUILD, GATHER});
-    _units[MELEE_ATTACKER] = _C(100, 100, 1, 0.1, 15, 1, 3, vector<int>{0, 15, 0, 0}, vector<CmdType>{MOVE, ATTACK});
-    _units[RANGE_ATTACKER] = _C(100, 50, 0, 0.2, 10, 5, 5, vector<int>{0, 10, 0, 0}, vector<CmdType>{MOVE, ATTACK});
-    _units[BARRACKS] = _C(200, 200, 1, 0.0, 0, 0, 5, vector<int>{0, 0, 0, 50}, vector<CmdType>{BUILD});
-    _units[BASE] = _C(500, 500, 2, 0.0, 0, 0, 5, {0, 0, 0, 50}, vector<CmdType>{BUILD});
+    _units[WORKER] = _C(50, 50, 0, 0.1, 2, 1, 0, vector<int>{0, 10, 40, 40}, vector<CmdType>{MOVE, ATTACK, BUILD, GATHER});
+    _units[MELEE_ATTACKER] = _C(50, 100, 1, 0.1, 15, 10, 0, vector<int>{0, 15, 0, 0}, vector<CmdType>{MOVE,ATTACK});
+    _units[RANGE_ATTACKER] = _C(100, 50, 0, 0.2, 0, 0, 15, vector<int>{0, 0, 0, 0}, vector<CmdType>{});
+    _units[BARRACKS] = _C(200, 200, 1, 0.0, 0, 0, 0, vector<int>{0, 0, 0, 50}, vector<CmdType>{BUILD});
+    _units[BASE] = _C(500, 500, 2, 0.0, 0, 0, 0, {0, 0, 0, 50}, vector<CmdType>{BUILD});
+
+
 }
 
 vector<pair<CmdBPtr, int> > GameDef::GetInitCmds(const RTSGameOptions&) const{
       vector<pair<CmdBPtr, int> > init_cmds;
-      init_cmds.push_back(make_pair(CmdBPtr(new CmdGenerateMap(INVALID, 0, 200)), 1));
-      init_cmds.push_back(make_pair(CmdBPtr(new CmdGenerateUnit(INVALID)), 2));
+      init_cmds.push_back(make_pair(CmdBPtr(new CmdGenerateMap(INVALID, 0, 200)), 1));  // 障碍 资源 
+      init_cmds.push_back(make_pair(CmdBPtr(new CmdGenerateUnit(INVALID)), 2));   
       return init_cmds;
 }
 
+// 通过判断最后一个拥有基地的玩家来确定胜利者
 PlayerId GameDef::CheckWinner(const GameEnv& env, bool /*exceeds_max_tick*/) const {
     return env.CheckBase(BASE);
 }
diff --git a/rts/game_MC/model.py b/rts/game_MC/model.py
old mode 100644
new mode 100755
index 04bf5f61..f1848f3c
--- a/rts/game_MC/model.py
+++ b/rts/game_MC/model.py
@@ -18,12 +18,13 @@ class Model_ActorCritic(Model):
     def __init__(self, args):
         super(Model_ActorCritic, self).__init__(args)
         self._init(args)
+        #self.isPrint = False
 
     def _init(self, args):
         params = args.params
         assert isinstance(params["num_action"], int), "num_action has to be a number. action = " + str(params["num_action"])
         self.params = params
-        self.net = MiniRTSNet(args)
+        self.net = MiniRTSNet(args)   # 卷积神经网络处理输入数据
         last_num_channel = self.net.num_channels[-1]
 
         if self.params.get("model_no_spatial", False):
@@ -31,9 +32,11 @@ def _init(self, args):
             linear_in_dim = last_num_channel
         else:
             linear_in_dim = last_num_channel * 25
+        
+        
 
-        self.linear_policy = nn.Linear(linear_in_dim, params["num_action"])
-        self.linear_value = nn.Linear(linear_in_dim, 1)
+        self.linear_policy = nn.Linear(linear_in_dim, params["num_action"])  # 策略函数
+        self.linear_value = nn.Linear(linear_in_dim, 1)   # 价值函数
 
         self.relu = nn.LeakyReLU(0.1)
 
@@ -49,13 +52,21 @@ def get_define_args():
     def forward(self, x):
         if self.params.get("model_no_spatial", False):
             # Replace a complicated network with a simple retraction.
-            # Input: batchsize, channel, height, width
+            # Input: batchsize, channel, height, width  Batch Object
             xreduced = x["s"].sum(2).sum(3).squeeze()
             xreduced[:, self.num_unit:] /= 20 * 20
             output = self._var(xreduced)
         else:
             output = self.net(self._var(x["s"]))
-
+        
+        #decide = self.decision(output)
+        #if not self.isPrint:
+            #print("x: ",x.batch)
+            # print("output: ",output)
+            # print("decision: ",decide)
+            # print("net: ",self)
+            #self.isPrint = True
+        #return decide
         return self.decision(output)
 
     def decision(self, h):
diff --git a/rts/game_MC/python_options.h b/rts/game_MC/python_options.h
old mode 100644
new mode 100755
index 52629a02..1eb52308
--- a/rts/game_MC/python_options.h
+++ b/rts/game_MC/python_options.h
@@ -22,6 +22,12 @@ struct GameState {
     using State = GameState;
     using Data = GameState;
 
+    /**
+     * 测试获取基地位置信息
+     * */
+     //float base_x, base_y;
+
+
     int32_t id;
     int32_t seq;
     int32_t game_counter;
@@ -52,7 +58,7 @@ struct GameState {
     std::vector<int64_t> uloc, tloc;
     std::vector<int64_t> bt, ct;
 
-    // Also we need to save distributions.
+    // Also we need to save distributions.  概率？
     std::vector<float> uloc_prob, tloc_prob;
     std::vector<float> bt_prob, ct_prob;
 
@@ -133,6 +139,7 @@ struct GameState {
 
     // These fields are used to exchange with Python side using tensor interface.
     DECLARE_FIELD(GameState, id, a, V, pi, last_r, s, rv, terminal, seq, game_counter, last_terminal, uloc, tloc, bt, ct, uloc_prob, tloc_prob, bt_prob, ct_prob, reduced_s, reduced_next_s);
+    //DECLARE_FIELD(GameState, id, a, V, pi, last_r, s, rv, terminal, seq, game_counter, last_terminal, uloc, tloc, bt, ct, uloc_prob, tloc_prob, bt_prob, ct_prob, reduced_s, reduced_next_s,base_x,base_y);
     REGISTER_PYBIND_FIELDS(id);
 };
 
diff --git a/rts/game_MC/python_wrapper.cc b/rts/game_MC/python_wrapper.cc
old mode 100644
new mode 100755
index 3788527c..013453f5
--- a/rts/game_MC/python_wrapper.cc
+++ b/rts/game_MC/python_wrapper.cc
@@ -33,32 +33,36 @@ class GameContext {
         GameDef::GlobalInit();
         _context.reset(new GC{context_options, options});
 
-        _num_frames_in_state = 1;
+        _num_frames_in_state = 1;   
         for (const AIOptions& opt : options.ai_options) {
             _num_frames_in_state = max(_num_frames_in_state, opt.num_frames_in_state);
         }
     }
 
     void Start() {
+        std::cout<<"--------------GameContext Start-----------"<<std::endl;
         _context->Start(
             [this](int game_idx, const ContextOptions &context_options, const PythonOptions &options, const elf::Signal &signal, Comm *comm) {
                     auto params = this->GetParams();
+                    if(game_idx == 1){
+                      std::cout<<"game_"<<game_idx<<" run GameStartFunc"<<std::endl;
+                    }
                     this->_wrapper.thread_main(game_idx, context_options, options, signal, &params, comm);
             });
     }
 
     std::map<std::string, int> GetParams() const {
         return std::map<std::string, int>{
-            { "num_action", GameDef::GetNumAction() },
-            { "num_unit_type", GameDef::GetNumUnitType() },
-            { "num_planes_per_time_stamp", MCExtractor::Size() },
-            { "num_planes", MCExtractor::Size() * _num_frames_in_state },
-            { "resource_dim", 2 * NUM_RES_SLOT },
-            { "max_unit_cmd", _context->options().max_unit_cmd },
-            { "map_x", _context->options().map_size_x },
-            { "map_y", _context->options().map_size_y },
-            { "num_cmd_type", CmdInput::CI_NUM_CMDS },
-            { "reduced_dim", MCExtractor::Size() * 5 * 5 }
+            { "num_action", GameDef::GetNumAction() },       // 9
+            { "num_unit_type", GameDef::GetNumUnitType() },   // 6  
+            { "num_planes_per_time_stamp", MCExtractor::Size() },  // 22  每一个时间戳中的 planes数？
+            { "num_planes", MCExtractor::Size() * _num_frames_in_state },  //  22 每一个状态包含一帧的数据
+            { "resource_dim", 2 * NUM_RES_SLOT }, // 10
+            { "max_unit_cmd", _context->options().max_unit_cmd }, // 1
+            { "map_x", _context->options().map_size_x }, // 20
+            { "map_y", _context->options().map_size_y }, // 20  
+            { "num_cmd_type", CmdInput::CI_NUM_CMDS },   // 4
+            { "reduced_dim", MCExtractor::Size() * 5 * 5 }// 22*5*5
         };
     }
 
@@ -88,7 +92,9 @@ class GameContext {
         else if (key == "ct_prob") return EntryInfo(key, type_name, { max_unit_cmd, CmdInput::CI_NUM_CMDS });
         else if (key == "reduced_s") return EntryInfo(key, type_name, { reduced_size });
         else if (key == "reduced_next_s") return EntryInfo(key, type_name, { reduced_size });
+        else if (key == "base_x" || key == "base_y") return EntryInfo(key, type_name);
 
+ 
         return EntryInfo();
     }
 
diff --git a/rts/game_MC/state_feature.cc b/rts/game_MC/state_feature.cc
old mode 100644
new mode 100755
index 5ff62a8a..3e822adc
--- a/rts/game_MC/state_feature.cc
+++ b/rts/game_MC/state_feature.cc
@@ -28,6 +28,14 @@ void MCExtractor::SaveInfo(const RTSState &s, PlayerId player_id, GameState *gs)
     gs->terminal = s.env().GetTermination() ? 1 : 0;
 
     gs->last_r = 0.0;
+
+    // 测试获取基地位置
+    // UnitId baseId = s.env().FindClosestBase(player_id);
+    //const Unit* base = s.env().GetUnit(s.env().FindClosestBase(player_id));
+    //gs->base_x = s.env().GetUnit(s.env().FindClosestBase(player_id))->GetPointF().x;
+    //gs->base_y = s.env().GetUnit(s.env().FindClosestBase(player_id))->GetPointF().y;
+    // 测试获取基地位置
+
     int winner = s.env().GetWinnerId();
     if (winner != INVALID) {
       if (winner == player_id) gs->last_r = 1.0;
diff --git a/rts/game_MC/state_feature.h b/rts/game_MC/state_feature.h
index f627a2ab..ea81b6a1 100644
--- a/rts/game_MC/state_feature.h
+++ b/rts/game_MC/state_feature.h
@@ -22,6 +22,7 @@ using namespace std;
 
 #define NUM_RES_SLOT 5
 
+// MCExtractor设置
 struct MCExtractorOptions {
     bool use_time_decay = true;
     bool save_prev_seen_units = false;
@@ -122,15 +123,15 @@ class MCExtractorInfo {
     void Reset(const MCExtractorOptions &opt) {
         extractors_.clear();
 
-        int num_unit_type = GameDef::GetNumUnitType();
+        int num_unit_type = GameDef::GetNumUnitType();  
         total_dim_ = 0;
-        total_dim_ += _add_extractor("UnitType", new ExtractorSpan(total_dim_, num_unit_type));
-        total_dim_ += _add_extractor("Feature", new ExtractorSpan(total_dim_, NUM_FEATURE));
+        total_dim_ += _add_extractor("UnitType", new ExtractorSpan(total_dim_, num_unit_type));  // 6
+        total_dim_ += _add_extractor("Feature", new ExtractorSpan(total_dim_, NUM_FEATURE));     // 4
 
         std::initializer_list<int> ticks = { 200, 500, 1000, 2000, 5000, 10000 };
 
         if (opt.use_time_decay) {
-            total_dim_ += _add_extractor("HistBin", new ExtractorSeq(total_dim_, ticks));
+            total_dim_ += _add_extractor("HistBin", new ExtractorSeq(total_dim_, ticks));   // 6+1
         }
 
         if (opt.save_prev_seen_units) {
@@ -138,7 +139,7 @@ class MCExtractorInfo {
             total_dim_ += _add_extractor("HistBinPrevSeen", new ExtractorSeq(total_dim_, ticks));
         }
 
-        total_dim_ += _add_extractor("Resource", new ExtractorSpan(total_dim_, NUM_RES_SLOT, 50));
+        total_dim_ += _add_extractor("Resource", new ExtractorSpan(total_dim_, NUM_RES_SLOT, 50));  // 5
     }
 
     int size() const { return total_dim_; }
@@ -150,7 +151,7 @@ class MCExtractorInfo {
 
 private:
     int total_dim_;
-    map<string, unique_ptr<Extractor>> extractors_;
+    map<string, unique_ptr<Extractor>> extractors_;   //
 
     int _add_extractor(const std::string &name, Extractor *e) {
         extractors_[name].reset(e);
diff --git a/rts/game_MC/trunk.py b/rts/game_MC/trunk.py
old mode 100644
new mode 100755
index e685481a..a226026d
--- a/rts/game_MC/trunk.py
+++ b/rts/game_MC/trunk.py
@@ -16,6 +16,7 @@ def __init__(self, args, output1d=True):
         super(MiniRTSNet, self).__init__(args)
         self._init(args)
         self.output1d = output1d
+        #self.isPrint = False
 
     def _init(self, args):
         self.m = args.params.get("num_planes_per_time_stamp", 13)
@@ -64,7 +65,13 @@ def get_define_args():
     def forward(self, input):
         # BN and LeakyReLU are from Wendy's code.
         x = input.view(input.size(0), self.input_channel, self.mapy, self.mapx)
-
+        # if not self.isPrint:
+        #     print("input size: ",input.size())
+        #     print("input: ",input)
+        #     print("x size:",x.size())
+        #     print("x: ",x)
+        #     print("Net",self)
+        #     self.isPrint = True
         counts = Counter()
         for i in range(len(self.arch)):
             if self.arch[i] == "c":
@@ -78,5 +85,10 @@ def forward(self, input):
 
         if self.output1d:
             x = x.view(x.size(0), -1)
-
-        return x
+        
+        # if not self.isPrint:
+        #     print("x",x)
+        #     print("x.size ",x.size())
+        #     self.isPrint = True
+        
+        return x     # 64 x 550
diff --git a/rts/game_MC/wrapper_callback.cc b/rts/game_MC/wrapper_callback.cc
old mode 100644
new mode 100755
index 40a06ab4..80195b71
--- a/rts/game_MC/wrapper_callback.cc
+++ b/rts/game_MC/wrapper_callback.cc
@@ -65,11 +65,11 @@ void WrapperCallbacks::OnGameOptions(RTSGameOptions *rts_options) {
     rts_options->handicap_level = _options.handicap_level;
 }
 
-void WrapperCallbacks::OnGameInit(RTSGame *game, const std::map<std::string, int> *more_params) {
+void WrapperCallbacks::OnGameInit(RTSGame *game, const std::map<std::string, int> *more_params,bool isPrint) {
     // std::cout << "Initialize opponent" << std::endl;
     std::vector<AI *> ais;
     for (const AIOptions &ai_opt : _options.ai_options) {
-        Context::AIComm *ai_comm = new Context::AIComm(_game_idx, _comm);
+        Context::AIComm *ai_comm = new Context::AIComm(_game_idx, _comm);  //设置 AI 和 Main_Loop通信的工具
         _ai_comms.emplace_back(ai_comm);
         initialize_ai_comm(*ai_comm, more_params);
         ais.push_back(get_ai(_game_idx, _context_options.mcts_options, ai_opt, ai_comm));
@@ -85,12 +85,17 @@ void WrapperCallbacks::OnGameInit(RTSGame *game, const std::map<std::string, int
         std::shuffle(orders.begin(), orders.end(), g);
         // cout << "[" << _game_idx << "] Shuffle is done: " << orders[0] << " " << orders[1] << endl;
     }
-
+    
     for (size_t i = 0; i < ais.size(); ++i) {
         int idx = orders[i];
         game->AddBot(ais[idx], _options.ai_options[idx].fs);
         game->GetState().AppendPlayer("player " + std::to_string(idx));
     }
+    // 输出玩家信息
+    if(isPrint){
+        std::cout<<"ais size = "<<ais.size()<<std::endl;
+        std::cout<<game->GetState().env().PrintPlayerInfo()<<std::endl;
+    }
 }
 
 void WrapperCallbacks::OnEpisodeStart(int k, std::mt19937 *rng, RTSGame *game) {
diff --git a/rts/game_MC/wrapper_callback.h b/rts/game_MC/wrapper_callback.h
old mode 100644
new mode 100755
index 1cc4c87c..0700c918
--- a/rts/game_MC/wrapper_callback.h
+++ b/rts/game_MC/wrapper_callback.h
@@ -36,7 +36,7 @@ class WrapperCallbacks {
     }
 
     void OnGameOptions(RTSGameOptions *rts_options);
-    void OnGameInit(RTSGame *game, const std::map<std::string, int> *more_params);
+    void OnGameInit(RTSGame *game, const std::map<std::string, int> *more_params,bool isPrint = false);
 
     void OnEpisodeStart(int k, std::mt19937 *rng, RTSGame *game);
 };
diff --git a/train.py b/train.py
old mode 100644
new mode 100755
index 8b30c168..17fcba0d
--- a/train.py
+++ b/train.py
@@ -16,8 +16,26 @@
 if __name__ == '__main__':
     trainer = Trainer()
     runner = SingleProcessRun()
-    env, all_args = load_env(os.environ, trainer=trainer, runner=runner)
-
+    '''
+        设置环境参数
+            env : dice [
+                game = game game.py
+                sampler = sampler
+                model_loaders =  存ModelLoader类的字典，每一个ModelLoader中有一个model
+                mi=mi
+            ]
+
+            all_args: 所有定义的参数
+    '''
+    env, all_args = load_env(os.environ, trainer=trainer, runner=runner)  
+    
+    '''
+       GC = GCWrapper(GC, co, desc, gpu=args.gpu, use_numpy=False, params=params)
+          GC      /gameMC/python_wrapper.cc GameContext
+          co      /elf/python_options_utils_cpp.h ContextOption
+          desc    actor 和 critic 的 Batch定义  
+          {'actor': {'batchsize': 64, 'input': {'T': 1, 'keys': {'s', 'terminal', 'last_r'}}, 'reply': {'T': 1, 'keys': {'a', 'pi', 'V'}}}}
+    '''
     GC = env["game"].initialize()
 
     model = env["model_loaders"][0].load_model(GC.params)
@@ -27,7 +45,15 @@
     trainer.setup(sampler=env["sampler"], mi=env["mi"], rl_method=env["method"])
 
     GC.reg_callback("train", trainer.train)
+    # def train(batch):
+    #   print(batch)
+    #   import pdb
+    #   pdb.set_trace()
+    #   return trainer.train(batch)
+
+    # GC.reg_callback("train", train)
     GC.reg_callback("actor", trainer.actor)
+
     runner.setup(GC, episode_summary=trainer.episode_summary,
                 episode_start=trainer.episode_start)
 
diff --git a/train_minirts.sh b/train_minirts.sh
index 1fa620da..fee80a44 100755
--- a/train_minirts.sh
+++ b/train_minirts.sh
@@ -1,3 +1,3 @@
 #!/bin/bash
 
-game=./rts/game_MC/game model=actor_critic model_file=./rts/game_MC/model python3 train.py --batchsize 128 --freq_update 1 --players "type=AI_NN,fs=50,args=backup/AI_SIMPLE|start/500|decay/0.99;type=AI_SIMPLE,fs=20" --num_games 1024 --tqdm --T 20 --additional_labels id,last_terminal --trainer_stats winrate --keys_in_reply V "$@"
+game=./rts/game_MC/game model=actor_critic model_file=./rts/game_MC/model python3 train.py --batchsize 32 --freq_update 1 --players "type=AI_NN,fs=50,args=backup/AI_SIMPLE|start/500|decay/0.99;type=AI_SIMPLE,fs=20" --num_games 1024 --tqdm --T 20 --additional_labels id,last_terminal --trainer_stats winrate --keys_in_reply V "$@"