✨: rm sensitive info in config.yaml, add readme of makefile. support old model_path config

.gitignore

@@ -17,4 +17,5 @@ compile_commands.json
 *dist/
 ktransformers/server/local_store/
 ktransformers/server_test1.db
-*.patch
\ No newline at end of file
+*.patch
+img/
\ No newline at end of file

README.md

@@ -151,7 +151,7 @@ Some preparation:
      ```
      install.bat
      ```
-
+4. If you are developer, you can make use of the makefile to compile and format the code. <br> the detailed usage of makefile is [here](./doc/en/makefile_usage.md) 
 <h3>Local Chat</h3>
 We provide a simple command-line local chat Python script that you can run for testing.
 

doc/en/makefile_usage.md

+# Makefile
+## Target
+### flake_find:
+```bash
+make flake_find
+```
+find all the python files under ./ktransformers dir and find the Error, Warning, Fatal... (their codes) into a list that are not consistent with the pep8 standard. For now we have get all this list in the .flake8 file's extend-ignore section in order to let flakes8 ignore them temporarily.(we may improve them in the future)
+### format:
+```bash
+make format
+```
+we use black to format all the python files under ./ktransformers dir. It obeys the pep8 standard 
+but we modify the line length to 120 by add 
+```toml
+[tool.black]
+line-length = 120
+preview = true
+unstable = true
+```
+in the pyproject.toml file.
+
+### dev_install:
+```bash
+make dev_install
+```
+install the package in the development mode. It means that the package is installed in the editable mode. So if you modify the code, you don't need to reinstall the package. We recommend the developer to use this method to install the package.
\ No newline at end of file

ktransformers/configs/config.yaml

@@ -24,9 +24,7 @@ model:
   type: ktransformers
 
   name: DeepSeek-Coder-V2-Instruct
-  # path: /mnt/data/model/DeepSeek-Coder-V2-Instruct/
   path: deepseek-ai/DeepSeek-V2-Lite-Chat
-  # gguf_path: /mnt/data/model/DeepSeek-Coder-V2-GGUF-WJH/
   gguf_path: ./DeepSeek-V2-Lite-Chat-GGUF
 
   device: cuda:0

ktransformers/server/args.py

@@ -14,7 +14,8 @@ class ArgumentParser:
         parser.add_argument("--ssl_certfile", type=str)
         parser.add_argument("--web", type=bool, default=self.cfg.mount_web)
         parser.add_argument("--model_name", type=str, default=self.cfg.model_name)
-        parser.add_argument("--model_dir", type=str, default=self.cfg.model_dir)
+        parser.add_argument("--model_dir", type=str)
+        parser.add_argument("--model_path", type=str)
         parser.add_argument(
             "--device", type=str, default=self.cfg.model_device, help="Warning: Abandoning this parameter"
         )
@@ -100,6 +101,16 @@ class ArgumentParser:
         parser.add_argument("--prompt_file", type=str, default=self.cfg.prompt_file)
 
         args = parser.parse_args()
+        if (args.model_dir is not None):
+            if (args.model_path is not None):
+                # if pass model_dir and model_path, we use model_path
+                args.model_dir = args.model_path
+            else:
+                # if only pass model_dir, we use model_dir
+                args.model_path = args.model_dir
+        else:
+            args.model_dir = self.cfg.model_dir
+            args.model_path = self.cfg.model_path
         # set config from args
         for key, value in vars(args).items():
             if value is not None and hasattr(self.cfg, key):

ktransformers/server/config/config.py

@@ -88,6 +88,8 @@ class Config(metaclass=Singleton):
         self.model: dict = cfg.get("model", {})
         self.backend_type: str = self.model.get("type", "transformers")
         self.model_dir: str = self.model.get("path", "")
+        # to make sure it consistent with previous version
+        self.model_path: str = self.model_dir
         self.model_name: str = self.model.get("name", "")
         self.model_device: str = self.model.get("device", "cuda:0")
         self.gguf_path: Optional[str] = self.model.get("gguf_path", None)