auto-coder 0.1.251__tar.gz → 0.1.253__tar.gz

{auto_coder-0.1.251 → auto_coder-0.1.253}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: auto-coder
-Version: 0.1.251
+Version: 0.1.253
 Summary: AutoCoder: AutoCoder
 Author: allwefantasy
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
@@ -26,7 +26,7 @@ Requires-Dist: tabulate
 Requires-Dist: jupyter_client
 Requires-Dist: prompt-toolkit
 Requires-Dist: tokenizers
-Requires-Dist: byzerllm[saas]>=0.1.161
+Requires-Dist: byzerllm[saas]>=0.1.163
 Requires-Dist: patch
 Requires-Dist: diff_match_patch
 Requires-Dist: GitPython

{auto_coder-0.1.251 → auto_coder-0.1.253}/src/auto_coder.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: auto-coder
-Version: 0.1.251
+Version: 0.1.253
 Summary: AutoCoder: AutoCoder
 Author: allwefantasy
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
@@ -26,7 +26,7 @@ Requires-Dist: tabulate
 Requires-Dist: jupyter_client
 Requires-Dist: prompt-toolkit
 Requires-Dist: tokenizers
-Requires-Dist: byzerllm[saas]>=0.1.161
+Requires-Dist: byzerllm[saas]>=0.1.163
 Requires-Dist: patch
 Requires-Dist: diff_match_patch
 Requires-Dist: GitPython

{auto_coder-0.1.251 → auto_coder-0.1.253}/src/auto_coder.egg-info/SOURCES.txt

@@ -57,6 +57,7 @@ src/autocoder/common/command_templates.py
 src/autocoder/common/const.py
 src/autocoder/common/files.py
 src/autocoder/common/git_utils.py
+src/autocoder/common/global_cancel.py
 src/autocoder/common/image_to_page.py
 src/autocoder/common/interpreter.py
 src/autocoder/common/llm_rerank.py
@@ -144,6 +145,7 @@ src/autocoder/utils/request_event_queue.py
 src/autocoder/utils/request_queue.py
 src/autocoder/utils/rest.py
 src/autocoder/utils/tests.py
+src/autocoder/utils/thread_utils.py
 src/autocoder/utils/types.py
 src/autocoder/utils/auto_coder_utils/__init__.py
 src/autocoder/utils/auto_coder_utils/chat_stream_out.py

{auto_coder-0.1.251 → auto_coder-0.1.253}/src/auto_coder.egg-info/requires.txt

@@ -16,7 +16,7 @@ tabulate
 jupyter_client
 prompt-toolkit
 tokenizers
-byzerllm[saas]>=0.1.161
+byzerllm[saas]>=0.1.163
 patch
 diff_match_patch
 GitPython

{auto_coder-0.1.251 → auto_coder-0.1.253}/src/autocoder/auto_coder.py

@@ -6,6 +6,7 @@ from autocoder.common import git_utils, code_auto_execute
 from autocoder.utils.llm_client_interceptors import token_counter_interceptor
 from autocoder.db.store import Store
 
+from autocoder.utils.llms import get_llm_names
 from autocoder.utils.queue_communicate import (
     queue_communicate,
     CommunicateEvent,
@@ -765,6 +766,23 @@ def main(input_args: Optional[List[str]] = None):
                 )
                 llm.setup_sub_client("planner_model", planner_model)
 
+            if args.commit_model:
+                model_name = args.commit_model.strip()
+                model_info = models_module.get_model_by_name(model_name)
+                commit_model = byzerllm.SimpleByzerLLM(default_model_name=model_name)
+                commit_model.deploy(
+                    model_path="",
+                    pretrained_model_type=model_info["model_type"],
+                    udf_name=model_name,
+                    infer_params={
+                        "saas.base_url": model_info["base_url"],
+                        "saas.api_key": model_info["api_key"],
+                        "saas.model": model_info["model_name"],
+                        "saas.is_reasoning": model_info["is_reasoning"]
+                    }
+                )
+                llm.setup_sub_client("commit_model", commit_model)    
+
             if args.designer_model:
                 model_name = args.designer_model.strip()
                 model_info = models_module.get_model_by_name(model_name)
@@ -875,6 +893,11 @@ def main(input_args: Optional[List[str]] = None):
                 designer_model.setup_default_model_name(args.designer_model)
                 llm.setup_sub_client("designer_model", designer_model)
 
+            if args.commit_model:
+                commit_model = byzerllm.ByzerLLM()
+                commit_model.setup_default_model_name(args.commit_model)
+                llm.setup_sub_client("commit_model", commit_model)
+
     else:
         llm = None
 
@@ -1086,8 +1109,7 @@ def main(input_args: Optional[List[str]] = None):
             return
 
         elif raw_args.agent_command == "chat":
-            from autocoder.rag.rag_entry import RAGFactory
-
+            
             memory_dir = os.path.join(args.source_dir, ".auto-coder", "memory")
             os.makedirs(memory_dir, exist_ok=True)
             memory_file = os.path.join(memory_dir, "chat_history.json")
@@ -1312,6 +1334,7 @@ def main(input_args: Optional[List[str]] = None):
             start_time = time.time()    
                         
             if "rag" in args.action:
+                from autocoder.rag.rag_entry import RAGFactory
                 args.enable_rag_search = True
                 args.enable_rag_context = False
                 rag = RAGFactory.get_rag(llm=chat_llm, args=args, path="")
@@ -1325,7 +1348,8 @@ def main(input_args: Optional[List[str]] = None):
                 response = mcp_server.send_request(
                     McpRequest(
                         query=args.query,
-                        model=args.inference_model or args.model
+                        model=args.inference_model or args.model,
+                        product_mode=args.product_mode
                     )
                 )
                 v = [[response.result,None]]
@@ -1348,9 +1372,7 @@ def main(input_args: Optional[List[str]] = None):
                 )              
 
 
-            model_name = getattr(chat_llm, 'default_model_name', None)
-            if not model_name:
-                model_name = "unknown(without default model name)"    
+            model_name = ",".join(get_llm_names(chat_llm))
             
             assistant_response, last_meta = stream_out(
                     v, 
@@ -1364,10 +1386,13 @@ def main(input_args: Optional[List[str]] = None):
             if last_meta:
                 elapsed_time = time.time() - start_time
                 printer = Printer()
+                speed = last_meta.generated_tokens_count / elapsed_time
                 printer.print_in_terminal("stream_out_stats", 
                                     elapsed_time=elapsed_time,
+                                    first_token_time=last_meta.first_token_time,
                                     input_tokens=last_meta.input_tokens_count,
-                                    output_tokens=last_meta.generated_tokens_count)
+                                    output_tokens=last_meta.generated_tokens_count,
+                                    speed=round(speed, 2))
             
             chat_history["ask_conversation"].append(
                 {"role": "assistant", "content": assistant_response}

{auto_coder-0.1.251 → auto_coder-0.1.253}/src/autocoder/auto_coder_rag.py

@@ -32,7 +32,10 @@ if platform.system() == "Windows":
     init()
 
 
-def initialize_system():
+def initialize_system(args):
+    if args.product_mode == "lite":
+        return
+
     print(f"\n\033[1;34m{get_message('initializing')}\033[0m")
 
     def print_status(message, status):
@@ -316,13 +319,30 @@ def main(input_args: Optional[List[str]] = None):
         action="store_true",
         help="Whether to return responses without contexts. only works when pro plugin is installed",
     )
-    serve_parser.add_argument(
-        "--data_cells_max_num",
+
+    serve_parser.add_argument("--data_cells_max_num",
         type=int,
         default=2000,
         help="Maximum number of data cells to process",
     )
 
+    serve_parser.add_argument(
+        "--product_mode",
+        type=str,
+        default="pro",
+        help="The mode of the auto-coder.rag, lite/pro default is pro",
+    )
+    serve_parser.add_argument(
+        "--lite",
+        action="store_true",
+        help="Run in lite mode (equivalent to --product_mode=lite)",
+    )
+    serve_parser.add_argument(
+        "--pro",
+        action="store_true",
+        help="Run in pro mode (equivalent to --product_mode=pro)",
+    )
+
     serve_parser.add_argument(
         "--recall_model",
         default="",
@@ -373,6 +393,22 @@ def main(input_args: Optional[List[str]] = None):
     # Tools command
     tools_parser = subparsers.add_parser("tools", help="Various tools")
     tools_subparsers = tools_parser.add_subparsers(dest="tool", help="Available tools")
+    tools_parser.add_argument(
+        "--product_mode",
+        type=str,
+        default="pro",
+        help="The mode of the auto-coder.rag, lite/pro default is pro",
+    )
+    tools_parser.add_argument(
+        "--lite",
+        action="store_true",
+        help="Run in lite mode (equivalent to --product_mode=lite)",
+    )
+    tools_parser.add_argument(
+        "--pro",
+        action="store_true",
+        help="Run in pro mode (equivalent to --product_mode=pro)",
+    )
 
     # Count tool
     count_parser = tools_subparsers.add_parser("count", help="Count tokens in a file")
@@ -431,8 +467,15 @@ def main(input_args: Optional[List[str]] = None):
             benchmark_byzerllm(args.model, args.parallel, args.rounds, args.query)
 
     elif args.command == "serve":
+         # Handle lite/pro flags
+        if args.lite:
+            args.product_mode = "lite"
+        elif args.pro:
+            args.product_mode = "pro"
+
         if not args.quick:
-            initialize_system()
+            initialize_system(args)
+       
         server_args = ServerArgs(
             **{
                 arg: getattr(args, arg)
@@ -448,7 +491,11 @@ def main(input_args: Optional[List[str]] = None):
             }
         )
 
-        if auto_coder_args.enable_hybrid_index:
+
+        if auto_coder_args.enable_hybrid_index and args.product_mode == "lite":
+            raise Exception("Hybrid index is not supported in lite mode")
+
+        if auto_coder_args.enable_hybrid_index and args.product_mode == "pro":
             # 尝试连接storage
             try:
                 from byzerllm.apps.byzer_storage.simple_api import ByzerStorage
@@ -460,36 +507,107 @@ def main(input_args: Optional[List[str]] = None):
                     "When enable_hybrid_index is true, ByzerStorage must be started"
                 )
                 logger.error("Please run 'byzerllm storage start' first")
-                return
-        else:
-            byzerllm.connect_cluster(address=args.ray_address)
-        llm = byzerllm.ByzerLLM()
-        llm.setup_default_model_name(args.model)
-
-        # Setup sub models if specified
-        if args.recall_model:
-            recall_model = byzerllm.ByzerLLM()
-            recall_model.setup_default_model_name(args.recall_model)
-            llm.setup_sub_client("recall_model", recall_model)
-
-        if args.chunk_model:
-            chunk_model = byzerllm.ByzerLLM()
-            chunk_model.setup_default_model_name(args.chunk_model)
-            llm.setup_sub_client("chunk_model", chunk_model)
+                return                        
+        
+        
 
-        if args.qa_model:
-            qa_model = byzerllm.ByzerLLM()
-            qa_model.setup_default_model_name(args.qa_model)
-            llm.setup_sub_client("qa_model", qa_model)
+        if args.product_mode == "pro":
+            byzerllm.connect_cluster(address=args.ray_address)
+            llm = byzerllm.ByzerLLM()
+            llm.skip_nontext_check = True
+            llm.setup_default_model_name(args.model)
+
+            # Setup sub models if specified
+            if args.recall_model:
+                recall_model = byzerllm.ByzerLLM()
+                recall_model.setup_default_model_name(args.recall_model)
+                recall_model.skip_nontext_check = True
+                llm.setup_sub_client("recall_model", recall_model)
+
+            if args.chunk_model:
+                chunk_model = byzerllm.ByzerLLM()
+                chunk_model.setup_default_model_name(args.chunk_model)
+                llm.setup_sub_client("chunk_model", chunk_model)
+
+            if args.qa_model:
+                qa_model = byzerllm.ByzerLLM()
+                qa_model.setup_default_model_name(args.qa_model)
+                qa_model.skip_nontext_check = True
+                llm.setup_sub_client("qa_model", qa_model)
+
+            # 当启用hybrid_index时,检查必要的组件
+            if auto_coder_args.enable_hybrid_index:
+                if not llm.is_model_exist("emb"):
+                    logger.error(
+                        "When enable_hybrid_index is true, an 'emb' model must be deployed"
+                    )
+                    return
+                llm.setup_default_emb_model_name("emb")
+
+        elif args.product_mode == "lite":
+            from autocoder import models as models_module
+            model_info = models_module.get_model_by_name(args.model)
+            llm = byzerllm.SimpleByzerLLM(default_model_name=args.model)
+            llm.deploy(
+                model_path="",
+                pretrained_model_type=model_info["model_type"],
+                udf_name=args.model,
+                infer_params={
+                    "saas.base_url": model_info["base_url"],
+                    "saas.api_key": model_info["api_key"],
+                    "saas.model": model_info["model_name"],
+                    "saas.is_reasoning": model_info["is_reasoning"]
+                }
+            )
 
-        # 当启用hybrid_index时,检查必要的组件
-        if auto_coder_args.enable_hybrid_index:
-            if not llm.is_model_exist("emb"):
-                logger.error(
-                    "When enable_hybrid_index is true, an 'emb' model must be deployed"
+            # Setup sub models if specified
+            if args.recall_model:
+                model_info = models_module.get_model_by_name(args.recall_model)
+                recall_model = byzerllm.SimpleByzerLLM(default_model_name=args.recall_model)
+                recall_model.deploy(
+                    model_path="",
+                    pretrained_model_type=model_info["model_type"],
+                    udf_name=args.recall_model,
+                    infer_params={
+                        "saas.base_url": model_info["base_url"],
+                        "saas.api_key": model_info["api_key"],
+                        "saas.model": model_info["model_name"],
+                        "saas.is_reasoning": model_info["is_reasoning"]
+                    }
                 )
-                return
-            llm.setup_default_emb_model_name("emb")
+                llm.setup_sub_client("recall_model", recall_model)
+
+            if args.chunk_model:
+                model_info = models_module.get_model_by_name(args.chunk_model)
+                chunk_model = byzerllm.SimpleByzerLLM(default_model_name=args.chunk_model)
+                chunk_model.deploy(
+                    model_path="",
+                    pretrained_model_type=model_info["model_type"],
+                    udf_name=args.chunk_model,
+                    infer_params={
+                        "saas.base_url": model_info["base_url"],
+                        "saas.api_key": model_info["api_key"],
+                        "saas.model": model_info["model_name"],
+                        "saas.is_reasoning": model_info["is_reasoning"]
+                    }
+                )
+                llm.setup_sub_client("chunk_model", chunk_model)
+
+            if args.qa_model:
+                model_info = models_module.get_model_by_name(args.qa_model)
+                qa_model = byzerllm.SimpleByzerLLM(default_model_name=args.qa_model)
+                qa_model.deploy(
+                    model_path="",
+                    pretrained_model_type=model_info["model_type"],
+                    udf_name=args.qa_model,
+                    infer_params={
+                        "saas.base_url": model_info["base_url"],
+                        "saas.api_key": model_info["api_key"],
+                        "saas.model": model_info["model_name"],
+                        "saas.is_reasoning": model_info["is_reasoning"]
+                    }
+                )
+                llm.setup_sub_client("qa_model", qa_model)
 
         if server_args.doc_dir:
             auto_coder_args.rag_type = "simple"
@@ -577,16 +695,61 @@ def main(input_args: Optional[List[str]] = None):
             count_tokens(args.tokenizer_path, args.file)
         elif args.tool == "recall":
             from .common.recall_validation import validate_recall
-
-            llm = byzerllm.ByzerLLM.from_default_model(args.model)
+            from autocoder import models as models_module
+
+            # Handle lite/pro flags
+            if args.lite:
+                args.product_mode = "lite"
+            elif args.pro:
+                args.product_mode = "pro"
+
+            if args.product_mode == "pro":
+                llm = byzerllm.ByzerLLM.from_default_model(args.model)
+            else:  # lite mode
+                model_info = models_module.get_model_by_name(args.model)
+                llm = byzerllm.SimpleByzerLLM(default_model_name=args.model)
+                llm.deploy(
+                    model_path="",
+                    pretrained_model_type=model_info["model_type"],
+                    udf_name=args.model,
+                    infer_params={
+                        "saas.base_url": model_info["base_url"],
+                        "saas.api_key": model_info["api_key"],
+                        "saas.model": model_info["model_name"],
+                        "saas.is_reasoning": model_info["is_reasoning"]
+                    }
+                )
 
             content = None if not args.content else [args.content]
             result = validate_recall(llm, content=content, query=args.query)
             print(f"Recall Validation Result:\n{result}")
+
         elif args.tool == "chunk":
             from .common.chunk_validation import validate_chunk
+            from autocoder import models as models_module
+            
+            if args.lite:
+                args.product_mode = "lite"
+            elif args.pro:
+                args.product_mode = "pro"
+
+            if args.product_mode == "pro":
+                llm = byzerllm.ByzerLLM.from_default_model(args.model)
+            else:  # lite mode
+                model_info = models_module.get_model_by_name(args.model)
+                llm = byzerllm.SimpleByzerLLM(default_model_name=args.model)
+                llm.deploy(
+                    model_path="",
+                    pretrained_model_type=model_info["model_type"],
+                    udf_name=args.model,
+                    infer_params={
+                        "saas.base_url": model_info["base_url"],
+                        "saas.api_key": model_info["api_key"],
+                        "saas.model": model_info["model_name"],
+                        "saas.is_reasoning": model_info["is_reasoning"]
+                    }
+                )
 
-            llm = byzerllm.ByzerLLM.from_default_model(args.model)
             content = None if not args.content else [args.content]
             result = validate_chunk(llm, content=content, query=args.query)
             print(f"Chunk Model Validation Result:\n{result}")

{auto_coder-0.1.251 → auto_coder-0.1.253}/src/autocoder/chat_auto_coder.py

@@ -49,12 +49,12 @@ from autocoder.common.mcp_server import get_mcp_server, McpRequest, McpInstallRe
 import byzerllm
 from byzerllm.utils import format_str_jinja2
 from autocoder.common.memory_manager import get_global_memory_file_paths 
-from autocoder import models
+from autocoder import models as models_module
 import shlex
 from autocoder.utils.llms import get_single_llm
 import pkg_resources
 from autocoder.common.printer import Printer
-from byzerllm.utils.langutil import run_in_thread,run_in_raw_thread
+from autocoder.utils.thread_utils import run_in_thread,run_in_raw_thread
 
 class SymbolItem(BaseModel):
     symbol_name: str
@@ -136,6 +136,8 @@ commands = [
 
 
 def show_help():
+    print(f"\033[1m{get_message('official_doc')}\033[0m")
+    print()
     print(f"\033[1m{get_message('supported_commands')}\033[0m")
     print()
     print(
@@ -1537,11 +1539,12 @@ def mcp(query: str):
         if os.path.exists(temp_yaml):
             os.remove(temp_yaml)
 
-    mcp_server = get_mcp_server()
+    mcp_server = get_mcp_server()    
     response = mcp_server.send_request(
         McpRequest(
             query=query,
-            model=args.inference_model or args.model
+            model=args.inference_model or args.model,
+            product_mode=args.product_mode
         )
     )
 
@@ -1562,19 +1565,19 @@ def mcp(query: str):
         file_path = os.path.join(mcp_dir, f"{timestamp}.md")
         
         # Format response as markdown
-        markdown_content = f"# {printer.get_message_from_key('mcp_response_title')}\n\n{response.result}"
+        markdown_content = response.result
         
         # Save to file
         with open(file_path, "w", encoding="utf-8") as f:
             f.write(markdown_content)
             
-        # Print with markdown formatting
-        printer.print_panel(
-            Markdown(markdown_content),
-            text_options={"justify": "left"},
-            panel_options={
-                "border_style": "green"
-            }
+        console = Console()
+        console.print(
+            Panel(
+                Markdown(markdown_content, justify="left"),
+                title=printer.get_message_from_key('mcp_response_title'),
+                border_style="green"
+            )
         )
 
 
@@ -1713,7 +1716,7 @@ def commit(query: str):
                 if os.path.exists(temp_yaml):
                     os.remove(temp_yaml)
             
-            target_model = args.code_model or args.model
+            target_model = args.commit_model or args.model
             llm = get_single_llm(target_model, product_mode)
             printer = Printer()
             printer.print_in_terminal("commit_generating", style="yellow", model_name=target_model)
@@ -1739,7 +1742,7 @@ def commit(query: str):
             md5 = hashlib.md5(file_content.encode("utf-8")).hexdigest()
             file_name = os.path.basename(execute_file)
             commit_result = git_utils.commit_changes(
-                ".", f"auto_coder_{file_name}_{md5}"
+                ".", f"auto_coder_{file_name}_{md5}\n{commit_message}"
             )
             git_utils.print_commit_info(commit_result=commit_result)
             if commit_message:
@@ -2171,7 +2174,7 @@ def manage_models(params, query: str):
         printer.print_in_terminal("models_lite_only", style="red")
         return
         
-    models_data = models.load_models()
+    models_data = models_module.load_models()
     subcmd = ""
     if "/list" in query:
         subcmd = "/list"
@@ -2208,6 +2211,9 @@ def manage_models(params, query: str):
             table.add_column("Name", style="cyan", width=40, no_wrap=False)
             table.add_column("Model Name", style="magenta", width=30, overflow="fold")
             table.add_column("Base URL", style="white", width=50, overflow="fold")
+            table.add_column("Input Price (M)", style="magenta", width=15)
+            table.add_column("Output Price (M)", style="magenta", width=15)
+            table.add_column("Speed (s/req)", style="blue", width=15)
             for m in models_data:
                 # Check if api_key_path exists and file exists
                 is_api_key_set = "api_key" in m  
@@ -2221,19 +2227,67 @@ def manage_models(params, query: str):
                 table.add_row(
                     name,                    
                     m.get("model_name", ""),                    
-                    m.get("base_url", "")
+                    m.get("base_url", ""),
+                    f"{m.get('input_price', 0.0):.2f}",
+                    f"{m.get('output_price', 0.0):.2f}",
+                    f"{m.get('average_speed', 0.0):.3f}"
                 )
             console.print(table)
         else:
             printer.print_in_terminal("models_no_models", style="yellow")
 
+    elif subcmd == "/input_price":
+        args = query.strip().split()
+        if len(args) >= 2:
+            name = args[0]
+            try:
+                price = float(args[1])
+                if models_module.update_model_input_price(name, price):
+                    printer.print_in_terminal("models_input_price_updated", style="green", name=name, price=price)
+                else:
+                    printer.print_in_terminal("models_not_found", style="red", name=name)
+            except ValueError as e:
+                printer.print_in_terminal("models_invalid_price", style="red", error=str(e))
+        else:
+            printer.print_in_terminal("models_input_price_usage", style="red")
+            
+    elif subcmd == "/output_price":
+        args = query.strip().split()
+        if len(args) >= 2:
+            name = args[0]
+            try:
+                price = float(args[1])
+                if models_module.update_model_output_price(name, price):
+                    printer.print_in_terminal("models_output_price_updated", style="green", name=name, price=price)
+                else:
+                    printer.print_in_terminal("models_not_found", style="red", name=name)
+            except ValueError as e:
+                printer.print_in_terminal("models_invalid_price", style="red", error=str(e))
+        else:
+            printer.print_in_terminal("models_output_price_usage", style="red")
+
+    elif subcmd == "/speed":
+        args = query.strip().split()
+        if len(args) >= 2:
+            name = args[0]
+            try:
+                speed = float(args[1])
+                if models_module.update_model_speed(name, speed):
+                    printer.print_in_terminal("models_speed_updated", style="green", name=name, speed=speed)
+                else:
+                    printer.print_in_terminal("models_not_found", style="red", name=name)
+            except ValueError as e:
+                printer.print_in_terminal("models_invalid_speed", style="red", error=str(e))
+        else:
+            printer.print_in_terminal("models_speed_usage", style="red")
+            
     elif subcmd == "/add":
         # Support both simplified and legacy formats
         args = query.strip().split(" ")        
         if len(args) == 2:
             # Simplified: /models /add <name> <api_key>
             name, api_key = args[0], args[1]            
-            result = models.update_model_with_api_key(name, api_key)
+            result = models_module.update_model_with_api_key(name, api_key)
             if result:
                 printer.print_in_terminal("models_added", style="green", name=name)
             else:
@@ -2275,7 +2329,7 @@ def manage_models(params, query: str):
         }
 
         models_data.append(final_model)
-        models.save_models(models_data)
+        models_module.save_models(models_data)
         printer.print_in_terminal("models_add_model_success", style="green", name=data_dict["name"])
 
     elif subcmd == "/remove":
@@ -2288,7 +2342,7 @@ def manage_models(params, query: str):
         if len(filtered_models) == len(models_data):
             printer.print_in_terminal("models_add_model_remove", style="yellow", name=name)
             return
-        models.save_models(filtered_models)
+        models_module.save_models(filtered_models)
         printer.print_in_terminal("models_add_model_removed", style="green", name=name)
 
     else: