auto-coder 0.1.251__py3-none-any.whl → 0.1.252__py3-none-any.whl

autocoder/chat_auto_coder_lang.py

@@ -85,7 +85,7 @@ MESSAGES = {
         "design_desc": "Generate SVG image based on the provided description",
         "commit_desc": "Auto generate yaml file and commit changes based on user's manual changes",
         "models_desc": "Manage model configurations, only available in lite mode",
-        "models_usage": "Usage: /models /list|/add|/add_model|/remove ...",
+        "models_usage": "Usage: /models /list|/add|/add_model|/remove|/price|/speed ...",
         "models_added": "Added/Updated model '{{name}}' successfully.",
         "models_add_failed": "Failed to add model '{{name}}'. Model not found in defaults.",
         "models_add_usage": "Usage: /models /add <name> <api_key> or\n/models /add <name> <model_type> <model_name> <base_url> <api_key_path> [description]",
@@ -96,6 +96,14 @@ MESSAGES = {
         "models_add_model_remove": "Model '{{name}}' not found.",
         "models_add_model_removed": "Removed model: {{name}}",
         "models_unknown_subcmd": "Unknown subcommand: {{subcmd}}",
+        "models_input_price_updated": "Updated input price for model {{name}} to {{price}} M/token",
+        "models_output_price_updated": "Updated output price for model {{name}} to {{price}} M/token",
+        "models_invalid_price": "Invalid price value: {{error}}",
+        "models_input_price_usage": "Usage: /models /input_price <name> <value>",
+        "models_output_price_usage": "Usage: /models /output_price <name> <value>",
+        "models_speed_updated": "Updated speed for model {{name}} to {{speed}} s/request",
+        "models_invalid_speed": "Invalid speed value: {{error}}",
+        "models_speed_usage": "Usage: /models /speed <name> <value>",
         "models_title": "All Models (内置 + models.json)",
         "models_no_models": "No models found.",
         "models_lite_only": "The /models command is only available in lite mode",
@@ -117,6 +125,7 @@ MESSAGES = {
         "commit_message": "{{ model_name }} Generated commit message: {{ message }}",
         "commit_failed": "{{ model_name }} Failed to generate commit message: {{ error }}",
         "confirm_execute": "Do you want to execute this script?",
+        "official_doc": "Official Documentation: https://uelng8wukz.feishu.cn/wiki/NhPNwSRcWimKFIkQINIckloBncI",
     },
     "zh": {
         "commit_generating": "{{ model_name }} 正在生成提交信息...", 
@@ -204,7 +213,7 @@ MESSAGES = {
         "conf_value": "值",
         "conf_title": "配置设置",
         "conf_subtitle": "使用 /conf <key>:<value> 修改这些设置",
-        "models_usage": "用法: /models /list|/add|/add_model|/remove ...",
+        "models_usage": "用法: /models /list|/add|/add_model|/remove|/price|/speed ...",
         "models_added": "成功添加/更新模型 '{{name}}'。",
         "models_add_failed": "添加模型 '{{name}}' 失败。在默认模型中未找到该模型。",
         "models_add_usage": "用法: /models /add <name> <api_key> 或\n/models /add <name> <model_type> <model_name> <base_url> <api_key_path> [description]",
@@ -215,6 +224,14 @@ MESSAGES = {
         "models_add_model_remove": "找不到模型 '{{name}}'。",
         "models_add_model_removed": "已移除模型: {{name}}",
         "models_unknown_subcmd": "未知的子命令: {{subcmd}}",
+        "models_input_price_updated": "已更新模型 {{name}} 的输入价格为 {{price}} M/token",
+        "models_output_price_updated": "已更新模型 {{name}} 的输出价格为 {{price}} M/token",
+        "models_invalid_price": "无效的价格值: {{error}}",
+        "models_input_price_usage": "用法: /models /input_price <name> <value>",
+        "models_output_price_usage": "用法: /models /output_price <name> <value>",
+        "models_speed_updated": "已更新模型 {{name}} 的速度为 {{speed}} 秒/请求",
+        "models_invalid_speed": "无效的速度值: {{error}}",
+        "models_speed_usage": "用法: /models /speed <name> <value>",
         "models_title": "所有模型 (内置 + models.json)",
         "models_no_models": "未找到任何模型。",
         "models_lite_only": "/models 命令仅在 lite 模式下可用",
@@ -232,7 +249,8 @@ MESSAGES = {
         "remove_files_none": "没有文件被移除。",
         "files_removed": "移除的文件",
         "models_api_key_empty": "警告:  {{name}}  API key 为空。请设置一个有效的 API key。",
-        "confirm_execute": "是否执行此脚本？",
+        "confirm_execute": "是否执行此脚本?",
+        "official_doc": "官方文档: https://uelng8wukz.feishu.cn/wiki/NhPNwSRcWimKFIkQINIckloBncI",
     }
 }
 

autocoder/common/__init__.py

@@ -254,6 +254,7 @@ class AutoCoderArgs(pydantic.BaseModel):
     planner_model: Optional[str] = ""
     voice2text_model: Optional[str] = ""
     text2voice_model: Optional[str] = ""
+    commit_model: Optional[str] = ""
     
     skip_build_index: Optional[bool] = False
     skip_filter_index: Optional[bool] = False

autocoder/common/auto_coder_lang.py

@@ -3,6 +3,7 @@ from byzerllm.utils import format_str_jinja2
 
 MESSAGES = {
     "en": {
+        "generation_cancelled": "[Interrupted] Generation cancelled",
         "model_not_found": "Model {{model_name}} not found",
         "generating_shell_script": "Generating Shell Script",
         "new_session_started": "New session started. Previous chat history has been archived.",    
@@ -51,7 +52,7 @@ MESSAGES = {
             "Paste the answer to the input box below, use '/break' to exit, '/clear' to clear the screen, '/eof' to submit."
         ),
         "code_generation_start": "Auto generate the code...",
-        "code_generation_complete": "Code generation completed in {{ duration }} seconds, input_tokens_count: {{ input_tokens }}, generated_tokens_count: {{ output_tokens }}",
+        "code_generation_complete": "Code generation completed in {{ duration }} seconds, input_tokens_count: {{ input_tokens }}, generated_tokens_count: {{ output_tokens }}, speed: {{ speed }} tokens/s",
         "code_merge_start": "Auto merge the code...",
         "code_execution_warning": "Content(send to model) is {{ content_length }} tokens (you may collect too much files), which is larger than the maximum input length {{ max_length }}",
         "quick_filter_start": "{{ model_name }} Starting filter context(quick_filter)...",
@@ -73,7 +74,7 @@ MESSAGES = {
         "ranking_process_failed": "Ranking process failed: {{ error }}",
         "ranking_failed": "Ranking failed in {{ elapsed }}s, using original order",
         "begin_index_source_code": "🚀 Begin to index source code in {{ source_dir }}",
-        "stream_out_stats": "Elapsed time {{ elapsed_time }} seconds, input tokens: {{ input_tokens }}, output tokens: {{ output_tokens }}",
+        "stream_out_stats": "Elapsed time {{ elapsed_time }} seconds, first token time: {{ first_token_time }} seconds, input tokens: {{ input_tokens }}, output tokens: {{ output_tokens }}, speed: {{ speed }} tokens/s",
         "quick_filter_stats": "快速过滤器完成，耗时 {{ elapsed_time }} 秒，输入token数: {{ input_tokens }}, 输出token数: {{ output_tokens }}",
         "upsert_file": "✅ Updated file: {{ file_path }}",  
         "unmerged_blocks_title": "Unmerged Blocks",
@@ -91,6 +92,7 @@ MESSAGES = {
         "estimated_input_tokens_in_generate": "Estimated input tokens in generate ({{ generate_mode }}): {{ estimated_input_tokens }}",
     },
     "zh": {
+        "generation_cancelled": "[已中断] 生成已取消",
         "model_not_found": "未找到模型: {{model_name}}",
         "generating_shell_script": "正在生成 Shell 脚本",
         "new_session_started": "新会话已开始。之前的聊天历史已存档。",
@@ -139,7 +141,7 @@ MESSAGES = {
             "将获得答案黏贴到下面的输入框，换行后，使用 '/break' 退出，'/clear' 清屏，'/eof' 提交。"
         ),
         "code_generation_start": "正在自动生成代码...",
-        "code_generation_complete": "代码生成完成，耗时 {{ duration }} 秒，输入token数: {{ input_tokens }}, 输出token数: {{ output_tokens }}",
+        "code_generation_complete": "代码生成完成，耗时 {{ duration }} 秒，输入token数: {{ input_tokens }}, 输出token数: {{ output_tokens }}, 速度: {{ speed }} tokens/秒",
         "code_merge_start": "正在自动合并代码...",
         "code_execution_warning": "发送给模型的内容长度为 {{ content_length }} tokens（您可能收集了太多文件），超过了最大输入长度 {{ max_length }}",
         "quick_filter_start": "{{ model_name }} 开始查找上下文(quick_filter)...",
@@ -171,7 +173,7 @@ MESSAGES = {
         "ranking_complete": "排序完成，耗时 {{ elapsed }} 秒，总投票数: {{ total_tasks }}，最佳候选索引: {{ best_candidate }}，得分: {{ scores }}，输入token数: {{ input_tokens }}，输出token数: {{ output_tokens }}",
         "ranking_process_failed": "排序过程失败: {{ error }}",
         "ranking_failed": "排序失败，耗时 {{ elapsed }} 秒，使用原始顺序",
-        "stream_out_stats": "耗时 {{ elapsed_time }} 秒，输入token数: {{ input_tokens }}, 输出token数: {{ output_tokens }}",
+        "stream_out_stats": "总耗时 {{ elapsed_time }} 秒，首token时间: {{ first_token_time }} 秒，输入token数: {{ input_tokens }}, 输出token数: {{ output_tokens }}, 速度: {{ speed }} tokens/秒",
         "quick_filter_stats": "Quick filter completed in {{ elapsed_time }} seconds, input tokens: {{ input_tokens }}, output tokens: {{ output_tokens }}",
         "quick_filter_title": "{{ model_name }} 正在分析如何筛选上下文...",
         "quick_filter_failed": "❌ 快速过滤器失败: {{ error }}. ",

autocoder/common/code_modification_ranker.py

@@ -8,6 +8,8 @@ from concurrent.futures import ThreadPoolExecutor, as_completed
 import traceback
 from autocoder.common.utils_code_auto_generate import chat_with_continue
 from byzerllm.utils.str2model import to_model
+
+from autocoder.utils.llms import get_llm_names
 class RankResult(BaseModel):
     rank_result: List[int]
 
@@ -78,9 +80,7 @@ class CodeModificationRanker:
                 # Submit tasks for each model and generate_times
                 futures = []
                 for llm in self.llms:                    
-                    model_name = getattr(llm, 'default_model_name', None)                    
-                    if not model_name:
-                        model_name = "unknown(without default model name)"
+                    model_name = ",".join(get_llm_names(llm))
                     self.printer.print_in_terminal(
                         "ranking_start", style="blue", count=len(generate_result.contents), model_name=model_name)
                     

autocoder/common/global_cancel.py

@@ -0,0 +1,21 @@
+import threading
+
+class GlobalCancel:
+    def __init__(self):
+        self._flag = False
+        self._lock = threading.Lock()
+    
+    @property
+    def requested(self):
+        with self._lock:
+            return self._flag
+    
+    def set(self):
+        with self._lock:
+            self._flag = True
+    
+    def reset(self):
+        with self._lock:
+            self._flag = False
+
+global_cancel = GlobalCancel()

autocoder/dispacher/actions/action.py

@@ -26,6 +26,7 @@ from autocoder.utils.conversation_store import store_code_model_conversation
 from loguru import logger
 import time
 from autocoder.common.printer import Printer
+from autocoder.utils.llms import get_llm_names
 
 
 class BaseAction:    
@@ -123,11 +124,16 @@ class ActionTSProject(BaseAction):
                 generate_result = generate.single_round_run(
                     query=args.query, source_content=content
                 )
+            elapsed_time = time.time() - start_time
+            speed = generate_result.metadata.get('generated_tokens_count', 0) / elapsed_time if elapsed_time > 0 else 0            
+            model_names = ",".join(get_llm_names(self.llm))
             self.printer.print_in_terminal(
                 "code_generation_complete",
-                duration=time.time() - start_time,
+                duration=elapsed_time,
                 input_tokens=generate_result.metadata.get('input_tokens_count', 0),
-                output_tokens=generate_result.metadata.get('generated_tokens_count', 0)
+                output_tokens=generate_result.metadata.get('generated_tokens_count', 0),
+                speed=round(speed, 2),
+                model_names=model_names
             )
             merge_result = None
             if args.execute and args.auto_merge:
@@ -213,11 +219,16 @@ class ActionPyScriptProject(BaseAction):
                     query=args.query, source_content=content
                 )
 
+            elapsed_time = time.time() - start_time
+            speed = generate_result.metadata.get('generated_tokens_count', 0) / elapsed_time if elapsed_time > 0 else 0
+            model_names = ",".join(get_llm_names(self.llm))
             self.printer.print_in_terminal(
                 "code_generation_complete",
-                duration=time.time() - start_time,
+                duration=elapsed_time,
                 input_tokens=generate_result.metadata.get('input_tokens_count', 0),
-                output_tokens=generate_result.metadata.get('generated_tokens_count', 0)
+                output_tokens=generate_result.metadata.get('generated_tokens_count', 0),
+                speed=round(speed, 2),
+                model_names=model_names
             )
             merge_result = None
             if args.execute and args.auto_merge:
@@ -335,11 +346,16 @@ class ActionPyProject(BaseAction):
                 generate_result = generate.single_round_run(
                     query=args.query, source_content=content
                 )
+            elapsed_time = time.time() - start_time
+            speed = generate_result.metadata.get('generated_tokens_count', 0) / elapsed_time if elapsed_time > 0 else 0
+            model_names = ",".join(get_llm_names(self.llm))
             self.printer.print_in_terminal(
                 "code_generation_complete",
-                duration=time.time() - start_time,
+                duration=elapsed_time,
                 input_tokens=generate_result.metadata.get('input_tokens_count', 0),
-                output_tokens=generate_result.metadata.get('generated_tokens_count', 0)
+                output_tokens=generate_result.metadata.get('generated_tokens_count', 0),
+                speed=round(speed, 2),
+                model_names=model_names
             )
             merge_result = None
             if args.execute and args.auto_merge:
@@ -440,11 +456,16 @@ class ActionSuffixProject(BaseAction):
                     query=args.query, source_content=content
                 )
               
+        elapsed_time = time.time() - start_time
+        speed = generate_result.metadata.get('generated_tokens_count', 0) / elapsed_time if elapsed_time > 0 else 0
+        model_names = ",".join(get_llm_names(self.llm))
         self.printer.print_in_terminal(
             "code_generation_complete",
-            duration=time.time() - start_time,
+            duration=elapsed_time,
             input_tokens=generate_result.metadata.get('input_tokens_count', 0),
-            output_tokens=generate_result.metadata.get('generated_tokens_count', 0)
+            output_tokens=generate_result.metadata.get('generated_tokens_count', 0),
+            speed=round(speed, 2),
+            model_names=model_names
         )
         merge_result = None
         if args.execute and args.auto_merge:

autocoder/dispacher/actions/plugins/action_regex_project.py

@@ -12,9 +12,10 @@ from autocoder.common.code_auto_generate_editblock import CodeAutoGenerateEditBl
 from autocoder.index.entry import build_index_and_filter_files
 from autocoder.regexproject import RegexProject
 from autocoder.utils.conversation_store import store_code_model_conversation
-from loguru import logger
+from autocoder.common.printer import Printer
 import time
-
+from autocoder.utils.llms import get_llm_names
+from loguru import logger
 class ActionRegexProject:
     def __init__(
         self, args: AutoCoderArgs, llm: Optional[byzerllm.ByzerLLM] = None
@@ -22,6 +23,7 @@ class ActionRegexProject:
         self.args = args
         self.llm = llm
         self.pp = None
+        self.printer = Printer()
 
     def run(self):
         args = self.args
@@ -58,7 +60,7 @@ class ActionRegexProject:
 
         start_time = time.time()
         if args.execute:
-            logger.info("Auto generate the code...")
+            self.printer.print_in_terminal("code_generation_start")
             
             if args.auto_merge == "diff":
                 generate = CodeAutoGenerateDiff(
@@ -83,10 +85,20 @@ class ActionRegexProject:
                     query=args.query, source_content=content
                 )
 
-            logger.info(f"Code generation completed in {time.time() - start_time:.2f} seconds, input_tokens_count: {generate_result.metadata.get('input_tokens_count', 0)}, generated_tokens_count: {generate_result.metadata.get('generated_tokens_count', 0)}")    
+            elapsed_time = time.time() - start_time
+            speed = generate_result.metadata.get('generated_tokens_count', 0) / elapsed_time if elapsed_time > 0 else 0
+            model_names = ",".join(get_llm_names(self.llm))
+            self.printer.print_in_terminal(
+                "code_generation_complete",
+                duration=elapsed_time,
+                input_tokens=generate_result.metadata.get('input_tokens_count', 0),
+                output_tokens=generate_result.metadata.get('generated_tokens_count', 0),
+                speed=round(speed, 2),
+                model_names=model_names
+            )
             merge_result = None
             if args.execute and args.auto_merge:
-                logger.info("Auto merge the code...")
+                self.printer.print_in_terminal("code_merge_start")
                 if args.auto_merge == "diff":
                     code_merge = CodeAutoMergeDiff(llm=self.llm, args=self.args)
                     merge_result = code_merge.merge_code(generate_result=generate_result)

autocoder/index/filter/quick_filter.py

@@ -17,6 +17,8 @@ from autocoder.common.printer import Printer
 from concurrent.futures import ThreadPoolExecutor
 import threading
 
+from autocoder.utils.llms import get_llm_names
+
 
 def get_file_path(file_path):
     if file_path.startswith("##"):
@@ -70,9 +72,7 @@ class QuickFilter():
 
         def process_chunk(chunk_index: int, chunk: List[IndexItem]) -> None:
             try:
-                model_name = getattr(self.index_manager.index_filter_llm, 'default_model_name', None)
-                if not model_name:
-                    model_name = "unknown(without default model name)"
+                model_name = ",".join(get_llm_names(self.index_manager.index_filter_llm))
                 
                 if chunk_index == 0:
                     # 第一个chunk使用流式输出
@@ -180,9 +180,7 @@ class QuickFilter():
             return self.big_filter(index_items)
         
         try:
-            model_name = getattr(self.index_manager.index_filter_llm, 'default_model_name', None)
-            if not model_name:
-                model_name = "unknown(without default model name)"
+            model_name = ",".join(get_llm_names(self.index_manager.index_filter_llm))
             
             # 渲染 Prompt 模板
             query = self.quick_filter_files.prompt(index_items, self.args.query)

autocoder/index/index.py

@@ -22,7 +22,8 @@ from autocoder.index.types import (
     TargetFile,
     FileList,
 )
-
+from autocoder.common.global_cancel import global_cancel
+from autocoder.utils.llms import get_llm_names
 class IndexManager:
     def __init__(
         self, llm: byzerllm.ByzerLLM, sources: List[SourceCode], args: AutoCoderArgs
@@ -195,7 +196,10 @@ class IndexManager:
             return True
         return False
 
-    def build_index_for_single_source(self, source: SourceCode):        
+    def build_index_for_single_source(self, source: SourceCode):   
+        if global_cancel.requested:
+            return None
+
         file_path = source.module_name
         if not os.path.exists(file_path):
             return None
@@ -205,9 +209,7 @@ class IndexManager:
 
         md5 = hashlib.md5(source.source_code.encode("utf-8")).hexdigest()
 
-        model_name = getattr(self.index_llm, 'default_model_name', None)
-        if not model_name:
-            model_name = "unknown(without default model name)"
+        model_name = ",".join(get_llm_names(self.index_llm))
 
         try:
             start_time = time.monotonic()
@@ -314,6 +316,9 @@ class IndexManager:
                 ):
                     wait_to_build_files.append(source)
 
+            # Remove duplicates based on module_name
+            wait_to_build_files = list({source.module_name: source for source in wait_to_build_files}.values())
+
             counter = 0
             num_files = len(wait_to_build_files)
             total_files = len(self.sources)
@@ -329,6 +334,8 @@ class IndexManager:
                 for source in wait_to_build_files
             ]
             for future in as_completed(futures):
+                if global_cancel.requested:
+                    break
                 result = future.result()
                 if result is not None:
                     counter += 1
@@ -345,7 +352,7 @@ class IndexManager:
                         with open(self.index_file, "w") as file:
                             json.dump(index_data, file, ensure_ascii=False, indent=2)
                         updated_sources = []
-                        
+        
         # 如果 updated_sources 或 keys_to_remove 有值，则保存索引文件
         if updated_sources or keys_to_remove:
             with open(self.index_file, "w") as file:

autocoder/models.py

@@ -2,7 +2,6 @@ import os
 import json
 from typing import List, Dict
 from urllib.parse import urlparse
-from autocoder.common.auto_coder_lang import get_message_with_format
 
 MODELS_JSON = os.path.expanduser("~/.auto-coder/keys/models.json")
 
@@ -15,7 +14,10 @@ default_models_list = [
         "model_type": "saas/openai",
         "base_url": "https://api.deepseek.com/v1",
         "api_key_path": "api.deepseek.com",
-        "is_reasoning": True
+        "is_reasoning": True,
+        "input_price": 0.0,  # 单位:M/百万 input tokens
+        "output_price": 0.0,  # 单位:M/百万 output tokens
+        "average_speed": 0.0  # 单位:秒/请求
     },    
     {
         "name": "deepseek_chat",
@@ -24,7 +26,10 @@ default_models_list = [
         "model_type": "saas/openai",
         "base_url": "https://api.deepseek.com/v1",
         "api_key_path": "api.deepseek.com",
-        "is_reasoning": False
+        "is_reasoning": False,
+        "input_price": 0.0,
+        "output_price": 0.0,
+        "average_speed": 0.0
     },
     {
         "name":"o1",
@@ -33,7 +38,10 @@ default_models_list = [
         "model_type": "saas/openai",
         "base_url": "https://api.openai.com/v1",
         "api_key_path": "",
-        "is_reasoning": True
+        "is_reasoning": True,
+        "input_price": 0.0,
+        "output_price": 0.0,
+        "average_speed": 0.0
     }
 ]
 
@@ -106,6 +114,7 @@ def get_model_by_name(name: str) -> Dict:
     """
     根据模型名称查找模型
     """
+    from autocoder.common.auto_coder_lang import get_message_with_format
     models = load_models()
     v = [m for m in models if m["name"] == name.strip()]
     
@@ -114,6 +123,78 @@ def get_model_by_name(name: str) -> Dict:
     return v[0]
 
 
+def update_model_input_price(name: str, price: float) -> bool:
+    """更新模型输入价格
+    
+    Args:
+        name: 模型名称
+        price: 输入价格(M/百万input tokens)
+        
+    Returns:
+        bool: 是否更新成功
+    """
+    if price < 0:
+        raise ValueError("Price cannot be negative")
+        
+    models = load_models()
+    updated = False
+    for model in models:
+        if model["name"] == name:
+            model["input_price"] = float(price)
+            updated = True
+            break
+    if updated:
+        save_models(models)
+    return updated
+
+def update_model_output_price(name: str, price: float) -> bool:
+    """更新模型输出价格
+    
+    Args:
+        name: 模型名称
+        price: 输出价格(M/百万output tokens)
+        
+    Returns:
+        bool: 是否更新成功
+    """
+    if price < 0:
+        raise ValueError("Price cannot be negative")
+        
+    models = load_models()
+    updated = False
+    for model in models:
+        if model["name"] == name:
+            model["output_price"] = float(price)
+            updated = True
+            break
+    if updated:
+        save_models(models)
+    return updated
+
+def update_model_speed(name: str, speed: float) -> bool:
+    """更新模型平均速度
+    
+    Args:
+        name: 模型名称
+        speed: 速度(秒/请求)
+        
+    Returns:
+        bool: 是否更新成功
+    """
+    if speed <= 0:
+        raise ValueError("Speed must be positive")
+        
+    models = load_models()
+    updated = False
+    for model in models:
+        if model["name"] == name:
+            model["average_speed"] = float(speed)
+            updated = True
+            break
+    if updated:
+        save_models(models)
+    return updated
+
 def check_model_exists(name: str) -> bool:
     """
     检查模型是否存在
@@ -124,14 +205,14 @@ def check_model_exists(name: str) -> bool:
 def update_model_with_api_key(name: str, api_key: str) -> Dict:
     """
     根据模型名称查找并更新模型的 api_key_path。
-    如果找到模型，会根据其 base_url 处理 api_key_path。
+    如果找到模型,会根据其 base_url 处理 api_key_path。
     
     Args:
         name: 模型名称
         api_key: API密钥
         
     Returns:
-        Dict: 更新后的模型信息，如果未找到则返回None
+        Dict: 更新后的模型信息,如果未找到则返回None
     """
     models = load_models()
     

autocoder/rag/doc_filter.py

@@ -91,9 +91,7 @@ class DocFilter:
                 def _run(conversations, docs):
                     submit_time_1 = time.time()
                     try:
-                        llm = ByzerLLM()
-                        llm.skip_nontext_check = True
-                        llm.setup_default_model_name(self.recall_llm.default_model_name)
+                        llm = self.recall_llm
 
                         v = (
                             _check_relevance_with_conversation.with_llm(

autocoder/rag/long_context_rag.py

@@ -52,11 +52,13 @@ class LongContextRAG:
     ) -> None:
         self.llm = llm
         self.args = args
-
-        self.index_model = byzerllm.ByzerLLM()
-        self.index_model.setup_default_model_name(
-            args.index_model or self.llm.default_model_name
-        )
+        if args.product_mode == "pro":
+            self.index_model = byzerllm.ByzerLLM()
+            self.index_model.setup_default_model_name(
+                args.index_model or self.llm.default_model_name
+            )
+        else:
+            self.index_model = self.llm
 
         self.path = path
         self.relevant_score = self.args.rag_doc_filter_relevance or 5

autocoder/rag/token_limiter.py

@@ -224,9 +224,7 @@ class TokenLimiter:
                 for idx, line in enumerate(source_code_lines):
                     source_code_with_line_number += f"{idx+1} {line}\n"
                 
-                llm = ByzerLLM()
-                llm.skip_nontext_check = True
-                llm.setup_default_model_name(self.chunk_llm.default_model_name)
+                llm = self.chunk_llm
 
                 extracted_info = (
                     self.extract_relevance_range_from_docs_with_conversation.options(

autocoder/utils/auto_coder_utils/chat_stream_out.py

@@ -1,4 +1,5 @@
 from rich.console import Console
+from autocoder.common.printer import Printer
 from rich.live import Live
 from rich.panel import Panel
 from rich.markdown import Markdown
@@ -11,6 +12,7 @@ from autocoder.utils.request_queue import request_queue
 import time
 from byzerllm.utils.types import SingleOutputMeta
 from autocoder.common import AutoCoderArgs
+from autocoder.common.global_cancel import global_cancel
 
 MAX_HISTORY_LINES = 40  # 最大保留历史行数
 
@@ -172,7 +174,9 @@ def stream_out(
     current_line = ""  # 当前行
     assistant_response = ""
     last_meta = None
-    panel_title = title if title is not None else f"Response[ {model_name} ]"    
+    panel_title = title if title is not None else f"Response[ {model_name} ]"  
+    first_token_time = 0.0
+    first_token_time_start = time.time()
     try:        
         with Live(
             Panel("", title=panel_title, border_style="green"),
@@ -180,6 +184,10 @@ def stream_out(
             console=console
         ) as live:
             for res in stream_generator:
+                if global_cancel.requested:
+                    printer = Printer(console)
+                    printer.print_in_terminal("generation_cancelled")                    
+                    break
                 last_meta = res[1]                
                 content = res[0]
                 reasoning_content = last_meta.reasoning_content
@@ -187,6 +195,9 @@ def stream_out(
                 if reasoning_content == "" and content == "":
                     continue
                 
+                if first_token_time == 0.0:
+                    first_token_time = time.time() - first_token_time_start
+
                 if keep_reasoning_content:
                     # 处理思考内容
                     if reasoning_content:
@@ -280,5 +291,5 @@ def stream_out(
                     status=RequestOption.COMPLETED
                 ),
             )
-            
+    last_meta.first_token_time = first_token_time
     return assistant_response, last_meta

autocoder/utils/llms.py

@@ -1,7 +1,21 @@
 import byzerllm
-from autocoder.auto_coder import models_module
+from typing import Union,Optional
+
+def get_llm_names(llm: Union[byzerllm.ByzerLLM, byzerllm.SimpleByzerLLM,str],target_model_type:Optional[str]=None):
+   if target_model_type is None:
+      return [llm.default_model_name for llm in [llm] if llm.default_model_name]
+   llms = llm.get_sub_client(target_model_type) 
+   if llms is None:
+      return [llm.default_model_name for llm in [llm] if llm.default_model_name]
+   elif isinstance(llms, list):
+       return [llm.default_model_name for llm in llms if llm.default_model_name]
+   elif isinstance(llms,str) and llms:
+      return llms.split(",")      
+   else:
+      return [llm.default_model_name for llm in [llms] if llm.default_model_name]
 
 def get_single_llm(model_names: str, product_mode: str):
+    from autocoder import models as models_module
     if product_mode == "pro":
         if "," in model_names:
             # Multiple code models specified