jarvis-ai-assistant 0.1.193__py3-none-any.whl → 0.1.195__py3-none-any.whl

jarvis/jarvis_utils/config.py

@@ -1,16 +1,12 @@
 # -*- coding: utf-8 -*-
 import os
 from functools import lru_cache
-
 from typing import Any, Dict, List
 
-
 import yaml
 
-
 from jarvis.jarvis_utils.builtin_replace_map import BUILTIN_REPLACE_MAP
 
-
 # 全局环境变量存储
 
 GLOBAL_CONFIG_DATA: Dict[str, Any] = {}
@@ -21,6 +17,7 @@ def set_global_env_data(env_data: Dict[str, Any]) -> None:
     global GLOBAL_CONFIG_DATA
     GLOBAL_CONFIG_DATA = env_data
 
+
 def set_config(key: str, value: Any) -> None:
     """设置配置"""
     GLOBAL_CONFIG_DATA[key] = value
@@ -32,48 +29,53 @@ def set_config(key: str, value: Any) -> None:
 所有配置都从环境变量中读取，带有回退默认值。
 """
 
+
 def get_git_commit_prompt() -> str:
     """
     获取Git提交提示模板
-    
+
     返回:
         str: Git提交信息生成提示模板，如果未配置则返回空字符串
     """
     return GLOBAL_CONFIG_DATA.get("JARVIS_GIT_COMMIT_PROMPT", "")
 
+
 # 输出窗口预留大小
 INPUT_WINDOW_REVERSE_SIZE = 2048
 
+
 @lru_cache(maxsize=None)
 def get_replace_map() -> dict:
     """
     获取替换映射表。
-    
+
     优先使用GLOBAL_CONFIG_DATA['JARVIS_REPLACE_MAP']的配置，
     如果没有则从数据目录下的replace_map.yaml文件中读取替换映射表，
     如果文件不存在则返回内置替换映射表。
-    
+
     返回:
         dict: 合并后的替换映射表字典(内置+文件中的映射表)
     """
-    if 'JARVIS_REPLACE_MAP' in GLOBAL_CONFIG_DATA:
-        return {**BUILTIN_REPLACE_MAP, **GLOBAL_CONFIG_DATA['JARVIS_REPLACE_MAP']}
-        
-    replace_map_path = os.path.join(get_data_dir(), 'replace_map.yaml')
+    if "JARVIS_REPLACE_MAP" in GLOBAL_CONFIG_DATA:
+        return {**BUILTIN_REPLACE_MAP, **GLOBAL_CONFIG_DATA["JARVIS_REPLACE_MAP"]}
+
+    replace_map_path = os.path.join(get_data_dir(), "replace_map.yaml")
     if not os.path.exists(replace_map_path):
         return BUILTIN_REPLACE_MAP.copy()
-    
-    from jarvis.jarvis_utils.output import PrettyOutput, OutputType
+
+    from jarvis.jarvis_utils.output import OutputType, PrettyOutput
+
     PrettyOutput.print(
         "警告：使用replace_map.yaml进行配置的方式已被弃用，将在未来版本中移除。"
         "请迁移到使用GLOBAL_CONFIG_DATA中的JARVIS_REPLACE_MAP配置。",
-        output_type=OutputType.WARNING
+        output_type=OutputType.WARNING,
     )
-    
-    with open(replace_map_path, 'r', encoding='utf-8', errors='ignore') as file:
+
+    with open(replace_map_path, "r", encoding="utf-8", errors="ignore") as file:
         file_map = yaml.safe_load(file) or {}
         return {**BUILTIN_REPLACE_MAP, **file_map}
 
+
 def get_max_token_count() -> int:
     """
     获取模型允许的最大token数量。
@@ -81,7 +83,8 @@ def get_max_token_count() -> int:
     返回:
         int: 模型能处理的最大token数量。
     """
-    return int(GLOBAL_CONFIG_DATA.get('JARVIS_MAX_TOKEN_COUNT', '960000'))
+    return int(GLOBAL_CONFIG_DATA.get("JARVIS_MAX_TOKEN_COUNT", "960000"))
+
 
 def get_max_input_token_count() -> int:
     """
@@ -90,7 +93,7 @@ def get_max_input_token_count() -> int:
     返回:
         int: 模型能处理的最大输入token数量。
     """
-    return int(GLOBAL_CONFIG_DATA.get('JARVIS_MAX_INPUT_TOKEN_COUNT', '32000'))
+    return int(GLOBAL_CONFIG_DATA.get("JARVIS_MAX_INPUT_TOKEN_COUNT", "32000"))
 
 
 def is_auto_complete() -> bool:
@@ -100,7 +103,7 @@ def is_auto_complete() -> bool:
     返回：
         bool: 如果启用了自动补全则返回True，默认为False
     """
-    return GLOBAL_CONFIG_DATA.get('JARVIS_AUTO_COMPLETE', False) == True
+    return GLOBAL_CONFIG_DATA.get("JARVIS_AUTO_COMPLETE", False) == True
 
 
 def get_shell_name() -> str:
@@ -110,8 +113,10 @@ def get_shell_name() -> str:
     返回：
         str: Shell名称（例如bash, zsh），默认为bash
     """
-    shell_path = GLOBAL_CONFIG_DATA.get('SHELL', '/bin/bash')
+    shell_path = GLOBAL_CONFIG_DATA.get("SHELL", "/bin/bash")
     return os.path.basename(shell_path)
+
+
 def get_normal_platform_name() -> str:
     """
     获取正常操作的平台名称。
@@ -119,7 +124,9 @@ def get_normal_platform_name() -> str:
     返回：
         str: 平台名称，默认为'yuanbao'
     """
-    return GLOBAL_CONFIG_DATA.get('JARVIS_PLATFORM', 'yuanbao')
+    return GLOBAL_CONFIG_DATA.get("JARVIS_PLATFORM", "yuanbao")
+
+
 def get_normal_model_name() -> str:
     """
     获取正常操作的模型名称。
@@ -127,7 +134,7 @@ def get_normal_model_name() -> str:
     返回：
         str: 模型名称，默认为'deep_seek'
     """
-    return GLOBAL_CONFIG_DATA.get('JARVIS_MODEL', 'deep_seek_v3')
+    return GLOBAL_CONFIG_DATA.get("JARVIS_MODEL", "deep_seek_v3")
 
 
 def get_thinking_platform_name() -> str:
@@ -137,7 +144,11 @@ def get_thinking_platform_name() -> str:
     返回：
         str: 平台名称，默认为'yuanbao'
     """
-    return GLOBAL_CONFIG_DATA.get('JARVIS_THINKING_PLATFORM', GLOBAL_CONFIG_DATA.get('JARVIS_PLATFORM', 'yuanbao'))
+    return GLOBAL_CONFIG_DATA.get(
+        "JARVIS_THINKING_PLATFORM", GLOBAL_CONFIG_DATA.get("JARVIS_PLATFORM", "yuanbao")
+    )
+
+
 def get_thinking_model_name() -> str:
     """
     获取思考操作的模型名称。
@@ -145,7 +156,10 @@ def get_thinking_model_name() -> str:
     返回：
         str: 模型名称，默认为'deep_seek'
     """
-    return GLOBAL_CONFIG_DATA.get('JARVIS_THINKING_MODEL', GLOBAL_CONFIG_DATA.get('JARVIS_MODEL', 'deep_seek'))
+    return GLOBAL_CONFIG_DATA.get(
+        "JARVIS_THINKING_MODEL", GLOBAL_CONFIG_DATA.get("JARVIS_MODEL", "deep_seek")
+    )
+
 
 def is_execute_tool_confirm() -> bool:
     """
@@ -154,7 +168,9 @@ def is_execute_tool_confirm() -> bool:
     返回：
         bool: 如果需要确认则返回True，默认为False
     """
-    return GLOBAL_CONFIG_DATA.get('JARVIS_EXECUTE_TOOL_CONFIRM', False) == True
+    return GLOBAL_CONFIG_DATA.get("JARVIS_EXECUTE_TOOL_CONFIRM", False) == True
+
+
 def is_confirm_before_apply_patch() -> bool:
     """
     检查应用补丁前是否需要确认。
@@ -162,7 +178,8 @@ def is_confirm_before_apply_patch() -> bool:
     返回：
         bool: 如果需要确认则返回True，默认为False
     """
-    return GLOBAL_CONFIG_DATA.get('JARVIS_CONFIRM_BEFORE_APPLY_PATCH', True) == True
+    return GLOBAL_CONFIG_DATA.get("JARVIS_CONFIRM_BEFORE_APPLY_PATCH", True) == True
+
 
 def get_max_tool_call_count() -> int:
     """
@@ -171,22 +188,23 @@ def get_max_tool_call_count() -> int:
     返回：
         int: 最大连续工具调用次数，默认为20
     """
-    return int(GLOBAL_CONFIG_DATA.get('JARVIS_MAX_TOOL_CALL_COUNT', '20'))
+    return int(GLOBAL_CONFIG_DATA.get("JARVIS_MAX_TOOL_CALL_COUNT", "20"))
 
 
 def get_data_dir() -> str:
     """
     获取Jarvis数据存储目录路径。
-    
+
     返回:
         str: 数据目录路径，优先从JARVIS_DATA_PATH环境变量获取，
              如果未设置或为空，则使用~/.jarvis作为默认值
     """
-    data_path = GLOBAL_CONFIG_DATA.get('JARVIS_DATA_PATH', '').strip()
+    data_path = GLOBAL_CONFIG_DATA.get("JARVIS_DATA_PATH", "").strip()
     if not data_path:
-        return os.path.expanduser('~/.jarvis')
+        return os.path.expanduser("~/.jarvis")
     return data_path
 
+
 def get_auto_update() -> bool:
     """
     获取是否自动更新git仓库。
@@ -194,7 +212,8 @@ def get_auto_update() -> bool:
     返回：
         bool: 如果需要自动更新则返回True，默认为True
     """
-    return GLOBAL_CONFIG_DATA.get('JARVIS_AUTO_UPDATE', True) == True
+    return GLOBAL_CONFIG_DATA.get("JARVIS_AUTO_UPDATE", True) == True
+
 
 def get_max_big_content_size() -> int:
     """
@@ -203,7 +222,8 @@ def get_max_big_content_size() -> int:
     返回：
         int: 最大大内容大小
     """
-    return int(GLOBAL_CONFIG_DATA.get('JARVIS_MAX_BIG_CONTENT_SIZE', '160000'))
+    return int(GLOBAL_CONFIG_DATA.get("JARVIS_MAX_BIG_CONTENT_SIZE", "160000"))
+
 
 def get_pretty_output() -> bool:
     """
@@ -212,7 +232,8 @@ def get_pretty_output() -> bool:
     返回：
         bool: 如果启用PrettyOutput则返回True，默认为True
     """
-    return GLOBAL_CONFIG_DATA.get('JARVIS_PRETTY_OUTPUT', False) == True
+    return GLOBAL_CONFIG_DATA.get("JARVIS_PRETTY_OUTPUT", False) == True
+
 
 def is_use_methodology() -> bool:
     """
@@ -221,7 +242,8 @@ def is_use_methodology() -> bool:
     返回：
         bool: 如果启用方法论则返回True，默认为True
     """
-    return GLOBAL_CONFIG_DATA.get('JARVIS_USE_METHODOLOGY', True) == True
+    return GLOBAL_CONFIG_DATA.get("JARVIS_USE_METHODOLOGY", True) == True
+
 
 def is_use_analysis() -> bool:
     """
@@ -230,7 +252,8 @@ def is_use_analysis() -> bool:
     返回：
         bool: 如果启用任务分析则返回True，默认为True
     """
-    return GLOBAL_CONFIG_DATA.get('JARVIS_USE_ANALYSIS', True) == True
+    return GLOBAL_CONFIG_DATA.get("JARVIS_USE_ANALYSIS", True) == True
+
 
 def is_print_prompt() -> bool:
     """
@@ -239,7 +262,7 @@ def is_print_prompt() -> bool:
     返回：
         bool: 如果打印提示则返回True，默认为True
     """
-    return GLOBAL_CONFIG_DATA.get('JARVIS_PRINT_PROMPT', False) == True
+    return GLOBAL_CONFIG_DATA.get("JARVIS_PRINT_PROMPT", False) == True
 
 
 def get_mcp_config() -> List[Dict[str, Any]]:
@@ -249,4 +272,4 @@ def get_mcp_config() -> List[Dict[str, Any]]:
     返回:
         List[Dict[str, Any]]: MCP配置项列表，如果未配置则返回空列表
     """
-    return GLOBAL_CONFIG_DATA.get("JARVIS_MCP", [])
+    return GLOBAL_CONFIG_DATA.get("JARVIS_MCP", [])

jarvis/jarvis_utils/embedding.py

@@ -1,9 +1,7 @@
 # -*- coding: utf-8 -*-
 import functools
 import os
-from typing import List
-
-from transformers import AutoTokenizer
+from typing import Any, List
 
 from jarvis.jarvis_utils.config import get_data_dir
 from jarvis.jarvis_utils.output import OutputType, PrettyOutput
@@ -11,6 +9,7 @@ from jarvis.jarvis_utils.output import OutputType, PrettyOutput
 # 全局缓存，避免重复加载模型
 _global_tokenizers = {}
 
+
 def get_context_token_count(text: str) -> int:
     """使用分词器获取文本的token数量。
 
@@ -21,12 +20,13 @@ def get_context_token_count(text: str) -> int:
         int: 文本中的token数量
     """
     try:
-        tokenizer = load_tokenizer()
+        from transformers import AutoTokenizer  # type: ignore
+        tokenizer : AutoTokenizer = load_tokenizer()
         # 分批处理长文本，确保不超过模型最大长度
         total_tokens = 0
         chunk_size = 100  # 每次处理100个字符，避免超过模型最大长度（考虑到中文字符可能被编码成多个token）
         for i in range(0, len(text), chunk_size):
-            chunk = text[i:i + chunk_size]
+            chunk = text[i : i + chunk_size]
             tokens = tokenizer.encode(chunk)  # type: ignore
             total_tokens += len(tokens)
         return total_tokens
@@ -34,7 +34,10 @@ def get_context_token_count(text: str) -> int:
         PrettyOutput.print(f"计算token失败: {str(e)}", OutputType.WARNING)
         return len(text) // 4  # 每个token大约4个字符的粗略估计
 
-def split_text_into_chunks(text: str, max_length: int = 512, min_length: int = 50) -> List[str]:
+
+def split_text_into_chunks(
+    text: str, max_length: int = 512, min_length: int = 50
+) -> List[str]:
     """将文本分割成块，基于token数量进行切割。
 
     参数：
@@ -52,15 +55,18 @@ def split_text_into_chunks(text: str, max_length: int = 512, min_length: int = 5
         chunks = []
         current_chunk = ""
         current_tokens = 0
-        
+
         # 按较大的块处理文本，避免破坏token边界
         chunk_size = 50  # 每次处理50个字符
         for i in range(0, len(text), chunk_size):
-            chunk = text[i:i + chunk_size]
+            chunk = text[i : i + chunk_size]
             chunk_tokens = get_context_token_count(chunk)
-            
+
             # 如果当前块加上新块会超过最大长度，且当前块已经达到最小长度，则保存当前块
-            if current_tokens + chunk_tokens > max_length and current_tokens >= min_length:
+            if (
+                current_tokens + chunk_tokens > max_length
+                and current_tokens >= min_length
+            ):
                 chunks.append(current_chunk)
                 current_chunk = chunk
                 current_tokens = chunk_tokens
@@ -77,17 +83,20 @@ def split_text_into_chunks(text: str, max_length: int = 512, min_length: int = 5
     except Exception as e:
         PrettyOutput.print(f"文本分割失败: {str(e)}", OutputType.WARNING)
         # 发生错误时回退到简单的字符分割
-        return [text[i:i + max_length] for i in range(0, len(text), max_length)]
+        return [text[i : i + max_length] for i in range(0, len(text), max_length)]
 
 
 @functools.lru_cache(maxsize=1)
-def load_tokenizer() -> AutoTokenizer:
+def load_tokenizer() -> Any:
     """
     加载用于文本处理的分词器，使用缓存避免重复加载。
 
     返回：
         AutoTokenizer: 加载的分词器
     """
+
+    from transformers import AutoTokenizer  # type: ignore
+
     model_name = "gpt2"
     cache_dir = os.path.join(get_data_dir(), "huggingface", "hub")
 
@@ -97,18 +106,14 @@ def load_tokenizer() -> AutoTokenizer:
 
     try:
         tokenizer = AutoTokenizer.from_pretrained(
-            model_name,
-            cache_dir=cache_dir,
-            local_files_only=True
+            model_name, cache_dir=cache_dir, local_files_only=True
         )
     except Exception:
         tokenizer = AutoTokenizer.from_pretrained(
-            model_name,
-            cache_dir=cache_dir,
-            local_files_only=False
+            model_name, cache_dir=cache_dir, local_files_only=False
         )
 
     # 保存到全局缓存
     _global_tokenizers[model_name] = tokenizer
 
-    return tokenizer # type: ignore
+    return tokenizer  # type: ignore

jarvis/jarvis_utils/file_processors.py

@@ -4,6 +4,7 @@ import unicodedata
 
 class FileProcessor:
     """Base class for file processor"""
+
     @staticmethod
     def can_handle(file_path: str) -> bool:
         """Determine if the file can be processed"""
@@ -14,9 +15,11 @@ class FileProcessor:
         """Extract file text content"""
         raise NotImplementedError
 
+
 class TextFileProcessor(FileProcessor):
     """Text file processor"""
-    ENCODINGS = ['utf-8', 'gbk', 'gb2312', 'latin1']
+
+    ENCODINGS = ["utf-8", "gbk", "gb2312", "latin1"]
     SAMPLE_SIZE = 8192  # Read the first 8KB to detect encoding
 
     @staticmethod
@@ -24,16 +27,20 @@ class TextFileProcessor(FileProcessor):
         """Determine if the file is a text file by trying to decode it"""
         try:
             # Read the first part of the file to detect encoding
-            with open(file_path, 'rb') as f:
+            with open(file_path, "rb") as f:
                 sample = f.read(TextFileProcessor.SAMPLE_SIZE)
 
             # Check if it contains null bytes (usually represents a binary file)
-            if b'\x00' in sample:
+            if b"\x00" in sample:
                 return False
 
             # Check if it contains too many non-printable characters (usually represents a binary file)
-            non_printable = sum(1 for byte in sample if byte < 32 and byte not in (9, 10, 13))  # tab, newline, carriage return
-            if non_printable / len(sample) > 0.3:  # If non-printable characters exceed 30%, it is considered a binary file
+            non_printable = sum(
+                1 for byte in sample if byte < 32 and byte not in (9, 10, 13)
+            )  # tab, newline, carriage return
+            if (
+                non_printable / len(sample) > 0.3
+            ):  # If non-printable characters exceed 30%, it is considered a binary file
                 return False
 
             # Try to decode with different encodings
@@ -55,7 +62,7 @@ class TextFileProcessor(FileProcessor):
         detected_encoding = None
         try:
             # First try to detect encoding
-            with open(file_path, 'rb') as f:
+            with open(file_path, "rb") as f:
                 raw_data = f.read()
 
             # Try different encodings
@@ -68,14 +75,14 @@ class TextFileProcessor(FileProcessor):
                     continue
 
             if not detected_encoding:
-                raise UnicodeDecodeError(f"Failed to decode file with supported encodings: {file_path}") # type: ignore
+                raise UnicodeDecodeError(f"Failed to decode file with supported encodings: {file_path}")  # type: ignore
 
             # Use the detected encoding to read the file
-            with open(file_path, 'r', encoding=detected_encoding, errors='ignore') as f:
+            with open(file_path, "r", encoding=detected_encoding, errors="ignore") as f:
                 content = f.read()
 
             # Normalize Unicode characters
-            content = unicodedata.normalize('NFKC', content)
+            content = unicodedata.normalize("NFKC", content)
 
             return content