jarvis-ai-assistant 0.1.66__tar.gz → 0.1.72__tar.gz

{jarvis_ai_assistant-0.1.66/src/jarvis_ai_assistant.egg-info → jarvis_ai_assistant-0.1.72}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: jarvis-ai-assistant
-Version: 0.1.66
+Version: 0.1.72
 Summary: Jarvis: An AI assistant that uses tools to interact with the system
 Home-page: https://github.com/skyfireitdiy/Jarvis
 Author: skyfire
@@ -44,9 +44,10 @@ Requires-Dist: colorama>=0.4.6
 Requires-Dist: prompt_toolkit>=3.0.0
 Requires-Dist: openai>=1.20.0
 Requires-Dist: playwright>=1.41.1
-Requires-Dist: numpy>=1.26.0
-Requires-Dist: faiss-cpu>=1.8.1
+Requires-Dist: numpy>=1.24.0
+Requires-Dist: faiss-cpu>=1.8.0
 Requires-Dist: sentence-transformers>=2.2.2
+Requires-Dist: bs4>=0.0.1
 Provides-Extra: dev
 Requires-Dist: pytest; extra == "dev"
 Requires-Dist: black; extra == "dev"
@@ -123,7 +124,6 @@ Jarvis supports configuration through environment variables that can be set in t
 |---------|------|--------|------|
 | JARVIS_PLATFORM | AI platform to use, supports kimi/openai/ai8 etc | kimi | Yes |
 | JARVIS_MODEL | Model name to use | - | No |
-
 | JARVIS_CODEGEN_PLATFORM | AI platform for code generation | Same as JARVIS_PLATFORM | No |
 | JARVIS_CODEGEN_MODEL | Model name for code generation | Same as JARVIS_MODEL | No |
 | JARVIS_CHEAP_PLATFORM | AI platform for cheap operations | Same as JARVIS_PLATFORM | No |
@@ -155,17 +155,17 @@ jarvis -p openai  # Use OpenAI platform
 
 ### Code Modification
 ```bash
-jarvis coder --feature "Add new feature"  # Modify code to add new feature
+jarvis-coder --feature "Add new feature"  # Modify code to add new feature
 ```
 
 ### Codebase Search
 ```bash
-jarvis codebase --search "database connection"  # Search codebase
+jarvis-codebase --search "database connection"  # Search codebase
 ```
 
 ### Codebase Question
 ```bash
-jarvis codebase --ask "How to use the database?"  # Ask about codebase
+jarvis-codebase --ask "How to use the database?"  # Ask about codebase
 ```
 
 ### Keep Chat History

{jarvis_ai_assistant-0.1.66 → jarvis_ai_assistant-0.1.72}/README.md

@@ -65,7 +65,6 @@ Jarvis supports configuration through environment variables that can be set in t
 |---------|------|--------|------|
 | JARVIS_PLATFORM | AI platform to use, supports kimi/openai/ai8 etc | kimi | Yes |
 | JARVIS_MODEL | Model name to use | - | No |
-
 | JARVIS_CODEGEN_PLATFORM | AI platform for code generation | Same as JARVIS_PLATFORM | No |
 | JARVIS_CODEGEN_MODEL | Model name for code generation | Same as JARVIS_MODEL | No |
 | JARVIS_CHEAP_PLATFORM | AI platform for cheap operations | Same as JARVIS_PLATFORM | No |
@@ -97,17 +96,17 @@ jarvis -p openai  # Use OpenAI platform
 
 ### Code Modification
 ```bash
-jarvis coder --feature "Add new feature"  # Modify code to add new feature
+jarvis-coder --feature "Add new feature"  # Modify code to add new feature
 ```
 
 ### Codebase Search
 ```bash
-jarvis codebase --search "database connection"  # Search codebase
+jarvis-codebase --search "database connection"  # Search codebase
 ```
 
 ### Codebase Question
 ```bash
-jarvis codebase --ask "How to use the database?"  # Ask about codebase
+jarvis-codebase --ask "How to use the database?"  # Ask about codebase
 ```
 
 ### Keep Chat History

{jarvis_ai_assistant-0.1.66 → jarvis_ai_assistant-0.1.72}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "jarvis-ai-assistant"
-version = "0.1.66"
+version = "0.1.72"
 description = "Jarvis: An AI assistant that uses tools to interact with the system"
 readme = "README.md"
 authors = [{ name = "Your Name", email = "your.email@example.com" }]
@@ -26,9 +26,10 @@ dependencies = [
     "prompt_toolkit>=3.0.0",
     "openai>=1.20.0",
     "playwright>=1.41.1",
-    "numpy>=1.26.0",
-    "faiss-cpu>=1.8.1",
+    "numpy>=1.24.0",
+    "faiss-cpu>=1.8.0",
     "sentence-transformers>=2.2.2",
+    "bs4>=0.0.1",
 ]
 requires-python = ">=3.8"
 

{jarvis_ai_assistant-0.1.66 → jarvis_ai_assistant-0.1.72}/setup.py

@@ -2,7 +2,7 @@ from setuptools import setup, find_packages
 
 setup(
     name="jarvis-ai-assistant",
-    version="0.1.66",
+    version="0.1.72",
     author="skyfire",
     author_email="skyfireitdiy@hotmail.com",
     description="An AI assistant that uses various tools to interact with the system",
@@ -19,9 +19,10 @@ setup(
         "prompt_toolkit>=3.0.0",
         "openai>=1.20.0",
         "playwright>=1.41.1",
-        "numpy>=1.26.0",
-        "faiss-cpu>=1.8.1",
+        "numpy>=1.24.0",
+        "faiss-cpu>=1.8.0",
         "sentence-transformers>=2.2.2",
+        "bs4>=0.0.1",
     ],
     entry_points={
         "console_scripts": [

{jarvis_ai_assistant-0.1.66 → jarvis_ai_assistant-0.1.72}/src/jarvis/__init__.py

@@ -1,3 +1,3 @@
 """Jarvis AI Assistant"""
 
-__version__ = "0.1.66" 
+__version__ = "0.1.72" 

jarvis_ai_assistant-0.1.72/src/jarvis/jarvis_codebase/main.py

@@ -0,0 +1,520 @@
+import hashlib
+import os
+import numpy as np
+import faiss
+from typing import List, Tuple, Optional
+from jarvis.models.registry import PlatformRegistry
+import concurrent.futures
+from threading import Lock
+from concurrent.futures import ThreadPoolExecutor
+from jarvis.utils import OutputType, PrettyOutput, find_git_root
+from jarvis.utils import load_env_from_file
+import argparse
+from sentence_transformers import SentenceTransformer
+import pickle
+
+class CodeBase:
+    def __init__(self, root_dir: str):
+        load_env_from_file()
+        self.root_dir = root_dir
+        os.chdir(self.root_dir)
+        self.thread_count = int(os.environ.get("JARVIS_THREAD_COUNT") or 10)
+        self.cheap_platform = os.environ.get("JARVIS_CHEAP_PLATFORM") or os.environ.get("JARVIS_PLATFORM") or "kimi"
+        self.cheap_model = os.environ.get("JARVIS_CHEAP_MODEL") or os.environ.get("JARVIS_MODEL") or "kimi"
+        self.normal_platform = os.environ.get("JARVIS_PLATFORM") or "kimi"
+        self.codegen_platform = os.environ.get("JARVIS_CODEGEN_PLATFORM") or os.environ.get("JARVIS_PLATFORM") or "kimi"
+        self.codegen_model = os.environ.get("JARVIS_CODEGEN_MODEL") or os.environ.get("JARVIS_MODEL") or "kimi"
+        self.normal_model = os.environ.get("JARVIS_MODEL") or "kimi"
+        self.embedding_model_name = os.environ.get("JARVIS_EMBEDDING_MODEL") or "BAAI/bge-large-zh-v1.5"
+        if not self.cheap_platform or not self.cheap_model or not self.codegen_platform or not self.codegen_model or not self.embedding_model_name or not self.normal_platform or not self.normal_model:
+            raise ValueError("JARVIS_CHEAP_PLATFORM or JARVIS_CHEAP_MODEL or JARVIS_CODEGEN_PLATFORM or JARVIS_CODEGEN_MODEL or JARVIS_EMBEDDING_MODEL or JARVIS_PLATFORM or JARVIS_MODEL is not set")
+        
+        PrettyOutput.print(f"廉价模型使用平台: {self.cheap_platform} 模型: {self.cheap_model}", output_type=OutputType.INFO)
+        PrettyOutput.print(f"代码生成模型使用平台: {self.codegen_platform} 模型: {self.codegen_model}", output_type=OutputType.INFO)
+        PrettyOutput.print(f"分析模型使用平台: {self.normal_platform} 模型: {self.normal_model}", output_type=OutputType.INFO)
+        PrettyOutput.print(f"嵌入模型: {self.embedding_model_name}", output_type=OutputType.INFO)
+        PrettyOutput.print(f"索引建立线程数: {self.thread_count}", output_type=OutputType.INFO)
+        PrettyOutput.print(f"检索算法：分层导航小世界算法", output_type=OutputType.INFO)
+            
+        # 初始化数据目录
+        self.data_dir = os.path.join(self.root_dir, ".jarvis-codebase")
+        if not os.path.exists(self.data_dir):
+            os.makedirs(self.data_dir)
+            
+        # 初始化嵌入模型，使用系统默认缓存目录
+        try:
+            os.environ["TOKENIZERS_PARALLELISM"] = "false"
+            PrettyOutput.print("正在加载/下载模型，请稍候...", output_type=OutputType.INFO)
+            self.embedding_model = SentenceTransformer(self.embedding_model_name)
+            
+            # 强制完全加载所有模型组件
+            test_text = """
+这是一段测试文本，用于确保模型完全加载。
+包含多行内容，以模拟实际使用场景。
+"""
+            # 预热模型，确保所有组件都被加载
+            self.embedding_model.encode([test_text], 
+                                     convert_to_tensor=True,
+                                     normalize_embeddings=True)
+            PrettyOutput.print("模型加载完成", output_type=OutputType.SUCCESS)
+        except Exception as e:
+            PrettyOutput.print(f"加载模型失败: {str(e)}", output_type=OutputType.ERROR)
+            raise
+            
+        self.vector_dim = self.embedding_model.get_sentence_embedding_dimension()
+
+        self.git_file_list = self.get_git_file_list()
+        self.platform_registry = PlatformRegistry().get_global_platform_registry()
+        
+        # 初始化缓存和索引
+        self.cache_path = os.path.join(self.data_dir, "cache.pkl")
+        self.vector_cache = {}
+        self.file_paths = []
+        
+        # 加载缓存
+        if os.path.exists(self.cache_path):
+            try:
+                with open(self.cache_path, 'rb') as f:
+                    cache_data = pickle.load(f)
+                    self.vector_cache = cache_data["vectors"]
+                    self.file_paths = cache_data["file_paths"]
+                PrettyOutput.print(f"加载了 {len(self.vector_cache)} 个向量缓存", 
+                                 output_type=OutputType.INFO)
+                # 从缓存重建索引
+                self.build_index()
+            except Exception as e:
+                PrettyOutput.print(f"加载缓存失败: {str(e)}", 
+                                 output_type=OutputType.WARNING)
+                self.vector_cache = {}
+                self.file_paths = []
+                self.index = None
+
+    def get_git_file_list(self):
+        """获取 git 仓库中的文件列表，排除 .jarvis-codebase 目录"""
+        files = os.popen("git ls-files").read().splitlines()
+        # 过滤掉 .jarvis-codebase 目录下的文件
+        return [f for f in files if not f.startswith(".jarvis-codebase/")]
+
+    def is_text_file(self, file_path: str):
+        with open(file_path, "r", encoding="utf-8") as f:
+            try:
+                f.read()
+                return True
+            except UnicodeDecodeError:
+                return False
+
+    def make_description(self, file_path: str) -> str:
+        model = self.platform_registry.create_platform(self.cheap_platform)
+        model.set_model_name(self.cheap_model)
+        model.set_suppress_output(True)
+        content = open(file_path, "r", encoding="utf-8").read()
+        prompt = f"""请分析以下代码文件，并生成一个详细的描述。描述应该包含以下要点：
+
+1. 主要功能和用途
+2. 关键类和方法的作用
+3. 重要的依赖和技术特征（如使用了什么框架、算法、设计模式等）
+4. 代码处理的主要数据类型和数据结构
+5. 关键业务逻辑和处理流程
+6. 特殊功能点和亮点特性
+
+请用简洁专业的语言描述，突出代码的技术特征和功能特点，以便后续进行相似代码检索。
+
+文件路径：{file_path}
+代码内容：
+{content}
+"""
+        response = model.chat(prompt)
+        return response
+
+    def save_cache(self):
+        """保存缓存数据"""
+        try:
+            cache_data = {
+                "vectors": self.vector_cache,
+                "file_paths": self.file_paths
+            }
+            with open(self.cache_path, 'wb') as f:
+                pickle.dump(cache_data, f)
+            PrettyOutput.print(f"保存了 {len(self.vector_cache)} 个向量缓存", 
+                             output_type=OutputType.INFO)
+        except Exception as e:
+            PrettyOutput.print(f"保存缓存失败: {str(e)}", 
+                             output_type=OutputType.ERROR)
+
+    def get_cached_vector(self, file_path: str, description: str) -> Optional[np.ndarray]:
+        """从缓存获取文件的向量表示"""
+        if file_path not in self.vector_cache:
+            return None
+        
+        # 检查文件是否被修改
+        try:
+            with open(file_path, "rb") as f:
+                current_md5 = hashlib.md5(f.read()).hexdigest()
+        except Exception as e:
+            PrettyOutput.print(f"计算文件MD5失败 {file_path}: {str(e)}", 
+                              output_type=OutputType.ERROR)
+            return None
+        
+        cached_data = self.vector_cache[file_path]
+        if cached_data["md5"] != current_md5:
+            return None
+        
+        # 检查描述是否变化
+        if cached_data["description"] != description:
+            return None
+        
+        return cached_data["vector"]
+
+    def cache_vector(self, file_path: str, vector: np.ndarray, description: str):
+        """缓存文件的向量表示"""
+        try:
+            with open(file_path, "rb") as f:
+                file_md5 = hashlib.md5(f.read()).hexdigest()
+        except Exception as e:
+            PrettyOutput.print(f"计算文件MD5失败 {file_path}: {str(e)}", 
+                              output_type=OutputType.ERROR)
+            file_md5 = ""
+        
+        self.vector_cache[file_path] = {
+            "path": file_path,  # 保存文件路径
+            "md5": file_md5,    # 保存文件MD5
+            "description": description,  # 保存文件描述
+            "vector": vector    # 保存向量
+        }
+        
+        # 保存缓存到文件
+        try:
+            with open(self.cache_path, 'wb') as f:
+                cache_data = {
+                    "vectors": self.vector_cache,
+                    "file_paths": self.file_paths
+                }
+                pickle.dump(cache_data, f)
+        except Exception as e:
+            PrettyOutput.print(f"保存向量缓存失败: {str(e)}", 
+                              output_type=OutputType.ERROR)
+
+    def get_embedding(self, text: str) -> np.ndarray:
+        """使用 transformers 模型获取文本的向量表示"""
+        # 对长文本进行截断
+        max_length = 512  # 或其他合适的长度
+        text = ' '.join(text.split()[:max_length])
+        
+        # 获取嵌入向量
+        embedding = self.embedding_model.encode(text, 
+                                                 normalize_embeddings=True,  # L2归一化
+                                                 show_progress_bar=False)
+        vector = np.array(embedding, dtype=np.float32)
+        return vector
+
+    def vectorize_file(self, file_path: str, description: str) -> np.ndarray:
+        """将文件内容和描述向量化"""
+        try:
+            # 先尝试从缓存获取
+            cached_vector = self.get_cached_vector(file_path, description)
+            if cached_vector is not None:
+                return cached_vector
+                
+            # 组合文件信息
+            combined_text = f"""
+文件路径: {file_path}
+文件描述: {description}
+"""
+            vector = self.get_embedding(combined_text)
+            
+            # 保存到缓存
+            self.cache_vector(file_path, vector, description)
+            return vector
+        except Exception as e:
+            PrettyOutput.print(f"Error vectorizing file {file_path}: {str(e)}", 
+                             output_type=OutputType.ERROR)
+            return np.zeros(self.vector_dim, dtype=np.float32)
+
+    def clean_cache(self) -> bool:
+        """清理过期的缓存记录，返回是否有文件被删除"""
+        files_to_delete = []
+        for file_path in list(self.vector_cache.keys()):
+            if file_path not in self.git_file_list:
+                del self.vector_cache[file_path]
+                files_to_delete.append(file_path)
+        
+        if files_to_delete:
+            self.save_cache()
+            PrettyOutput.print(f"清理了 {len(files_to_delete)} 个文件的缓存", 
+                             output_type=OutputType.INFO)
+            return True
+        return False
+
+    def process_file(self, file_path: str):
+        """处理单个文件"""
+        try:
+            # 跳过不存在的文件
+            if not os.path.exists(file_path):
+                return None
+                
+            if not self.is_text_file(file_path):
+                return None
+                
+            md5 = hashlib.md5(open(file_path, "rb").read()).hexdigest()
+            
+            # 检查文件是否已经处理过且内容未变
+            if file_path in self.vector_cache:
+                if self.vector_cache[file_path].get("md5") == md5:
+                    return None
+                    
+            description = self.make_description(file_path)
+            vector = self.vectorize_file(file_path, description)
+            
+            # 保存到缓存，使用实际文件路径作为键
+            self.vector_cache[file_path] = {
+                "vector": vector,
+                "description": description,
+                "md5": md5
+            }
+            
+            return file_path
+            
+        except Exception as e:
+            PrettyOutput.print(f"处理文件失败 {file_path}: {str(e)}", 
+                             output_type=OutputType.ERROR,
+                             traceback=True)
+            return None
+
+    def build_index(self):
+        """从向量缓存构建 faiss 索引"""
+        # 创建底层 HNSW 索引
+        hnsw_index = faiss.IndexHNSWFlat(self.vector_dim, 16)
+        hnsw_index.hnsw.efConstruction = 40
+        hnsw_index.hnsw.efSearch = 16
+        
+        # 用 IndexIDMap 包装 HNSW 索引
+        self.index = faiss.IndexIDMap(hnsw_index)
+        
+        vectors = []
+        ids = []
+        self.file_paths = []  # 重置文件路径列表
+        
+        for i, (file_path, data) in enumerate(self.vector_cache.items()):
+            vectors.append(data["vector"].reshape(1, -1))
+            ids.append(i)
+            self.file_paths.append(file_path)
+            
+        if vectors:
+            vectors = np.vstack(vectors)
+            self.index.add_with_ids(vectors, np.array(ids))
+        else:
+            self.index = None
+
+    def gen_vector_db_from_cache(self):
+        """从缓存生成向量数据库"""
+        self.build_index()
+        self.save_cache()
+
+    def generate_codebase(self):
+        """生成代码库索引"""
+        files_deleted = self.clean_cache()  # 清理过期缓存
+        processed_files = []
+        
+        # 使用线程池处理文件
+        with ThreadPoolExecutor(max_workers=self.thread_count) as executor:
+            futures = [executor.submit(self.process_file, file) for file in self.git_file_list]
+            for future in concurrent.futures.as_completed(futures):
+                result = future.result()
+                if result:
+                    processed_files.append(result)
+                    PrettyOutput.print(f"索引文件: {result}", output_type=OutputType.INFO)
+
+        if files_deleted or processed_files:
+            PrettyOutput.print("重新生成向量数据库", output_type=OutputType.INFO)
+            self.gen_vector_db_from_cache()
+        else:
+            PrettyOutput.print("没有新的文件变更，跳过向量数据库生成", output_type=OutputType.INFO)
+            
+        PrettyOutput.print(f"成功为 {len(processed_files)} 个文件生成索引", output_type=OutputType.INFO)
+
+    def rerank_results(self, query: str, initial_results: List[Tuple[str, float, str]]) -> List[Tuple[str, float, str]]:
+        """使用大模型对搜索结果重新排序"""
+        if not initial_results:
+            return []
+
+        model = self.platform_registry.create_platform(self.normal_platform)
+        model.set_model_name(self.normal_model)
+        model.set_suppress_output(True)
+
+        try:
+            # 构建重排序的prompt
+            prompt = f"""请根据用户的查询，对以下代码文件进行相关性排序。对每个文件给出0-100的相关性分数，分数越高表示越相关。
+只需要输出每个文件的分数，格式为：
+<RERANK_START>
+文件路径: 分数
+文件路径: 分数
+<RERANK_END>
+
+用户查询: {query}
+
+待评估文件:
+"""
+            for path, _, desc in initial_results:
+                prompt += f"""
+文件: {path}
+描述: {desc}
+---
+"""
+            
+            response = model.chat(prompt)
+            
+            # 提取<RERANK_START>和<RERANK_END>之间的内容
+            start_tag = "<RERANK_START>"
+            end_tag = "<RERANK_END>"
+            if start_tag in response and end_tag in response:
+                response = response[response.find(start_tag) + len(start_tag):response.find(end_tag)]
+            
+            # 解析响应，提取文件路径和分数
+            scored_results = []
+            for line in response.split('\n'):
+                if ':' not in line:
+                    continue
+                try:
+                    file_path, score_str = line.split(':', 1)
+                    file_path = file_path.strip()
+                    score = float(score_str.strip())
+                    # 找到对应的原始描述
+                    desc = next((desc for p, _, desc in initial_results if p == file_path), "")
+                    scored_results.append((file_path, score/100.0, desc))
+                except:
+                    continue
+            
+            # 按分数降序排序
+            return sorted(scored_results, key=lambda x: x[1], reverse=True)
+            
+        finally:
+            model.delete_chat()
+        
+        return initial_results
+
+    def search_similar(self, query: str, top_k: int = 20) -> List[Tuple[str, float, str]]:
+        """搜索相似文件"""
+        model = self.platform_registry.create_platform(self.normal_platform)
+        model.set_model_name(self.normal_model)
+        model.set_suppress_output(True)
+        
+        try:
+            prompt = f"""请根据以下查询，生成意思完全相同的另一个表述。这个表述将用于代码搜索，所以要保持专业性和准确性。
+原始查询: {query}
+
+请直接输出新表述，不要有编号或其他标记。
+"""
+            
+            query = model.chat(prompt)
+            
+        finally:
+            model.delete_chat()
+
+        PrettyOutput.print(f"查询:  {query}", output_type=OutputType.INFO)
+        
+        # 为每个查询获取相似文件
+        all_results = {}  # 文件路径 -> (总分数, 出现次数, 描述)
+        
+        q_vector = self.get_embedding(query)
+        q_vector = q_vector.reshape(1, -1)
+            
+        distances, indices = self.index.search(q_vector, top_k)
+
+        PrettyOutput.print(f"查询 {query} 的结果: ", output_type=OutputType.INFO)
+
+        initial_results = []
+        
+        for i, distance in zip(indices[0], distances[0]):
+            if i == -1:  # faiss返回-1表示无效结果
+                continue
+                
+            similarity = 1.0 / (1.0 + float(distance))
+            PrettyOutput.print(f"  {self.file_paths[i]} : 距离 {distance:.3f}, 相似度 {similarity:.3f}", 
+                                output_type=OutputType.INFO)
+                
+            file_path = self.file_paths[i]
+            data = self.vector_cache[file_path]
+            initial_results.append((file_path, similarity, data["description"]))
+
+        # 使用大模型重新排序
+        PrettyOutput.print("使用大模型重新排序...", output_type=OutputType.INFO)
+        reranked_results = self.rerank_results(query, initial_results)
+        
+        return reranked_results
+
+    def ask_codebase(self, query: str, top_k: int=20) -> str:
+        """查询代码库"""
+        results = self.search_similar(query, top_k)
+        PrettyOutput.print(f"找到的关联文件: ", output_type=OutputType.SUCCESS)
+        for path, score, _ in results:
+            PrettyOutput.print(f"文件: {path} 关联度: {score:.3f}", 
+                             output_type=OutputType.INFO)
+        
+        prompt = f"""你是一个代码专家，请根据以下文件信息回答用户的问题：
+"""
+        for path, _, _ in results:
+            try:
+                if len(prompt) > 30 * 1024:
+                    PrettyOutput.print(f"避免上下文超限，丢弃低相关度文件：{path}", OutputType.WARNING)
+                    continue
+                content = open(path, "r", encoding="utf-8").read()
+                prompt += f"""
+文件路径: {path}prompt
+文件内容:
+{content}
+========================================
+"""
+            except Exception as e:
+                PrettyOutput.print(f"读取文件失败 {path}: {str(e)}", 
+                                 output_type=OutputType.ERROR)
+                continue
+                
+        prompt += f"""
+用户问题: {query}
+
+请用专业的语言回答用户的问题，如果给出的文件内容不足以回答用户的问题，请告诉用户，绝对不要胡编乱造。
+"""
+        model = self.platform_registry.create_platform(self.codegen_platform)
+        model.set_model_name(self.codegen_model)
+        try:
+            response = model.chat(prompt)
+            return response
+        finally:
+            model.delete_chat()
+
+
+def main():
+    parser = argparse.ArgumentParser(description='Codebase management and search tool')
+    parser.add_argument('--search', type=str, help='Search query to find similar code files')
+    parser.add_argument('--top-k', type=int, default=20, help='Number of results to return (default: 20)')
+    parser.add_argument('--ask', type=str, help='Ask a question about the codebase')
+    args = parser.parse_args()
+    
+    current_dir = find_git_root()
+    codebase = CodeBase(current_dir)
+
+    try:
+        codebase.generate_codebase()
+        PrettyOutput.print("\nCodebase generation completed", output_type=OutputType.SUCCESS)
+    except Exception as e:
+        PrettyOutput.print(f"Error during codebase generation: {str(e)}", output_type=OutputType.ERROR)
+    
+    if args.search:
+        results = codebase.search_similar(args.search, args.top_k)
+        if not results:
+            PrettyOutput.print("No similar files found", output_type=OutputType.WARNING)
+            return
+            
+        PrettyOutput.print("\nSearch Results:", output_type=OutputType.INFO)
+        for path, score, desc in results:
+            PrettyOutput.print("\n" + "="*50, output_type=OutputType.INFO)
+            PrettyOutput.print(f"File: {path}", output_type=OutputType.INFO)
+            PrettyOutput.print(f"Similarity: {score:.3f}", output_type=OutputType.INFO)
+            PrettyOutput.print(f"Description: {desc[100:]}", output_type=OutputType.INFO)
+
+    if args.ask:            
+        codebase.ask_codebase(args.ask, args.top_k)
+
+
+if __name__ == "__main__":
+    exit(main())

{jarvis_ai_assistant-0.1.66 → jarvis_ai_assistant-0.1.72}/src/jarvis/jarvis_coder/main.py

@@ -1,11 +1,8 @@
-from concurrent.futures import ThreadPoolExecutor, as_completed
-import hashlib
 import os
 import re
-import sqlite3
 import threading
 import time
-from typing import Dict, Any, List, Optional, Tuple
+from typing import Dict, Any, List, Tuple
 
 import yaml
 from jarvis.models.base import BasePlatform
@@ -127,7 +124,7 @@ class JarvisCoder:
             return [patch.replace('<PATCH_START>', '').replace('<PATCH_END>', '').strip() 
                    for patch in patches if patch.strip()]
         except Exception as e:
-            PrettyOutput.print(f"解析patch失败: {str(e)}", OutputType.ERROR)
+            PrettyOutput.print(f"解析patch失败: {str(e)}", OutputType.WARNING)
             return []
         
     def _make_patch(self, related_files: List[Dict], feature: str) -> List[str]:
@@ -141,20 +138,31 @@ class JarvisCoder:
 要替换的内容
 =======
 新的内容
-<<<<<<
+>>>>>>
 <PATCH_END>
 
-2. 如果是新文件，格式如下：
+2. 如果是新文件或者替换整个文件内容，格式如下：
 <PATCH_START>
 >>>>>> path/to/new/file
 =======
 新文件的完整内容
-<<<<<<
+>>>>>>
+<PATCH_END>
+
+3. 如果要删除文件中的某一段，格式如下：
+<PATCH_START>
+>>>>>> path/to/file
+要删除的内容
+=======
+>>>>>>
 <PATCH_END>
 
 文件列表如下：
 """
         for i, file in enumerate(related_files):
+            if len(prompt) > 30 * 1024:
+                PrettyOutput.print(f'避免上下文超限，丢弃低相关度文件：{file["file_path"]}', OutputType.WARNING)
+                continue
             prompt += f"""{i}. {file["file_path"]}\n"""
             prompt += f"""文件内容:\n"""
             prompt += f"<FILE_CONTENT_START>\n"
@@ -180,7 +188,7 @@ class JarvisCoder:
             return [patch.replace('<PATCH_START>', '').replace('<PATCH_END>', '').strip() 
                    for patch in patches if patch.strip()]
         except Exception as e:
-            PrettyOutput.print(f"解析patch失败: {str(e)}", OutputType.ERROR)
+            PrettyOutput.print(f"解析patch失败: {str(e)}", OutputType.WARNING)
             return []
 
     def _apply_patch(self, related_files: List[Dict], patches: List[str]) -> Tuple[bool, str]:
@@ -219,7 +227,7 @@ class JarvisCoder:
                     return False, "\n".join(error_info)
                 
                 old_content = parts[0]
-                new_content = parts[1].split("<<<<<<")[0]
+                new_content = parts[1].split(">>>>>>")[0]
                 
                 # 处理新文件
                 if not old_content:
@@ -308,9 +316,9 @@ class JarvisCoder:
     def _load_related_files(self, feature: str) -> List[Dict]:
         """加载相关文件内容"""
         ret = []
-        related_files = self._codebase.search_similar(feature, top_k=5)
+        related_files = self._codebase.search_similar(feature)
         for file, score, _ in related_files:
-            PrettyOutput.print(f"相关文件: {file} 相关度: {score:.3f}", OutputType.INFO)
+            PrettyOutput.print(f"相关文件: {file} 相关度: {score:.3f}", OutputType.SUCCESS)
             with open(file, "r", encoding="utf-8") as f:
                 content = f.read()
             ret.append({"file_path": file, "file_content": content})
@@ -441,7 +449,7 @@ def main():
                 PrettyOutput.print(result["stdout"], OutputType.SUCCESS)
             else:
                 if result["stderr"]:
-                    PrettyOutput.print(result["stderr"], OutputType.ERROR)
+                    PrettyOutput.print(result["stderr"], OutputType.WARNING)
                 if result["error"]:
                     PrettyOutput.print(f"错误类型: {type(result['error']).__name__}", OutputType.WARNING)
                 

{jarvis_ai_assistant-0.1.66 → jarvis_ai_assistant-0.1.72}/src/jarvis/models/registry.py

@@ -121,7 +121,7 @@ class PlatformRegistry:
                             # 检查平台实现
                             if not PlatformRegistry.check_platform_implementation(obj):
                                 continue
-
+                            PrettyOutput.print(f"从 {os.path.join(directory, filename)} 加载平台：{obj.platform_name}", OutputType.SUCCESS)
                             platforms[obj.platform_name] = obj
                             break
                 except Exception as e:

jarvis_ai_assistant-0.1.72/src/jarvis/tools/__init__.py

@@ -0,0 +1,6 @@
+from .registry import ToolRegistry
+
+__all__ = [
+    'ToolRegistry',
+] 
+

{jarvis_ai_assistant-0.1.66 → jarvis_ai_assistant-0.1.72}/src/jarvis/tools/registry.py

@@ -99,7 +99,7 @@ class ToolRegistry:
                         parameters=tool_instance.parameters,
                         func=tool_instance.execute
                     )
-                    PrettyOutput.print(f"从 {file_path} 加载工具: {tool_instance.name}: {tool_instance.description}", OutputType.INFO)
+                    PrettyOutput.print(f"从 {file_path} 加载工具: {tool_instance.name}: {tool_instance.description}", OutputType.SUCCESS)
                     tool_found = True
                     break
                     

{jarvis_ai_assistant-0.1.66 → jarvis_ai_assistant-0.1.72}/src/jarvis/utils.py

@@ -158,7 +158,7 @@ def get_multiline_input(tip: str) -> str:
             lines.append(line)
             
     except KeyboardInterrupt:
-        PrettyOutput.print("\n输入已取消", OutputType.ERROR)
+        PrettyOutput.print("\n输入已取消", OutputType.INFO)
         return "__interrupt__"
     
     return "\n".join(lines)

{jarvis_ai_assistant-0.1.66 → jarvis_ai_assistant-0.1.72/src/jarvis_ai_assistant.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: jarvis-ai-assistant
-Version: 0.1.66
+Version: 0.1.72
 Summary: Jarvis: An AI assistant that uses tools to interact with the system
 Home-page: https://github.com/skyfireitdiy/Jarvis
 Author: skyfire
@@ -44,9 +44,10 @@ Requires-Dist: colorama>=0.4.6
 Requires-Dist: prompt_toolkit>=3.0.0
 Requires-Dist: openai>=1.20.0
 Requires-Dist: playwright>=1.41.1
-Requires-Dist: numpy>=1.26.0
-Requires-Dist: faiss-cpu>=1.8.1
+Requires-Dist: numpy>=1.24.0
+Requires-Dist: faiss-cpu>=1.8.0
 Requires-Dist: sentence-transformers>=2.2.2
+Requires-Dist: bs4>=0.0.1
 Provides-Extra: dev
 Requires-Dist: pytest; extra == "dev"
 Requires-Dist: black; extra == "dev"
@@ -123,7 +124,6 @@ Jarvis supports configuration through environment variables that can be set in t
 |---------|------|--------|------|
 | JARVIS_PLATFORM | AI platform to use, supports kimi/openai/ai8 etc | kimi | Yes |
 | JARVIS_MODEL | Model name to use | - | No |
-
 | JARVIS_CODEGEN_PLATFORM | AI platform for code generation | Same as JARVIS_PLATFORM | No |
 | JARVIS_CODEGEN_MODEL | Model name for code generation | Same as JARVIS_MODEL | No |
 | JARVIS_CHEAP_PLATFORM | AI platform for cheap operations | Same as JARVIS_PLATFORM | No |
@@ -155,17 +155,17 @@ jarvis -p openai  # Use OpenAI platform
 
 ### Code Modification
 ```bash
-jarvis coder --feature "Add new feature"  # Modify code to add new feature
+jarvis-coder --feature "Add new feature"  # Modify code to add new feature
 ```
 
 ### Codebase Search
 ```bash
-jarvis codebase --search "database connection"  # Search codebase
+jarvis-codebase --search "database connection"  # Search codebase
 ```
 
 ### Codebase Question
 ```bash
-jarvis codebase --ask "How to use the database?"  # Ask about codebase
+jarvis-codebase --ask "How to use the database?"  # Ask about codebase
 ```
 
 ### Keep Chat History

{jarvis_ai_assistant-0.1.66 → jarvis_ai_assistant-0.1.72}/src/jarvis_ai_assistant.egg-info/SOURCES.txt

@@ -9,6 +9,7 @@ src/jarvis/main.py
 src/jarvis/utils.py
 src/jarvis/jarvis_codebase/__init__.py
 src/jarvis/jarvis_codebase/main.py
+src/jarvis/jarvis_coder/__init__.py
 src/jarvis/jarvis_coder/main.py
 src/jarvis/models/__init__.py
 src/jarvis/models/ai8.py

{jarvis_ai_assistant-0.1.66 → jarvis_ai_assistant-0.1.72}/src/jarvis_ai_assistant.egg-info/requires.txt

@@ -4,9 +4,10 @@ colorama>=0.4.6
 prompt_toolkit>=3.0.0
 openai>=1.20.0
 playwright>=1.41.1
-numpy>=1.26.0
-faiss-cpu>=1.8.1
+numpy>=1.24.0
+faiss-cpu>=1.8.0
 sentence-transformers>=2.2.2
+bs4>=0.0.1
 
 [dev]
 pytest

jarvis_ai_assistant-0.1.66/src/jarvis/jarvis_codebase/main.py

@@ -1,342 +0,0 @@
-import hashlib
-import os
-import sqlite3
-import time
-import numpy as np
-import faiss
-from typing import List, Tuple, Optional
-from jarvis.models.registry import PlatformRegistry
-import concurrent.futures
-from threading import Lock
-from concurrent.futures import ThreadPoolExecutor
-from jarvis.utils import OutputType, PrettyOutput, find_git_root
-from jarvis.utils import load_env_from_file
-import argparse
-from sentence_transformers import SentenceTransformer
-
-class CodeBase:
-    def __init__(self, root_dir: str, thread_count: int = 10):
-        load_env_from_file()
-        self.root_dir = root_dir
-        os.chdir(self.root_dir)
-        self.thread_count = thread_count
-        self.cheap_platform = os.environ.get("JARVIS_CHEAP_PLATFORM") or os.environ.get("JARVIS_PLATFORM") or "kimi"
-        self.cheap_model = os.environ.get("JARVIS_CHEAP_MODEL") or os.environ.get("JARVIS_MODEL") or "kimi"
-        self.normal_platform = os.environ.get("JARVIS_PLATFORM") or "kimi"
-        self.normal_model = os.environ.get("JARVIS_MODEL") or "kimi"
-        self.embedding_model_name = os.environ.get("JARVIS_EMBEDDING_MODEL") or "BAAI/bge-large-zh-v1.5"
-        if not self.cheap_platform or not self.cheap_model or not self.embedding_model_name or not self.normal_platform or not self.normal_model:
-            raise ValueError("JARVIS_CHEAP_PLATFORM or JARVIS_CHEAP_MODEL or JARVIS_EMBEDDING_MODEL or JARVIS_PLATFORM or JARVIS_MODEL is not set")
-        
-        PrettyOutput.print(f"廉价模型使用平台: {self.cheap_platform} 模型: {self.cheap_model}", output_type=OutputType.INFO)
-        PrettyOutput.print(f"分析模型使用平台: {self.normal_platform} 模型: {self.normal_model}", output_type=OutputType.INFO)
-        PrettyOutput.print(f"嵌入模型: {self.embedding_model_name}", output_type=OutputType.INFO)
-        PrettyOutput.print(f"检索算法：分层导航小世界算法", output_type=OutputType.INFO)
-            
-        # 初始化数据目录
-        self.data_dir = os.path.join(self.root_dir, ".jarvis-codebase")
-        if not os.path.exists(self.data_dir):
-            os.makedirs(self.data_dir)
-            
-        # 初始化嵌入模型，使用系统默认缓存目录
-        try:
-            PrettyOutput.print("正在加载/下载模型，请稍候...", output_type=OutputType.INFO)
-            self.embedding_model = SentenceTransformer(self.embedding_model_name)
-            
-            # 强制完全加载所有模型组件
-            test_text = """
-这是一段测试文本，用于确保模型完全加载。
-包含多行内容，以模拟实际使用场景。
-"""
-            # 预热模型，确保所有组件都被加载
-            self.embedding_model.encode([test_text], 
-                                     convert_to_tensor=True,
-                                     normalize_embeddings=True)
-            PrettyOutput.print("模型加载完成", output_type=OutputType.SUCCESS)
-        except Exception as e:
-            PrettyOutput.print(f"加载模型失败: {str(e)}", output_type=OutputType.ERROR)
-            raise
-            
-        self.vector_dim = self.embedding_model.get_sentence_embedding_dimension()
-
-            
-        self.db_path = os.path.join(self.data_dir, "codebase.db")
-        if not os.path.exists(self.db_path):
-            self.create_db()
-        self.git_file_list = self.get_git_file_list()
-        self.platform_registry = PlatformRegistry().get_global_platform_registry()
-        self.index_path = os.path.join(self.data_dir, "vectors.index")
-        self.index = None
-        if  os.path.exists(self.index_path):
-            PrettyOutput.print("正在加载向量数据库", output_type=OutputType.INFO)
-            self.index = faiss.read_index(self.index_path)
-
-    def get_git_file_list(self):
-        return os.popen("git ls-files").read().splitlines()
-
-    def get_db_connection(self):
-        """创建并返回一个新的数据库连接"""
-        return sqlite3.connect(self.db_path)
-
-    def clean_db(self) -> bool:
-        """清理数据库和向量索引中的过期记录"""
-        db = self.get_db_connection()
-        try:
-            # 获取所有数据库记录
-            all_records = db.execute("SELECT path FROM codebase").fetchall()
-            files_to_delete = []
-            
-            # 找出需要删除的文件
-            for row in all_records:
-                if row[0] not in self.git_file_list:
-                    files_to_delete.append(row[0])
-            
-            if not files_to_delete:
-                return False
-                
-            for file_path in files_to_delete:
-                db.execute("DELETE FROM codebase WHERE path = ?", (file_path,))
-            
-            db.commit()
-            
-            PrettyOutput.print(f"清理了 {len(files_to_delete)} 个文件的记录", 
-                             output_type=OutputType.INFO)
-            return True
-        finally:
-            db.close()
-        
-    def create_db(self):
-        db = self.get_db_connection()
-        try:
-            db.execute("CREATE TABLE IF NOT EXISTS codebase (path TEXT, md5 TEXT ,description TEXT)")
-            db.commit()
-        finally:
-            db.close()
-
-    def is_text_file(self, file_path: str):
-        with open(file_path, "r", encoding="utf-8") as f:
-            try:
-                f.read()
-                return True
-            except UnicodeDecodeError:
-                return False
-
-    def make_description(self, file_path: str) -> str:
-        model = self.platform_registry.create_platform(self.cheap_platform)
-        model.set_model_name(self.cheap_model)
-        model.set_suppress_output(True)
-        content = open(file_path, "r", encoding="utf-8").read()
-        prompt = f"""请分析以下代码文件，并生成一个详细的描述。描述应该包含以下要点：
-
-1. 主要功能和用途
-2. 关键类和方法的作用
-3. 重要的依赖和技术特征（如使用了什么框架、算法、设计模式等）
-4. 代码处理的主要数据类型和数据结构
-5. 关键业务逻辑和处理流程
-6. 特殊功能点和亮点特性
-
-请用简洁专业的语言描述，突出代码的技术特征和功能特点，以便后续进行相似代码检索。
-
-文件路径：{file_path}
-代码内容：
-{content}
-"""
-        response = model.chat(prompt)
-        return response
-
-    def get_embedding(self, text: str) -> np.ndarray:
-        """使用 transformers 模型获取文本的向量表示"""
-        # 对长文本进行截断
-        max_length = 512  # 或其他合适的长度
-        text = ' '.join(text.split()[:max_length])
-        
-        # 获取嵌入向量
-        embedding = self.embedding_model.encode(text, 
-                                             normalize_embeddings=True,  # L2归一化
-                                             show_progress_bar=False)
-        return np.array(embedding, dtype=np.float32)
-
-    def vectorize_file(self, file_path: str, description: str) -> np.ndarray:
-        """将文件内容和描述向量化"""
-        try:            
-            # 组合文件信息
-            combined_text = f"""
-文件路径: {file_path}
-文件描述: {description}
-"""
-            return self.get_embedding(combined_text)
-        except Exception as e:
-            PrettyOutput.print(f"Error vectorizing file {file_path}: {str(e)}", 
-                             output_type=OutputType.ERROR)
-            return np.zeros(self.vector_dim, dtype=np.float32)
-
-    def process_file(self, file):
-        """处理单个文件的辅助方法"""
-        db = self.get_db_connection()
-        try:
-            if not self.is_text_file(file):
-                return None
-            md5 = hashlib.md5(open(file, "rb").read()).hexdigest()
-            if db.execute("SELECT path FROM codebase WHERE md5 = ?", (md5,)).fetchone():
-                return None
-            description = self.make_description(file)
-            return (file, md5, description)
-        finally:
-            db.close()
-
-    def gen_vector_db_from_sqlite(self):
-        self.index = faiss.IndexHNSWFlat(self.vector_dim, 16)
-        self.index.hnsw.efConstruction = 40
-        self.index.hnsw.efSearch = 16
-        db = self.get_db_connection()
-        try:
-            all_records = db.execute("SELECT path, description FROM codebase").fetchall()
-            for row in all_records:
-                file, description = row
-                PrettyOutput.print(f"正在向量化文件: {file}", output_type=OutputType.INFO)
-                vector = self.vectorize_file(file, description)
-                vector = vector.reshape(1, -1)  
-                self.index.add(vector)
-            faiss.write_index(self.index, self.index_path)
-        finally:
-            db.close()
-
-    def generate_codebase(self):
-        updated =self.clean_db()
-        db_lock = Lock()
-        processed_files = []  # 用于跟踪已处理的文件
-        
-        def process_and_save(file):
-            result = self.process_file(file)
-            if result:
-                file, md5, description = result
-                db = self.get_db_connection()
-                try:
-                    with db_lock:
-                        db.execute("DELETE FROM codebase WHERE path = ?", (file,))
-                        db.execute("INSERT INTO codebase (path, md5, description) VALUES (?, ?, ?)", 
-                                 (file, md5, description))
-                        db.commit()
-                        PrettyOutput.print(f"索引文件: {file}", output_type=OutputType.INFO)
-                        processed_files.append(file)
-                finally:
-                    db.close()
-        
-        # 使用 ThreadPoolExecutor 并等待所有任务完成
-        with ThreadPoolExecutor(max_workers=self.thread_count) as executor:
-            futures = [executor.submit(process_and_save, file) for file in self.git_file_list]
-            # 等待所有任务完成
-            concurrent.futures.wait(futures)
-
-        if updated or len(processed_files) > 0:
-            PrettyOutput.print("有新的文件被删除或添加，正在重新生成向量数据库", output_type=OutputType.INFO)
-            self.gen_vector_db_from_sqlite()
-        else:
-            PrettyOutput.print("没有新的文件被删除或添加，跳过向量数据库生成", output_type=OutputType.INFO)
-            
-        PrettyOutput.print(f"成功索引 {len(processed_files)} 个文件", output_type=OutputType.INFO)
-
-    def search_similar(self, query: str, top_k: int = 5) -> List[Tuple[str, float, str]]:
-        """搜索与查询最相似的文件
-        
-        Args:
-            query: 查询文本
-            top_k: 返回结果数量
-            
-        Returns:
-            List of (file_path, similarity_score, description) tuples
-        """
-        # 获取查询文本的向量表示
-        query_vector = self.get_embedding(query)
-        query_vector = query_vector.reshape(1, -1)
-
-        # 搜索最相似的向量
-        distances, indices = self.index.search(query_vector, top_k)
-        
-        # 获取对应的文件信息
-        db = self.get_db_connection()
-        try:
-            results = []
-            for i, distance in zip(indices[0], distances[0]):
-                if i == -1:  # faiss返回-1表示无效结果
-                    continue
-                    
-                # 将numpy.int64转换为Python int
-                offset = int(i)
-                # 获取文件路径和描述
-                cursor = db.execute("SELECT path, description FROM codebase LIMIT 1 OFFSET ?", (offset,))
-                row = cursor.fetchone()
-                if row:
-                    path, description = row
-                    # 将distance转换为相似度分数（0-1之间）
-                    similarity = 1.0 / (1.0 + float(distance))  # 确保使用Python float
-                    results.append((path, similarity, description))
-            
-            return results
-        finally:
-            db.close()
-
-    def ask_codebase(self, query: str, top_k: int = 5) -> List[Tuple[str, float, str]]:
-        """Ask a question about the codebase"""
-        # 使用搜索函数获取相似文件
-        results = self.search_similar(query, top_k)
-        PrettyOutput.print(f"找到的关联文件: ", output_type=OutputType.INFO)
-        for path, score, _ in results:
-            PrettyOutput.print(f"文件: {path} 关联度: {score:.3f}", output_type=OutputType.INFO)
-        
-        prompt = f"""你是一个代码专家，请根据以下文件信息回答用户的问题：
-"""
-        for path, _, _ in results:
-            content = open(path, "r", encoding="utf-8").read()
-            prompt += f"""
-文件路径: {path}
-文件内容:
-{content}
-========================================
-"""
-        prompt += f"""
-用户问题: {query}
-
-请用专业的语言回答用户的问题，如果给出的文件内容不足以回答用户的问题，请告诉用户，绝对不要胡编乱造。
-"""
-        model = self.platform_registry.create_platform(self.normal_platform)
-        model.set_model_name(self.normal_model)
-        response = model.chat(prompt)
-        return response
-
-
-def main():
-    parser = argparse.ArgumentParser(description='Codebase management and search tool')
-    parser.add_argument('--search', type=str, help='Search query to find similar code files')
-    parser.add_argument('--top-k', type=int, default=5, help='Number of results to return (default: 5)')
-    parser.add_argument('--ask', type=str, help='Ask a question about the codebase')
-    args = parser.parse_args()
-    
-    current_dir = find_git_root()
-    codebase = CodeBase(current_dir)
-
-    try:
-        codebase.generate_codebase()
-        PrettyOutput.print("\nCodebase generation completed", output_type=OutputType.SUCCESS)
-    except Exception as e:
-        PrettyOutput.print(f"Error during codebase generation: {str(e)}", output_type=OutputType.ERROR)
-    
-    if args.search:
-        results = codebase.search_similar(args.search, args.top_k)
-        if not results:
-            PrettyOutput.print("No similar files found", output_type=OutputType.WARNING)
-            return
-            
-        PrettyOutput.print("\nSearch Results:", output_type=OutputType.INFO)
-        for path, score, desc in results:
-            PrettyOutput.print("\n" + "="*50, output_type=OutputType.INFO)
-            PrettyOutput.print(f"File: {path}", output_type=OutputType.INFO)
-            PrettyOutput.print(f"Similarity: {score:.3f}", output_type=OutputType.INFO)
-            PrettyOutput.print(f"Description: {desc[100:]}", output_type=OutputType.INFO)
-
-    if args.ask:            
-        codebase.ask_codebase(args.ask, args.top_k)
-
-
-if __name__ == "__main__":
-    exit(main())

jarvis_ai_assistant-0.1.66/src/jarvis/tools/__init__.py

@@ -1,9 +0,0 @@
-from .registry import ToolRegistry
-from jarvis.tools.codebase_qa import CodebaseQATool
-
-__all__ = [
-    'ToolRegistry',
-] 
-
-def register_tools():
-    register_tool(CodebaseQATool())