auto-coder 0.1.291__tar.gz → 0.1.293__tar.gz

{auto_coder-0.1.291 → auto_coder-0.1.293}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: auto-coder
-Version: 0.1.291
+Version: 0.1.293
 Summary: AutoCoder: AutoCoder
 Author: allwefantasy
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
@@ -36,6 +36,7 @@ Requires-Dist: google-generativeai
 Requires-Dist: protobuf
 Requires-Dist: azure-cognitiveservices-speech
 Requires-Dist: real_agent
+Requires-Dist: duckdb
 Requires-Dist: python-docx
 Requires-Dist: docx2txt
 Requires-Dist: pdf2image

{auto_coder-0.1.291 → auto_coder-0.1.293}/src/auto_coder.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: auto-coder
-Version: 0.1.291
+Version: 0.1.293
 Summary: AutoCoder: AutoCoder
 Author: allwefantasy
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
@@ -36,6 +36,7 @@ Requires-Dist: google-generativeai
 Requires-Dist: protobuf
 Requires-Dist: azure-cognitiveservices-speech
 Requires-Dist: real_agent
+Requires-Dist: duckdb
 Requires-Dist: python-docx
 Requires-Dist: docx2txt
 Requires-Dist: pdf2image

{auto_coder-0.1.291 → auto_coder-0.1.293}/src/auto_coder.egg-info/SOURCES.txt

@@ -92,6 +92,7 @@ src/autocoder/common/text.py
 src/autocoder/common/types.py
 src/autocoder/common/utils_code_auto_generate.py
 src/autocoder/common/mcp_servers/__init__.py
+src/autocoder/common/mcp_servers/mcp_server_gpt4o_mini_search.py
 src/autocoder/common/mcp_servers/mcp_server_perplexity.py
 src/autocoder/data/byzerllm.md
 src/autocoder/data/tokenizer.json
@@ -151,6 +152,7 @@ src/autocoder/rag/cache/byzer_storage_cache.py
 src/autocoder/rag/cache/cache_result_merge.py
 src/autocoder/rag/cache/file_monitor_cache.py
 src/autocoder/rag/cache/local_byzer_storage_cache.py
+src/autocoder/rag/cache/local_duckdb_storage_cache.py
 src/autocoder/rag/cache/rag_file_meta.py
 src/autocoder/rag/cache/simple_cache.py
 src/autocoder/rag/loaders/__init__.py

{auto_coder-0.1.291 → auto_coder-0.1.293}/src/auto_coder.egg-info/requires.txt

@@ -26,6 +26,7 @@ google-generativeai
 protobuf
 azure-cognitiveservices-speech
 real_agent
+duckdb
 python-docx
 docx2txt
 pdf2image

{auto_coder-0.1.291 → auto_coder-0.1.293}/src/autocoder/auto_coder_rag.py

@@ -179,6 +179,14 @@ def main(input_args: Optional[List[str]] = None):
     build_index_parser = subparsers.add_parser(
         "build_hybrid_index", help="Build hybrid index for RAG"
     )
+
+    build_index_parser.add_argument(
+        "--rag_storage_type",
+        type=str,
+        default="duckdb",
+        help="The storage type of the RAG, duckdb or byzer-storage",
+    )
+
     build_index_parser.add_argument(
         "--quick", action="store_true", help="Skip system initialization"
     )
@@ -329,6 +337,14 @@ def main(input_args: Optional[List[str]] = None):
         action="store_true",
         help="Enable hybrid index",
     )
+
+    serve_parser.add_argument(
+        "--rag_storage_type",
+        type=str,
+        default="duckdb",
+        help="The storage type of the RAG, duckdb or byzer-storage",
+    )    
+
     serve_parser.add_argument(
         "--hybrid_index_max_output_tokens",
         type=int,

{auto_coder-0.1.291 → auto_coder-0.1.293}/src/autocoder/auto_coder_server.py

@@ -34,6 +34,14 @@ import sys
 import io
 from autocoder.utils.log_capture import LogCapture
 
+# If support dotenv, use it
+if os.path.exists(".env"):
+    try:
+        from dotenv import load_dotenv
+        load_dotenv()
+    except ImportError:
+        pass
+
 def convert_yaml_config_to_str(yaml_config):
     yaml_content = yaml.safe_dump(
         yaml_config,

{auto_coder-0.1.291 → auto_coder-0.1.293}/src/autocoder/common/__init__.py

@@ -292,10 +292,14 @@ class AutoCoderArgs(pydantic.BaseModel):
         
     rag_url: Optional[str] = ""
     rag_token: Optional[str] = ""
-    rag_type: Optional[str] = "storage"   
-    rag_params_max_tokens: Optional[int] = 4096 
-    rag_doc_filter_relevance: Optional[int] = 5
-    rag_context_window_limit: Optional[int] = 120000 
+    rag_type: Optional[str] = "storage"
+    rag_storage_type: Optional[str] = "duckdb"  # 向量化存储类型 byzer-storage | duckdb
+    rag_params_max_tokens: Optional[int] = 500000 
+    rag_doc_filter_relevance: Optional[int] = 2
+    rag_context_window_limit: Optional[int] = 120000
+    rag_duckdb_vector_dim: Optional[int] = 1024  # DuckDB 向量化存储的维度
+    rag_duckdb_query_similarity: Optional[float] = 0.1  # DuckDB 向量化检索 相似度 阈值
+    rag_duckdb_query_top_k: Optional[int] = 10000  # DuckDB 向量化检索 返回 TopK个结果(且大于相似度)
     # rag 本地图床地址
     local_image_host: Optional[str] = ""
     rag_recall_max_queries: Optional[int] = 5

{auto_coder-0.1.291 → auto_coder-0.1.293}/src/autocoder/common/mcp_hub.py

@@ -2,8 +2,12 @@ import os
 import json
 import asyncio
 import aiohttp
+import importlib
+import pkgutil
+import re
+import inspect
 from datetime import datetime, timedelta
-from typing import Dict, List, Optional, Any, Set, Optional
+from typing import Dict, List, Optional, Any, Set, Optional, Tuple
 from pathlib import Path
 from pydantic import BaseModel, Field
 
@@ -62,23 +66,53 @@ class McpConnection:
         self.session = session        
 
 
-MCP_PERPLEXITY_SERVER = '''
-{
-    "perplexity": {
-        "command": "python",
-        "args": [
-            "-m", "autocoder.common.mcp_servers.mcp_server_perplexity"
-        ],
-        "env": {
-            "PERPLEXITY_API_KEY": "{{PERPLEXITY_API_KEY}}"
-        }
-    }
-}
-'''
-
-MCP_BUILD_IN_SERVERS = {
-    "perplexity": json.loads(MCP_PERPLEXITY_SERVER)["perplexity"]
-}
+def _generate_server_configs() -> Tuple[Dict[str, Any], Dict[str, str]]:
+    """
+    Scan the autocoder.common.mcp_servers directory for mcp_server_*.py files
+    and generate server configurations.
+    
+    Returns:
+        Tuple of (built-in servers dict, JSON templates dict)
+    """
+    servers = {}
+    templates = {}
+    
+    try:
+        package_name = "autocoder.common.mcp_servers"
+        package = importlib.import_module(package_name)
+        
+        # Find all modules in the package
+        for _, name, _ in pkgutil.iter_modules(package.__path__, package.__name__ + "."):
+            # Only process modules that start with "mcp_server_"
+            base_name = name.split(".")[-1]
+            if base_name.startswith("mcp_server_"):
+                # Generate a friendly server name
+                friendly_name = base_name[11:]
+                                
+                # Create env dictionary with placeholders
+                env_dict = {}
+                
+                # Create server configuration
+                config = {
+                    "command": "python",
+                    "args": ["-m", name],
+                    "env": env_dict
+                }                                
+                
+                # Store in dictionaries
+                servers[friendly_name] = config
+                templates[friendly_name] = json.dumps({friendly_name: config}, indent=4)
+                
+                logger.info(f"Detected MCP server: {friendly_name}")
+    
+    except Exception as e:
+        logger.error(f"Error generating server configs: {e}")
+    
+    return servers, templates
+
+
+# Automatically generate server configurations
+MCP_BUILD_IN_SERVERS, MCP_SERVER_TEMPLATES = _generate_server_configs()
 
 
 class McpHub:
@@ -422,3 +456,10 @@ class McpHub:
         """
         for name in list(self.connections.keys()):
             await self.delete_connection(name)
+
+    @classmethod
+    def get_server_templates(cls) -> Dict[str, str]:
+        """
+        Get all available server templates as JSON strings
+        """
+        return MCP_SERVER_TEMPLATES

auto_coder-0.1.293/src/autocoder/common/mcp_servers/mcp_server_gpt4o_mini_search.py

@@ -0,0 +1,153 @@
+from os import getenv
+from textwrap import dedent
+import sys
+
+import mcp.server.stdio
+import mcp.types as types
+from mcp.server import NotificationOptions, Server
+from mcp.server.models import InitializationOptions
+import json
+from openai import OpenAI
+
+OPENAI_API_KEY = getenv("OPENAI_API_KEY")
+# Check if API key is empty or None
+if not OPENAI_API_KEY:
+    print("Error: OPENAI_API_KEY environment variable is not set. Please set it before running this server.", file=sys.stderr)
+    sys.exit(1)
+
+OPENAI_API_BASE_URL = getenv(
+    "OPENAI_API_BASE_URL", "https://api.openai.com/v1")
+
+server = Server("mcp-server-gpt4o-mini-search")
+
+client = OpenAI(
+    api_key=OPENAI_API_KEY,
+    base_url=OPENAI_API_BASE_URL
+)
+
+
+@server.list_tools()
+async def handle_list_tools() -> list[types.Tool]:
+    return [
+        types.Tool(
+            name="gpt4o_mini_search",
+            description=dedent(
+                """
+                GPT-4o mini with search enables agents to gather information from the internet 
+                in real-time, providing up-to-date answers with source citations.
+                This tool is ideal for fact-checking, research, and accessing current information
+                that might not be in the model's training data.
+                
+                The search-enhanced responses include relevant web sources to support the information
+                provided, making it useful for obtaining verified and recent information.
+                
+                [Response structure]
+                - id: A unique identifier for the response
+                - model: The model used (gpt-4o-mini-search-preview)
+                - object: The object type ("chat.completion")
+                - created: The Unix timestamp when the completion was created
+                - choices[]: The list of completion choices generated
+                - usage: Usage statistics for the completion request
+                """
+            ),
+            inputSchema={
+                "type": "object",
+                "properties": {
+                    "system_message": {
+                        "type": "string",
+                        "description": "Optional custom system message. If not provided, a default search-optimized system message will be used.",
+                    },
+                    "messages": {
+                        "type": "array",
+                        "description": "A list of messages comprising the conversation so far (excluding system message which is handled separately).",
+                        "items": {
+                            "type": "object",
+                            "properties": {
+                                "content": {
+                                    "type": "string",
+                                    "description": "The contents of the message in this turn of conversation.",
+                                },
+                                "role": {
+                                    "type": "string",
+                                    "description": "The role of the speaker in this turn of conversation.",
+                                    "enum": ["user", "assistant"],
+                                },
+                            },
+                            "required": ["content", "role"],
+                        },
+                    },
+                },
+                "required": ["messages"],
+            },
+        )
+    ]
+
+
+@server.call_tool()
+async def handle_call_tool(
+    name: str, arguments: dict
+) -> list[types.TextContent | types.ImageContent | types.EmbeddedResource]:
+    if name != "gpt4o_mini_search":
+        raise ValueError(f"Unknown tool: {name}")
+
+    # Extract user messages
+    user_messages = arguments.get("messages", [])
+
+    # Define default system message if not provided
+    default_system_message = (
+        "你是专业搜索助手，需要：\n"
+        "1. 提供基于用户查询的清晰格式化信息\n"
+        "2. 使用[标题](URL)格式嵌入链接\n"
+        "3. 每条信息后附上来源\n"
+        "4. 用'---'分隔不同结果\n"
+        "5. 直接在文本中引用，不使用编号引用\n"
+        "6. 确保提供完整URL"
+    )
+
+    # Use custom system message if provided, otherwise use default
+    system_message = arguments.get("system_message", default_system_message)
+
+    # Prepare full message list with system message first
+    full_messages = [{"role": "system", "content": system_message}]
+    full_messages.extend(user_messages)
+
+    try:
+        # Initialize OpenAI client
+
+        # Make the API call using OpenAI SDK
+        completion = client.chat.completions.create(
+            model="gpt-4o-mini-search-preview",
+            messages=full_messages
+        )
+
+        # Extract content from response
+        content = completion.choices[0].message.content
+
+    except Exception as e:
+        raise RuntimeError(f"API error: {str(e)}")
+
+    return [types.TextContent(
+            type="text",
+            text=content,
+            )]
+
+
+async def main():
+    async with mcp.server.stdio.stdio_server() as (read_stream, write_stream):
+        await server.run(
+            read_stream,
+            write_stream,
+            InitializationOptions(
+                server_name="mcp-server-gpt4o-mini-search",
+                server_version="0.1.0",
+                capabilities=server.get_capabilities(
+                    notification_options=NotificationOptions(
+                        tools_changed=True),
+                    experimental_capabilities={},
+                ),
+            ),
+        )
+
+if __name__ == "__main__":
+    import asyncio
+    asyncio.run(main())

{auto_coder-0.1.291 → auto_coder-0.1.293}/src/autocoder/rag/api_server.py

@@ -31,6 +31,14 @@ from byzerllm.utils.client.entrypoints.openai.protocol import (
 from pydantic import BaseModel
 from typing import List,Optional
 
+# If support dotenv, use it
+if os.path.exists(".env"):
+    try:
+        from dotenv import load_dotenv
+        load_dotenv()
+    except ImportError:
+        pass
+
 logger = init_logger(__name__)
 
 llm_client: ByzerLLM = None