proxy for api-tool

-Only for api_tool for now, if this solution works well then implementation for other tools is required
- Need to check api keys creation with the current proxies
- Show connection string example at creation
- locale needs updates for other languages

Author: pabik

application/agents/agent_creator.py

@@ -11,4 +11,7 @@ def create_agent(cls, type, *args, **kwargs):
         agent_class = cls.agents.get(type.lower())
         if not agent_class:
             raise ValueError(f"No agent class found for type {type}")
+        config = kwargs.pop('config', None)
+        if isinstance(config, dict) and 'proxy_id' in config and 'proxy_id' not in kwargs:
+            kwargs['proxy_id'] = config['proxy_id']
         return agent_class(*args, **kwargs)

application/agents/base.py

@@ -17,6 +17,7 @@ def __init__(
         api_key,
         user_api_key=None,
         decoded_token=None,
+        proxy_id=None,
     ):
         self.endpoint = endpoint
         self.llm = LLMCreator.create_llm(
@@ -30,6 +31,7 @@ def __init__(
         self.tools = []
         self.tool_config = {}
         self.tool_calls = []
+        self.proxy_id = proxy_id
 
     def gen(self, *args, **kwargs) -> Generator[Dict, None, None]:
         raise NotImplementedError('Method "gen" must be implemented in the child class')
@@ -41,6 +43,11 @@ def _get_user_tools(self, user="local"):
         user_tools = user_tools_collection.find({"user": user, "status": True})
         user_tools = list(user_tools)
         tools_by_id = {str(tool["_id"]): tool for tool in user_tools}
+        if hasattr(self, 'proxy_id') and self.proxy_id:
+            for tool_id, tool in tools_by_id.items():
+                if 'config' not in tool:
+                    tool['config'] = {}
+                tool['config']['proxy_id'] = self.proxy_id
         return tools_by_id
 
     def _build_tool_parameters(self, action):
@@ -126,6 +133,7 @@ def _execute_tool_action(self, tools_dict, call):
                     "method": tool_data["config"]["actions"][action_name]["method"],
                     "headers": headers,
                     "query_params": query_params,
+                    "proxy_id": self.proxy_id,
                 }
                 if tool_data["name"] == "api_tool"
                 else tool_data["config"]

application/agents/classic_agent.py

@@ -18,9 +18,10 @@ def __init__(
         prompt="",
         chat_history=None,
         decoded_token=None,
+        proxy_id=None,
     ):
         super().__init__(
-            endpoint, llm_name, gpt_model, api_key, user_api_key, decoded_token
+            endpoint, llm_name, gpt_model, api_key, user_api_key, decoded_token, proxy_id
         )
         self.user = decoded_token.get("sub")
         self.prompt = prompt

application/agents/tools/api_tool.py

@@ -23,15 +23,43 @@ def execute_action(self, action_name, **kwargs):
         )
 
     def _make_api_call(self, url, method, headers, query_params, body):
+        sanitized_headers = {}
+        for key, value in headers.items():
+            if isinstance(value, str):
+                sanitized_value = value.encode('latin-1', errors='ignore').decode('latin-1')
+                sanitized_headers[key] = sanitized_value
+            else:
+                sanitized_headers[key] = value
+
         if query_params:
             url = f"{url}?{requests.compat.urlencode(query_params)}"
         if isinstance(body, dict):
             body = json.dumps(body)
+        response = None
         try:
             print(f"Making API call: {method} {url} with body: {body}")
             if body == "{}":
                 body = None
-            response = requests.request(method, url, headers=headers, data=body)
+
+            proxy_id = self.config.get("proxy_id", None)
+            request_kwargs = {
+                'method': method,
+                'url': url,
+                'headers': sanitized_headers,
+                'data': body
+            }
+            try:
+                if proxy_id:
+                    from application.agents.tools.proxy_handler import apply_proxy_to_request
+                    response = apply_proxy_to_request(
+                        requests.request,
+                        proxy_id=proxy_id,
+                        **request_kwargs
+                    )
+                else:
+                    response = requests.request(**request_kwargs)
+            except ImportError:
+                response = requests.request(**request_kwargs)
             response.raise_for_status()
             content_type = response.headers.get(
                 "Content-Type", "application/json"

application/agents/tools/proxy_handler.py

@@ -0,0 +1,63 @@
+import logging
+import requests
+from typing import Dict, Optional
+from bson.objectid import ObjectId
+
+from application.core.mongo_db import MongoDB
+
+logger = logging.getLogger(__name__)
+
+# Get MongoDB connection
+mongo = MongoDB.get_client()
+db = mongo["docsgpt"]
+proxies_collection = db["proxies"]
+
+def get_proxy_config(proxy_id: str) -> Optional[Dict[str, str]]:
+    """
+    Retrieve proxy configuration from the database.
+    
+    Args:
+        proxy_id: The ID of the proxy configuration
+        
+    Returns:
+        A dictionary with proxy configuration or None if not found
+    """
+    if not proxy_id or proxy_id == "none":
+        return None
+        
+    try:
+        if ObjectId.is_valid(proxy_id):
+            proxy_config = proxies_collection.find_one({"_id": ObjectId(proxy_id)})
+            if proxy_config and "connection" in proxy_config:
+                connection_str = proxy_config["connection"].strip()
+                if connection_str:
+                    # Format proxy for requests library
+                    return {
+                        "http": connection_str,
+                        "https": connection_str
+                    }
+        return None
+    except Exception as e:
+        logger.error(f"Error retrieving proxy configuration: {e}")
+        return None
+
+def apply_proxy_to_request(request_func, proxy_id=None, **kwargs):
+    """
+    Apply proxy configuration to a requests function if available.
+    This is a minimal wrapper that doesn't change the function signature.
+    
+    Args:
+        request_func: The requests function to call (e.g., requests.get, requests.post)
+        proxy_id: Optional proxy ID to use
+        **kwargs: Arguments to pass to the request function
+    
+    Returns:
+        The response from the request
+    """
+    if proxy_id:
+        proxy_config = get_proxy_config(proxy_id)
+        if proxy_config:
+            kwargs['proxies'] = proxy_config
+            logger.info(f"Using proxy for request")
+    
+    return request_func(**kwargs)

application/api/answer/routes.py

@@ -335,6 +335,9 @@ class Stream(Resource):
             "prompt_id": fields.String(
                 required=False, default="default", description="Prompt ID"
             ),
+            "proxy_id": fields.String(
+                required=False, description="Proxy ID to use for API calls"
+            ),
             "chunks": fields.Integer(
                 required=False, default=2, description="Number of chunks"
             ),
@@ -376,6 +379,7 @@ def post(self):
             )
             conversation_id = data.get("conversation_id")
             prompt_id = data.get("prompt_id", "default")
+            proxy_id = data.get("proxy_id", None)
 
             index = data.get("index", None)
             chunks = int(data.get("chunks", 2))
@@ -386,6 +390,7 @@ def post(self):
                 data_key = get_data_from_api_key(data["api_key"])
                 chunks = int(data_key.get("chunks", 2))
                 prompt_id = data_key.get("prompt_id", "default")
+                proxy_id = data_key.get("proxy_id", None)
                 source = {"active_docs": data_key.get("source")}
                 retriever_name = data_key.get("retriever", retriever_name)
                 user_api_key = data["api_key"]
@@ -422,6 +427,7 @@ def post(self):
                 api_key=settings.API_KEY,
                 user_api_key=user_api_key,
                 prompt=prompt,
+                proxy_id=proxy_id,
                 chat_history=history,
                 decoded_token=decoded_token,
             )
@@ -496,6 +502,9 @@ class Answer(Resource):
             "prompt_id": fields.String(
                 required=False, default="default", description="Prompt ID"
             ),
+            "proxy_id": fields.String(
+                required=False, description="Proxy ID to use for API calls"
+            ),
             "chunks": fields.Integer(
                 required=False, default=2, description="Number of chunks"
             ),
@@ -527,6 +536,7 @@ def post(self):
             )
             conversation_id = data.get("conversation_id")
             prompt_id = data.get("prompt_id", "default")
+            proxy_id = data.get("proxy_id", None)
             chunks = int(data.get("chunks", 2))
             token_limit = data.get("token_limit", settings.DEFAULT_MAX_HISTORY)
             retriever_name = data.get("retriever", "classic")
@@ -535,6 +545,7 @@ def post(self):
                 data_key = get_data_from_api_key(data["api_key"])
                 chunks = int(data_key.get("chunks", 2))
                 prompt_id = data_key.get("prompt_id", "default")
+                proxy_id = data_key.get("proxy_id", None)
                 source = {"active_docs": data_key.get("source")}
                 retriever_name = data_key.get("retriever", retriever_name)
                 user_api_key = data["api_key"]
@@ -569,6 +580,7 @@ def post(self):
                 api_key=settings.API_KEY,
                 user_api_key=user_api_key,
                 prompt=prompt,
+                proxy_id=proxy_id,
                 chat_history=history,
                 decoded_token=decoded_token,
             )