Merge pull request #150 from hchen2020/master

RoutingSettings.Provider and Model.

Author: Oceania2018

docs/agent/router.md

@@ -0,0 +1,11 @@
+# Router
+
+This section will explain in detail the usage of Router. Router has a dedicated configuration node for customization.
+
+```json
+"Router": {
+  "RouterId": "",
+  "Provider": "azure-openai",
+  "Model": "gpt-4"
+}
+```

docs/architecture/routing.md

@@ -1,3 +1,16 @@
 # Routing
 
-Routing is an important function that allows multiple Agents to work together to complete enterprise-level tasks. When you apply LLM to your business system, it is inevitable to develop agents with different functions. How to effectively manage so many agents with different responsibilities is a challenging task. From an engineering perspective, each Different teams are responsible for the development of different Agents, and the teams will face the problem of mutual isolation between Agents. The ideal situation is that they can be developed independently but cooperate with each other.
+Routing is an important function that allows multiple Agents to work together to complete enterprise-level tasks. When you apply LLM to your business system, it is inevitable to develop agents with different functions. How to effectively manage so many agents with different responsibilities is a challenging task. From an engineering perspective, each Different teams are responsible for the development of different Agents, and the teams will face the problem of mutual isolation between Agents. The ideal situation is that they can be developed independently but cooperate with each other.
+
+## Router
+
+The Routing feature is the core technology used by BotSharp to manage multiple Agents. BotSharp has a built-in intelligent Agent called `Router`. When you enable this function, all user requests will be pre-processed by the Router to determine which Agent to distribute the request to for processing. The advantage of Routing technology is that it can isolate different Agents and allow them to work together to achieve the user's goals. The adoption of `Routing` ensures that Agent can be scalable, flexible and robust enough in enterprise applications.
+
+## Reasoner
+
+For simple questions raised by users, the ordinary routing function can already handle it. However, for the scenario where the user has a long description and needs to disassemble the task, ordinary routing cannot handle it. At this time, the `Reasoning` feature needs to be turned on, and LLM will respond according to the problem. The complexity is broken down into different small tasks. These small tasks can be processed by the corresponding Agent. During the processing process, the Router will constantly adjust the next step plan to deal with the different results returned by the Agent.
+
+
+### How to register agent to router?
+
+When you add a new Agent, the Router can automatically read the Agent's configuration, but in order for the Router to distribute the Request to the new Agent, you must set the `AllowRouting` attribute to `True`. For more information on how to use Router, please refer to the Agent/Router chapter.

docs/index.rst

@@ -52,6 +52,7 @@ The main documentation for the site is organized into the following sections:
    agent/intro
    agent/conversation
    agent/state
+   agent/router
 
 .. _integration-docs:
 

src/Infrastructure/BotSharp.Abstraction/Conversations/Models/IncomingMessageModel.cs

@@ -1,5 +1,3 @@
-using System.Text.Json.Serialization;
-
 namespace BotSharp.Abstraction.Conversations.Models;
 
 public class IncomingMessageModel

src/Infrastructure/BotSharp.Abstraction/Functions/Models/FunctionCallFromLlm.cs

@@ -1,6 +1,4 @@
 using BotSharp.Abstraction.Routing.Models;
-using System.Text.Json;
-using System.Text.Json.Serialization;
 
 namespace BotSharp.Abstraction.Functions.Models;
 

src/Infrastructure/BotSharp.Abstraction/Routing/Models/RetrievalArgs.cs

@@ -1,5 +1,4 @@
 using System.Text.Json;
-using System.Text.Json.Serialization;
 
 namespace BotSharp.Abstraction.Routing.Models;
 
@@ -19,6 +18,6 @@ public class RetrievalArgs : RoutingArgs
 
     public override string ToString()
     {
-        return $" [{AgentName}]: {Question} ({JsonSerializer.Serialize(Arguments)}) => {Answer} ({Reason})";
+        return $"[{AgentName}, {Reason}]: ({JsonSerializer.Serialize(Arguments)}) {Question}";
     }
 }

src/Infrastructure/BotSharp.Abstraction/Routing/Settings/RoutingSettings.cs

@@ -8,4 +8,8 @@ public class RoutingSettings
     public string RouterId { get; set; } = string.Empty;
 
     public bool EnableReasoning {  get; set; } = false;
+
+    public string Provider {  get; set; } = string.Empty;
+
+    public string Model { get; set; } = string.Empty;
 }

src/Infrastructure/BotSharp.Core/Conversations/Services/ConversationStateService.cs

@@ -126,7 +126,7 @@ public string GetState(string name, string defaultValue = "")
             _states[name] = defaultValue ?? "";
         }
 
-        if (_states[name] == null)
+        if (string.IsNullOrEmpty(_states[name]))
         {
             return defaultValue;
         }

src/Infrastructure/BotSharp.Core/Infrastructures/CompletionProvider.cs

@@ -10,12 +10,12 @@ public static IChatCompletion GetChatCompletion(IServiceProvider services, strin
 
         var state = services.GetRequiredService<IConversationStateService>();
 
-        if (provider == null)
+        if (string.IsNullOrEmpty(provider))
         {
             provider = state.GetState("provider", "azure-openai");
         }
 
-        if (model == null)
+        if (string.IsNullOrEmpty(model))
         {
             model = state.GetState("model", "gpt-3.5-turbo");
         }

src/Infrastructure/BotSharp.Core/Routing/Prompts/router_prompt.liquid

@@ -11,6 +11,7 @@ You're a Agent Router with reasoning, you can dispatch request to different agen
 Route request to appropriate agent.
 Parameters:
   1. agent_name: the name of the agent;
+  2. reason: why route to this agent;
 
 # task_end
 Call this function when current task is completed.