samples/notebooks/dotnet/05-using-the-planner.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Introduction to the Planner\n",
    "\n",
    "The Planner is one of the fundamental concepts of the Semantic Kernel. It makes use of the collection of skills that have been registered to the kernel and using AI, will formulate a plan to execute a given ask.\n",
    "\n",
    "Read more about it [here](https://aka.ms/sk/concepts/planner)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "dotnet_interactive": {
     "language": "csharp"
    },
    "polyglot_notebook": {
     "kernelName": "csharp"
    }
   },
   "outputs": [],
   "source": [
    "#r \"nuget: Microsoft.SemanticKernel, 0.15.230531.5-preview\"\n",
    "\n",
    "#!import config/Settings.cs\n",
    "#!import config/Utils.cs\n",
    "\n",
    "using Microsoft.SemanticKernel;\n",
    "using Microsoft.SemanticKernel.CoreSkills;\n",
    "using Microsoft.SemanticKernel.Orchestration;\n",
    "using Microsoft.SemanticKernel.Planning;\n",
    "using Microsoft.SemanticKernel.Planning.Sequential;\n",
    "\n",
    "var builder = new KernelBuilder();\n",
    "\n",
    "// Configure AI backend used by the kernel\n",
    "var (useAzureOpenAI, model, azureEndpoint, apiKey, orgId) = Settings.LoadFromFile();\n",
    "\n",
    "if (useAzureOpenAI)\n",
    "    builder.WithAzureTextCompletionService(model, azureEndpoint, apiKey);\n",
    "else\n",
    "    builder.WithOpenAITextCompletionService(model, apiKey, orgId);\n",
    "\n",
    "var kernel = builder.Build();"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Setting Up the Planner\n",
    "The planner is located in the Semantic Kernel's CoreSkills and requires Orchestration"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "dotnet_interactive": {
     "language": "csharp"
    },
    "polyglot_notebook": {
     "kernelName": "csharp"
    }
   },
   "outputs": [],
   "source": [
    "// Load native skill into the kernel registry, sharing its functions with prompt templates\n",
    "var planner = new SequentialPlanner(kernel);"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Providing skills to the planner\n",
    "The planner needs to know what skills are available to it. Here we'll give it access to the `SummarizeSkill` and `WriterSkill` we have defined on disk."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "dotnet_interactive": {
     "language": "csharp"
    },
    "polyglot_notebook": {
     "kernelName": "csharp"
    }
   },
   "outputs": [],
   "source": [
    "var skillsDirectory = Path.Combine(System.IO.Directory.GetCurrentDirectory(), \"..\", \"..\", \"skills\");\n",
    "kernel.ImportSemanticSkillFromDirectory(skillsDirectory, \"SummarizeSkill\");\n",
    "kernel.ImportSemanticSkillFromDirectory(skillsDirectory, \"WriterSkill\");"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Define your ASK. What do you want the Kernel to do?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "dotnet_interactive": {
     "language": "csharp"
    },
    "polyglot_notebook": {
     "kernelName": "csharp"
    }
   },
   "outputs": [],
   "source": [
    "var ask = \"Tomorrow is Valentine's day. I need to come up with a few date ideas and e-mail them to my significant other.\";\n",
    "var originalPlan = await planner.CreatePlanAsync(ask);\n",
    "\n",
    "Console.WriteLine(\"Original plan:\\n\");\n",
    "Console.WriteLine(JsonSerializer.Serialize(originalPlan, new JsonSerializerOptions { WriteIndented = true }));"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As you can see in the above plan, the Planner has taken the user's ask and converted it into a Plan object detailing how the AI would go about solving this task.\n",
    "\n",
    "It makes use of the skills that the Kernel has available to it and determines which functions to call in order to fulfill the user's ask.\n",
    "\n",
    "The output of each step of the plan gets set as `setContextVariable` which makes it available as `input` to the next skill."
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's also define an inline skill and have it be available to the Planner.\n",
    "Be sure to give it a function name and skill name."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "dotnet_interactive": {
     "language": "csharp"
    },
    "polyglot_notebook": {
     "kernelName": "csharp"
    }
   },
   "outputs": [],
   "source": [
    "string skPrompt = \"\"\"\n",
    "{{$input}}\n",
    "\n",
    "Rewrite the above in the style of Shakespeare.\n",
    "\"\"\";\n",
    "var shakespeareFunction = kernel.CreateSemanticFunction(skPrompt, \"shakespeare\", \"ShakespeareSkill\", maxTokens: 2000, temperature: 0.2, topP: 0.5);"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's update our ask using this new skill."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "dotnet_interactive": {
     "language": "csharp"
    },
    "polyglot_notebook": {
     "kernelName": "csharp"
    }
   },
   "outputs": [],
   "source": [
    "var ask = @\"Tomorrow is Valentine's day. I need to come up with a few date ideas.\n",
    "She likes Shakespeare so write using his style. E-mail these ideas to my significant other\";\n",
    "\n",
    "var newPlan = await planner.CreatePlanAsync(ask);\n",
    "\n",
    "Console.WriteLine(\"Updated plan:\\n\");\n",
    "Console.WriteLine(JsonSerializer.Serialize(newPlan, new JsonSerializerOptions { WriteIndented = true }));"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Executing the plans\n",
    "\n",
    "Now that we have different plans, let's try to execute them! The Kernel can execute the plan using RunAsync."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "dotnet_interactive": {
     "language": "csharp"
    },
    "polyglot_notebook": {
     "kernelName": "csharp"
    }
   },
   "outputs": [],
   "source": [
    "var originalPlanResult = await originalPlan.InvokeAsync();\n",
    "\n",
    "Console.WriteLine(\"Original Plan results:\\n\");\n",
    "Console.WriteLine(Utils.WordWrap(originalPlanResult.Result, 100));"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now lets execute and print the new plan:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "dotnet_interactive": {
     "language": "csharp"
    },
    "polyglot_notebook": {
     "kernelName": "csharp"
    }
   },
   "outputs": [],
   "source": [
    "var newPlanResult = await kernel.RunAsync(newPlan);\n",
    "Console.WriteLine(\"New Plan results:\\n\");\n",
    "Console.WriteLine(newPlanResult.Result);"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": ".NET (C#)",
   "language": "C#",
   "name": ".net-csharp"
  },
  "language_info": {
   "name": "polyglot-notebook"
  },
  "polyglot_notebook": {
   "kernelInfo": {
    "defaultKernelName": "csharp",
    "items": [
     {
      "aliases": [],
      "name": "csharp"
     }
    ]
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}