Google colab workbook example

deployed-agents/Google Colab Workbook/Getting_started_with_Corvic.ipynb

@@ -0,0 +1,359 @@
+{
+ "nbformat": 4,
+ "nbformat_minor": 0,
+ "metadata": {
+  "colab": {
+   "provenance": []
+  },
+  "kernelspec": {
+   "name": "python3",
+   "display_name": "Python 3"
+  },
+  "language_info": {
+   "name": "python"
+  }
+ },
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "source": [
+    "**Getting Started with Corvic using Colab**\n",
+    "In this example, we will accomplish a simple query to Corvic. This example is similar to the one mentioned [here](https://app.corvic.ai/docs/mcpIntegrations/openaiMcpWithCorvic) wherein we enquire about video game sales\n",
+    "\n",
+    "The following two steps are involved:\n",
+    "<p>\n",
+    "STEP 1: Invoke Corvic agent using the Model Context Protocol and ask a question about video game sales\n",
+    "</p>\n",
+    "<p>\n",
+    "STEP 2: (Optional) Parse the markdown and separate out the text and image data\n",
+    "</p>"
+   ],
+   "metadata": {
+    "id": "XbHo_dz8weuI"
+   }
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "\n",
+    "\n",
+    "---\n",
+    "\n"
+   ],
+   "metadata": {
+    "id": "rIKZ3cASxrq_"
+   }
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "STEP 1: Invoke Corvic agent using the Model Context Protocol and ask a question about video game sales\n",
+    "<p/>\n",
+    "Please note:\n",
+    "The MCP library is used for corvic agent invocation. Corvic returns markdown text. We will convert markdown and beautiful soup libraries to parse text and images"
+   ],
+   "metadata": {
+    "id": "rgYEqIKixUmr"
+   }
+  },
+  {
+   "cell_type": "code",
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "XObia6wuZ6xe",
+    "outputId": "c3bbdc6f-3f9c-4989-c11f-27da08467500"
+   },
+   "source": [
+    "# mcp is used for corvic agent invocation\n",
+    "!pip install mcp\n",
+    "!pip install beautifulsoup4\n",
+    "!pip install markdown"
+   ],
+   "outputs": [],
+   "execution_count": null
+  },
+  {
+   "cell_type": "code",
+   "source": [
+    "import asyncio\n",
+    "from mcp import ClientSession\n",
+    "from mcp.client.sse import sse_client\n"
+   ],
+   "metadata": {
+    "id": "-lsdUACMaXRy"
+   },
+   "outputs": [],
+   "execution_count": null
+  },
+  {
+   "cell_type": "code",
+   "source": [
+    "url=\"YOUR CORVIC MCP ENDPOINT\"\n",
+    "headers={\n",
+    "          \"Authorization\": \"YOUR CORVIC API TOKEN\"\n",
+    "        }\n"
+   ],
+   "metadata": {
+    "id": "nkQV-Ch_yJjr"
+   },
+   "outputs": [],
+   "execution_count": null
+  },
+  {
+   "cell_type": "code",
+   "source": [
+    "async def run():\n",
+    "    async with sse_client(\n",
+    "            url,\n",
+    "            headers=headers,\n",
+    "    ) as (read, write):\n",
+    "        async with ClientSession(read, write) as session:\n",
+    "            # Initialize the connection\n",
+    "            await session.initialize()\n",
+    "\n",
+    "            # List available tools\n",
+    "            tools = await session.list_tools()\n",
+    "            print(tools)\n",
+    "\n",
+    "            # Call a tool\n",
+    "            result = await session.call_tool(\n",
+    "                \"query\", arguments={\n",
+    "                    \"query_content\":\n",
+    "                    \"Group all the data by name and find the top titles by \"\n",
+    "                    \"global sales. Output the name and the total global sales \"\n",
+    "                    \" in a html tabular format.\"\n",
+    "                    \"Provide a bar chart for the same table with the game title\"\n",
+    "                    \" on X axis and global sales on Y axis\"}\n",
+    "            )\n",
+    "            return result\n",
+    "\n"
+   ],
+   "metadata": {
+    "id": "g9gQh1GbagQ8"
+   },
+   "outputs": [],
+   "execution_count": null
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "The following line will invoke corvic and collect the markdown response in the corvic_response field"
+   ],
+   "metadata": {
+    "id": "0UWY8Vg7yfIG"
+   }
+  },
+  {
+   "cell_type": "code",
+   "source": [
+    "corvic_response=await run()"
+   ],
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "zgiYun6jasQf",
+    "outputId": "d4f13259-1d02-4b86-a900-0dddc35f9c3c"
+   },
+   "outputs": [],
+   "execution_count": null
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "You can write the Corvic response directly in an MD file and view it"
+   ],
+   "metadata": {
+    "id": "x_c1BBSy3P0j"
+   }
+  },
+  {
+   "cell_type": "code",
+   "source": [
+    "with open(\"vg_query.md\", \"a\") as file:\n",
+    "  for content in corvic_response.content:\n",
+    "      file.write(content.text)"
+   ],
+   "metadata": {
+    "id": "0-W9kSys3ZdN"
+   },
+   "outputs": [],
+   "execution_count": null
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "You can directly view the markdown output as below."
+   ],
+   "metadata": {
+    "id": "gipmqzY64Iyi"
+   }
+  },
+  {
+   "metadata": {},
+   "cell_type": "markdown",
+   "source": ""
+  },
+  {
+   "cell_type": "code",
+   "source": [
+    "from IPython.display import Markdown, display\n",
+    "display(Markdown(corvic_response.content[0].text))"
+   ],
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/",
+     "height": 1000
+    },
+    "id": "zd96XjmO3lMP",
+    "outputId": "5a6f7044-70a9-4204-82ea-d642f1ed0a4d"
+   },
+   "outputs": [],
+   "execution_count": null
+  },
+  {
+   "metadata": {},
+   "cell_type": "markdown",
+   "source": [
+    "This completes Step 1\n",
+    "\n",
+    "---"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "STEP 2: Parse the markdown and display the text and images separately"
+   ],
+   "metadata": {
+    "id": "DcFN8jLTzSoX"
+   }
+  },
+  {
+   "cell_type": "code",
+   "source": [
+    "from types import NoneType\n",
+    "from bs4 import BeautifulSoup\n",
+    "import markdown\n",
+    "import base64\n",
+    "\n",
+    "\n",
+    "html_content = markdown.markdown(corvic_response.content[0].text)\n",
+    "html_view = BeautifulSoup(\"<html>\"+html_content+\"</html>\", features=\"html.parser\")\n",
+    "\n",
+    "html_node = html_view.find('html')\n",
+    "#Find images and strip from original html\n",
+    "image_data=[]\n",
+    "text_output=''\n",
+    "for c1 in html_node.children:\n",
+    "  # print(str(type(c1)))\n",
+    "  if 'bs4.element.NavigableString' not in str(type(c1)):\n",
+    "    if len(list(c1.children)) > 0:\n",
+    "      for c2 in c1.children:\n",
+    "        # print(str(type(c2)) + \" \" + str(c2.name))\n",
+    "        if \"img\" == str(c2.name):\n",
+    "          b64encoded_img_string=c2.attrs['src'].replace('data:image/png;base64,', '')\n",
+    "          image_data.append(base64.b64decode(b64encoded_img_string))\n",
+    "          c2.decompose()\n",
+    "        else:\n",
+    "          # print(str(type(c2)) )\n",
+    "          if 'NoneType' not in str(type(c2)):\n",
+    "            text_output+=repr(c2)\n",
+    "  else:\n",
+    "    text_output+=repr(c1)"
+   ],
+   "metadata": {
+    "id": "LcQmN2unnod3"
+   },
+   "outputs": [],
+   "execution_count": null
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "Here is the text output in a html format for better readability"
+   ],
+   "metadata": {
+    "id": "w5BOH3Dv4yS1"
+   }
+  },
+  {
+   "cell_type": "code",
+   "source": [
+    "from IPython.core.display import display, HTML\n",
+    "\n",
+    "display(HTML(html_node.prettify()))"
+   ],
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/",
+     "height": 372
+    },
+    "id": "Q6EVovBvp9JK",
+    "outputId": "2d23bdde-e67c-4c5b-fefa-da9d0a859b1b"
+   },
+   "outputs": [],
+   "execution_count": null
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "Here is the first image from the Corvic response"
+   ],
+   "metadata": {
+    "id": "7oJEprsv5O3f"
+   }
+  },
+  {
+   "cell_type": "code",
+   "source": [
+    "from IPython import display\n",
+    "\n",
+    "display.Image(image_data[0])"
+   ],
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/",
+     "height": 422
+    },
+    "id": "2VblEuxLoxdp",
+    "outputId": "de8a6ede-37fd-4f67-cc94-28d49064cbc4"
+   },
+   "outputs": [],
+   "execution_count": null
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "\n",
+    "\n",
+    "---\n",
+    "\n"
+   ],
+   "metadata": {
+    "id": "JNfQ9V8W4tKC"
+   }
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "Summary: You have been able to invoke the Corvic MCP from this Colab workbook.\n",
+    "You also explored the response in different ways allowing you great flexibility in using the output in your AI pipelines"
+   ],
+   "metadata": {
+    "id": "jg7vEXZn4YQ5"
+   }
+  },
+  {
+   "cell_type": "code",
+   "source": [],
+   "metadata": {
+    "id": "EqEklB0q5IZZ"
+   },
+   "outputs": [],
+   "execution_count": null
+  }
+ ]
+}

deployed-agents/Google Colab Workbook/README.md

@@ -0,0 +1,32 @@
+# 🧪 Google Colab Workbook for interacting with Corvic Agents via MCP
+
+This example demonstrates how to work with a deployed Corvic agent via MCP
+We also explore how to view, and parse text and image outputs for downstream AI processing
+---
+
+## 📘 Use Case
+
+You want to explore Corvic Agent in an interactive Google Colab Environment
+You want to view different parts of the response, including text and images
+---
+
+## ✅ Steps
+
+**Pre-requisites**
+This example is similar to the one mentioned [here](https://app.corvic.ai/docs/mcpIntegrations/openaiMcpWithCorvic) 
+wherein we enquire about video game sales. Please complete the pre-requisites
+
+1. **Configure Corvic Agent Endpoint**
+   - Set `MCP_URL` to your deployed Corvic agent's endpoint.
+   - Set the `HEADERS` with your Corvic API token.
+
+2. **Understand the query Invoke the Corvic Agent via MCP**
+   - We invoke a question about video game sales and ask the agent to provide a HTML table and bar chart
+
+3. **Visualize the responses**
+   - We view the response from Corvic via a markdown viewer
+   - We parse text and image separately and view them
+
+---
+
+Need help? Contact [support@corvic.ai](mailto:support@corvic.ai) or visit [https://www.corvic.ai](https://www.corvic.ai).