Upload official_microsoft_graphrag_colab.ipynb with huggingface_hub

official_microsoft_graphrag_colab.ipynb

@@ -0,0 +1,224 @@
+{
+  "cells": [
+    {
+      "cell_type": "markdown",
+      "metadata": {},
+      "source": [
+        "# \ud83c\udfe2 Official Microsoft GraphRAG (MSR) on Google Colab\n",
+        "### End-to-End Pipeline: Entity Extraction \u2794 Leiden Community Partitioning \u2794 Local & Global Search\n",
+        "\n",
+        "[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/Prannesshkva/qdb-ai-benchmarks/blob/main/official_microsoft_graphrag_colab.ipynb)\n",
+        "\n",
+        "This notebook runs the **official Microsoft Research `graphrag` CLI pipeline** against OpenAI's API, testing both:\n",
+        "1. **Official Microsoft GraphRAG (Local Search & Global Search)**\n",
+        "2. **QDB Deductive Engine (`qdb-ai`)** on the exact same multi-hop challenge."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {},
+      "outputs": [],
+      "source": [
+        "# [1] Install Official Microsoft GraphRAG and QDB\n",
+        "!pip install graphrag qdb-ai==2.1.1 pandas -q\n",
+        "print(\"\u2705 Microsoft GraphRAG & QDB installed successfully!\")"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {},
+      "outputs": [],
+      "source": [
+        "# [2] Configure OpenAI API Key\n",
+        "import os, getpass\n",
+        "\n",
+        "# Enter your OpenAI API Key (input is masked for security)\n",
+        "api_key = getpass.getpass(\"\ud83d\udd11 Enter your OpenAI API Key (sk-...): \")\n",
+        "os.environ[\"OPENAI_API_KEY\"] = api_key.strip()\n",
+        "os.environ[\"GRAPHRAG_API_KEY\"] = api_key.strip()\n",
+        "\n",
+        "print(\"\u2705 OpenAI API Key loaded successfully!\")"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {},
+      "outputs": [],
+      "source": [
+        "# [3] Initialize Microsoft GraphRAG Workspace & settings.yaml\n",
+        "import shutil\n",
+        "\n",
+        "project_dir = \"./graphrag_project\"\n",
+        "input_dir = f\"{project_dir}/input\"\n",
+        "\n",
+        "if os.path.exists(project_dir):\n",
+        "    shutil.rmtree(project_dir)\n",
+        "os.makedirs(input_dir, exist_ok=True)\n",
+        "\n",
+        "# Write Microsoft GraphRAG settings.yaml\n",
+        "settings_yaml = f\"\"\"\n",
+        "encoding_model: cl100k_base\n",
+        "skip_workflows: []\n",
+        "\n",
+        "llm:\n",
+        "  api_key: \"{os.environ['OPENAI_API_KEY']}\"\n",
+        "  type: openai_chat\n",
+        "  model: gpt-4o-mini\n",
+        "  model_supports_json: true\n",
+        "  max_tokens: 4000\n",
+        "  temperature: 0.0\n",
+        "\n",
+        "embeddings:\n",
+        "  llm:\n",
+        "    api_key: \"{os.environ['OPENAI_API_KEY']}\"\n",
+        "    type: openai_embedding\n",
+        "    model: text-embedding-3-small\n",
+        "\n",
+        "chunks:\n",
+        "  size: 300\n",
+        "  overlap: 50\n",
+        "  group_by_columns: [id]\n",
+        "\n",
+        "input:\n",
+        "  type: file\n",
+        "  file_type: text\n",
+        "  base_dir: \"input\"\n",
+        "  file_pattern: \".*\\\\.txt$\"\n",
+        "\n",
+        "entity_extraction:\n",
+        "  max_gleanings: 1\n",
+        "\n",
+        "community_reports:\n",
+        "  max_length: 2000\n",
+        "  max_input_length: 8000\n",
+        "\n",
+        "snapshots:\n",
+        "  graphml: false\n",
+        "  embeddings: false\n",
+        "  transient: false\n",
+        "\"\"\"\n",
+        "\n",
+        "with open(f\"{project_dir}/settings.yaml\", \"w\", encoding=\"utf-8\") as f:\n",
+        "    f.write(settings_yaml.strip())\n",
+        "\n",
+        "print(\"\u2705 Microsoft GraphRAG settings.yaml initialized!\")"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {},
+      "outputs": [],
+      "source": [
+        "# [4] Write Benchmark Knowledge Base to input/data.txt\n",
+        "corpus_text = \"\"\"Nexus Dynamics engineered the Chronos Sensor Array in Cambridge during fiscal year 2021.\n",
+        "The Chronos Sensor Array utilizes sub-atomic resonance crystals manufactured by Aether Labs.\n",
+        "Aether Labs merged into Hyperion Aerospace during the international 2022 Geneva Summit.\n",
+        "Hyperion Aerospace contracted Project Valkyrie to deploy the orbital quantum transceiver network.\n",
+        "Project Valkyrie established its primary operational ground station in the Almaty facility, Kazakhstan.\n",
+        "The Almaty facility is directed by Dr. Elena Rostov who holds the master telemetry decryption key.\n",
+        "\n",
+        "In 2019, Hyperion Aerospace was headquartered in Berlin and maintained zero active contracts with Project Valkyrie.\n",
+        "Dr. Elena Rostov resigned from the Moscow Space Observatory in 2018 and has no telemetry access.\n",
+        "The Chronos Sensor Array was entirely decommissioned and destroyed in 2017 before deployment.\n",
+        "\n",
+        "Cambridge University published research on high-frequency resonance sensors in 2021.\n",
+        "Geneva hosts international aerospace summits annually to regulate satellite frequencies.\n",
+        "Kazakhstan operates several commercial communication relays across Central Asia.\n",
+        "Dr. Elena Rostov authored a textbook on orbital mechanics published by Springer in 2015.\n",
+        "\"\"\"\n",
+        "\n",
+        "with open(f\"{input_dir}/data.txt\", \"w\", encoding=\"utf-8\") as f:\n",
+        "    f.write(corpus_text)\n",
+        "\n",
+        "print(\"\u2705 Benchmark corpus written to ./graphrag_project/input/data.txt\")"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {},
+      "outputs": [],
+      "source": [
+        "# [5] Execute Official Microsoft GraphRAG Indexing Pipeline\n",
+        "print(\"\ud83d\ude80 Starting Microsoft GraphRAG Indexing (Calling OpenAI API)...\")\n",
+        "!python -m graphrag.index --root ./graphrag_project\n",
+        "print(\"\ud83c\udf89 Indexing complete! Generated parquet knowledge graph tables in ./graphrag_project/output/\")"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {},
+      "outputs": [],
+      "source": [
+        "# [6] Query Official Microsoft GraphRAG: LOCAL SEARCH (Community Subgraph)\n",
+        "query = \"Who holds the master telemetry decryption key for the sensor array designed by Nexus Dynamics, and where is the facility located?\"\n",
+        "\n",
+        "print(\"=\" * 80)\n",
+        "print(\"\ud83d\udd0d RUNNING OFFICIAL MICROSOFT GRAPHRAG (LOCAL SEARCH):\")\n",
+        "print(\"=\" * 80)\n",
+        "!python -m graphrag.query --root ./graphrag_project --method local --query \"Who holds the master telemetry decryption key for the sensor array designed by Nexus Dynamics, and where is the facility located?\""
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {},
+      "outputs": [],
+      "source": [
+        "# [7] Query Official Microsoft GraphRAG: GLOBAL SEARCH (Community Summaries)\n",
+        "print(\"=\" * 80)\n",
+        "print(\"\ud83c\udf10 RUNNING OFFICIAL MICROSOFT GRAPHRAG (GLOBAL SEARCH):\")\n",
+        "print(\"=\" * 80)\n",
+        "!python -m graphrag.query --root ./graphrag_project --method global --query \"Who holds the master telemetry decryption key for the sensor array designed by Nexus Dynamics, and where is the facility located?\""
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {},
+      "outputs": [],
+      "source": [
+        "# [8] Run QDB Deductive Engine on the Exact Same Challenge\n",
+        "from qdb import Vault\n",
+        "\n",
+        "vault = Vault(\"official_comparison_vault\", purge=True, embedder=\"fast\")\n",
+        "for line in corpus_text.strip().split(\"\\n\"):\n",
+        "    if len(line.strip()) > 10:\n",
+        "        vault.ingest(line.strip())\n",
+        "\n",
+        "print(\"=\" * 80)\n",
+        "print(\"\u269b\ufe0f RUNNING QDB DEDUCTIVE ENGINE (DISCRETE QCBO HAMILTONIAN + SQA):\")\n",
+        "print(\"=\" * 80)\n",
+        "qdb_answer = vault.ask(query, hops=6, budget=6, solver=\"auto\")\n",
+        "print(qdb_answer)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {},
+      "source": [
+        "## \ud83d\udcca Summary Comparison:\n",
+        "- **Microsoft GraphRAG Local Search**: Evaluates 1-2 hop neighborhood around seed entities.\n",
+        "- **Microsoft GraphRAG Global Search**: Aggregates macro community summaries.\n",
+        "- **QDB Deductive Engine**: Minimizes the global QCBO Hamiltonian across all 6 hops simultaneously while filtering contradictory claims with $+50.0$ energy barriers."
+      ]
+    }
+  ],
+  "metadata": {
+    "language_info": {
+      "name": "python"
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "language": "python",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 4
+}