Multimodal_PyMuPDF.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "gpuType": "A100",
      "authorship_tag": "ABX9TyNbGerV9yhRTcjA6bhXPSjG",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/benitomartin/multimodal-llm-pymupdf4llm/blob/main/Multimodal_PyMuPDF.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "collapsed": true,
        "id": "gjnsESDkLwwn",
        "outputId": "522f23e9-02bc-4f99-e4b8-a12d42ce65c5"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "  Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n"
          ]
        }
      ],
      "source": [
        "!pip install -qq pymupdf4llm\n",
        "!pip install -qq llama-index\n",
        "!pip install -qq llama-index qdrant-client\n",
        "!pip install -qq llama-index-vector-stores-qdrant\n",
        "!pip install -qq git+https://github.com/openai/CLIP.git\n",
        "!pip install -qq llama-index-embeddings-clip"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import os\n",
        "from google.colab import userdata\n",
        "import pymupdf4llm\n",
        "\n",
        "import qdrant_client\n",
        "from qdrant_client import QdrantClient\n",
        "from qdrant_client.http.models import VectorParams, Distance\n",
        "\n",
        "from llama_index.core import Document, SimpleDirectoryReader, StorageContext\n",
        "from llama_index.vector_stores.qdrant import QdrantVectorStore\n",
        "from llama_index.core.indices import MultiModalVectorStoreIndex\n",
        "from llama_index.core.node_parser import SentenceSplitter\n",
        "from llama_index.core.schema import ImageDocument, ImageNode\n",
        "from llama_index.multi_modal_llms.openai import OpenAIMultiModal\n",
        "from llama_index.core.response.notebook_utils import display_source_node\n",
        "\n",
        "import json"
      ],
      "metadata": {
        "id": "PAbIs5yz6ahR"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "os.environ[\"OPENAI_API_KEY\"] = userdata.get('OPENAI_API_KEY')"
      ],
      "metadata": {
        "id": "8lgAHnUp-a6T"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "You can download the document used in this notebook [here](https://arxiv.org/abs/1706.03762)"
      ],
      "metadata": {
        "id": "qRXo_vvIFI_m"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Perform the markdown conversion\n",
        "docs = pymupdf4llm.to_markdown(doc=\"/content/document.pdf\",\n",
        "                                  page_chunks = True,\n",
        "                                  write_images = True,\n",
        "                                  image_path = \"/content/images\",\n",
        "                                  image_format = \"jpg\",\n",
        "                                  )"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6Tsil3nEL1U8",
        "outputId": "274aa973-2aa8-4715-a557-ecbad066e982"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Processing /content/document.pdf...\n",
            "[                                        ] (0/15)\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b==\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b[==                                      ] ( 1/15)\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b===\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b[=====                                   ] ( 2/15)\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b==\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b[========                                ] ( 3/15)\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b===\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b[==========                              ] ( 4/15)\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b===\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b[=============                           ] ( 5/15)\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b==\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b[================                        ] ( 6/15)\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b===\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b[==================                      ] ( 7/15)\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b===\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b[=====================                   ] ( 8/15)\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b==\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b[========================                ] ( 9/15)\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b===\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b[==========================              ] (10/15)\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b===\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b[=============================           ] (11/15)\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b==\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b[================================        ] (12/15)\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b===\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b[==================================      ] (13/15)\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b===\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b[=====================================   ] (14/15)\b\b\b\b\b\b\b\b\b\b\b\b==\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b[========================================] (15/15)\b\b\b\b\b\b\b\b\b]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "llama_documents = []\n",
        "\n",
        "for document in docs:\n",
        "    # Extract just the 'metadata' field and convert certain elements as needed\n",
        "    metadata = {\n",
        "        \"file_path\": document[\"metadata\"].get(\"file_path\"),\n",
        "        \"page\": str(document[\"metadata\"].get(\"page\")),\n",
        "        \"images\": str(document.get(\"images\")),\n",
        "        \"toc_items\": str(document.get(\"toc_items\")),\n",
        "\n",
        "    }\n",
        "\n",
        "    # Create a Document object with just the text and the cleaned metadata\n",
        "    llama_document = Document(\n",
        "        text=document[\"text\"],\n",
        "        metadata=metadata,\n",
        "        text_template=\"Metadata: {metadata_str}\\n-----\\nContent: {content}\",\n",
        "    )\n",
        "\n",
        "    llama_documents.append(llama_document)"
      ],
      "metadata": {
        "id": "WAgbKthmNYDp"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "llama_documents[0]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "collapsed": true,
        "id": "Wz_MWZZCe3Md",
        "outputId": "24d673ee-54d6-40a1-b33d-9c54417d0c13"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Document(id_='2e043b1d-cdc4-47b1-81df-278c852dc534', embedding=None, metadata={'file_path': '/content/document.pdf', 'page': '1', 'images': '[]', 'toc_items': '[]'}, excluded_embed_metadata_keys=[], excluded_llm_metadata_keys=[], relationships={}, text='#### Provided proper attribution is provided, Google hereby grants permission to reproduce the tables and figures in this paper solely for use in journalistic or scholarly works.\\n\\n## Attention Is All You Need\\n\\n\\n**Ashish Vaswani[∗]**\\nGoogle Brain\\n```\\navaswani@google.com\\n\\n```\\n**Llion Jones[∗]**\\nGoogle Research\\n```\\n llion@google.com\\n\\n```\\n\\n**Noam Shazeer[∗]**\\nGoogle Brain\\n```\\nnoam@google.com\\n\\n```\\n\\n**Aidan N. Gomez[∗†]**\\nUniversity of Toronto\\n```\\naidan@cs.toronto.edu\\n\\n```\\n\\n**Niki Parmar[∗]**\\nGoogle Research\\n```\\nnikip@google.com\\n\\n```\\n\\n**Jakob Uszkoreit[∗]**\\nGoogle Research\\n```\\nusz@google.com\\n\\n```\\n\\n**Łukasz Kaiser[∗]**\\nGoogle Brain\\n```\\nlukaszkaiser@google.com\\n\\n```\\n\\n**Illia Polosukhin[∗‡]**\\n```\\nillia.polosukhin@gmail.com\\n\\n#### Abstract\\n\\n```\\n\\nThe dominant sequence transduction models are based on complex recurrent or\\nconvolutional neural networks that include an encoder and a decoder. The best\\nperforming models also connect the encoder and decoder through an attention\\nmechanism. We propose a new simple network architecture, the Transformer,\\nbased solely on attention mechanisms, dispensing with recurrence and convolutions\\nentirely. Experiments on two machine translation tasks show these models to\\nbe superior in quality while being more parallelizable and requiring significantly\\nless time to train. Our model achieves 28.4 BLEU on the WMT 2014 Englishto-German translation task, improving over the existing best results, including\\nensembles, by over 2 BLEU. On the WMT 2014 English-to-French translation task,\\nour model establishes a new single-model state-of-the-art BLEU score of 41.8 after\\ntraining for 3.5 days on eight GPUs, a small fraction of the training costs of the\\nbest models from the literature. We show that the Transformer generalizes well to\\nother tasks by applying it successfully to English constituency parsing both with\\nlarge and limited training data.\\n\\n_∗Equal contribution. Listing order is random. Jakob proposed replacing RNNs with self-attention and started_\\nthe effort to evaluate this idea. Ashish, with Illia, designed and implemented the first Transformer models and\\nhas been crucially involved in every aspect of this work. Noam proposed scaled dot-product attention, multi-head\\nattention and the parameter-free position representation and became the other person involved in nearly every\\ndetail. Niki designed, implemented, tuned and evaluated countless model variants in our original codebase and\\ntensor2tensor. Llion also experimented with novel model variants, was responsible for our initial codebase, and\\nefficient inference and visualizations. Lukasz and Aidan spent countless long days designing various parts of and\\nimplementing tensor2tensor, replacing our earlier codebase, greatly improving results and massively accelerating\\nour research.\\n_†Work performed while at Google Brain._\\n_‡Work performed while at Google Research._\\n\\n\\n31st Conference on Neural Information Processing Systems (NIPS 2017), Long Beach, CA, USA.\\n\\n\\n-----\\n\\n', mimetype='text/plain', start_char_idx=None, end_char_idx=None, text_template='Metadata: {metadata_str}\\n-----\\nContent: {content}', metadata_template='{key}: {value}', metadata_seperator='\\n')"
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "client = qdrant_client.QdrantClient(location=\":memory:\")"
      ],
      "metadata": {
        "collapsed": true,
        "id": "OGVhYNytL1W9"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Create a collection for text data\n",
        "client.create_collection(\n",
        "    collection_name=\"text_collection\",\n",
        "    vectors_config=VectorParams(size=1536, distance=Distance.COSINE)\n",
        ")\n",
        "\n",
        "# Create a collection for image data\n",
        "client.create_collection(\n",
        "    collection_name=\"image_collection\",\n",
        "    vectors_config=VectorParams(size=512, distance=Distance.COSINE)\n",
        ")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5BIWDXuhL1at",
        "outputId": "2fda917a-de0d-4d65-f453-47a1c34e5e23"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Initialize Collections\n",
        "text_store = QdrantVectorStore(\n",
        "    client=client, collection_name=\"text_collection\"\n",
        ")\n",
        "\n",
        "image_store = QdrantVectorStore(\n",
        "    client=client, collection_name=\"image_collection\"\n",
        ")"
      ],
      "metadata": {
        "id": "7h-nJOuML1cy"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Create the MultiModal index\n",
        "\n",
        "storage_context = StorageContext.from_defaults(\n",
        "    vector_store=text_store, image_store=image_store\n",
        ")\n",
        "\n",
        "# context images\n",
        "image_path = \"/content/images\"\n",
        "image_documents = SimpleDirectoryReader(image_path).load_data()\n",
        "\n",
        "index = MultiModalVectorStoreIndex.from_documents(\n",
        "    llama_documents + image_documents,\n",
        "    storage_context=storage_context)"
      ],
      "metadata": {
        "id": "VbjED6KUL1hF",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "29058678-b92e-4c3a-dfdf-dde7efb52be5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|████████████████████████████████████████| 338M/338M [00:02<00:00, 133MiB/s]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Set query and retriever\n",
        "query = \"Could you provide an image of the Multi-Head Attention?\"\n",
        "\n",
        "retriever = index.as_retriever(similarity_top_k=1, image_similarity_top_k=1)  # Adjust to retrieve enough images\n",
        "\n",
        "retrieval_results = retriever.retrieve(query)"
      ],
      "metadata": {
        "id": "u6htTs7n-Rg4"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "retrieval_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LM1q63lM-RjY",
        "outputId": "b4de99a5-f239-43f3-bcf7-613f8c09b53a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[NodeWithScore(node=TextNode(id_='70f127a0-5a71-4d31-abbe-b387efe87004', embedding=None, metadata={'file_path': '/content/document.pdf', 'page': '5', 'images': '[]', 'toc_items': \"[[3, 'Applications of Attention in our Model', 5], [2, 'Position-wise Feed-Forward Networks', 5], [2, 'Embeddings and Softmax', 5]]\"}, excluded_embed_metadata_keys=[], excluded_llm_metadata_keys=[], relationships={<NodeRelationship.SOURCE: '1'>: RelatedNodeInfo(node_id='1cef4562-f63b-4210-a5a1-9cfb73c243bb', node_type=<ObjectType.DOCUMENT: '4'>, metadata={'file_path': '/content/document.pdf', 'page': '5', 'images': '[]', 'toc_items': \"[[3, 'Applications of Attention in our Model', 5], [2, 'Position-wise Feed-Forward Networks', 5], [2, 'Embeddings and Softmax', 5]]\"}, hash='0f2ad0b590aa7f8bb81d763dae8e2d648e6530216a1c9e5b913656abeb4f9fe3')}, text='output values. These are concatenated and once again projected, resulting in the final values, as\\ndepicted in Figure 2.\\n\\nMulti-head attention allows the model to jointly attend to information from different representation\\nsubspaces at different positions. With a single attention head, averaging inhibits this.\\n\\nMultiHead(Q, K, V ) = Concat(head1, ..., headh)W _[O]_\\n\\nwhere headi = Attention(QWi[Q][, KW][ K]i _[, V W][ V]i_ [)]\\n\\nWhere the projections are parameter matrices Wi[Q] _∈_ R[d][model][×][d][k], Wi[K] _∈_ R[d][model][×][d][k], Wi[V] _∈_ R[d][model][×][d][v]\\nand W _[O]_ _∈_ R[hd][v][×][d][model].\\n\\nIn this work we employ h = 8 parallel attention layers, or heads. For each of these we use\\n_dk = dv = dmodel/h = 64. Due to the reduced dimension of each head, the total computational cost_\\nis similar to that of single-head attention with full dimensionality.\\n\\n**3.2.3** **Applications of Attention in our Model**\\n\\nThe Transformer uses multi-head attention in three different ways:\\n\\n    - In \"encoder-decoder attention\" layers, the queries come from the previous decoder layer,\\nand the memory keys and values come from the output of the encoder. This allows every\\nposition in the decoder to attend over all positions in the input sequence. This mimics the\\ntypical encoder-decoder attention mechanisms in sequence-to-sequence models such as\\n\\n[38, 2, 9].\\n\\n    - The encoder contains self-attention layers. In a self-attention layer all of the keys, values\\nand queries come from the same place, in this case, the output of the previous layer in the\\nencoder. Each position in the encoder can attend to all positions in the previous layer of the\\nencoder.\\n\\n    - Similarly, self-attention layers in the decoder allow each position in the decoder to attend to\\nall positions in the decoder up to and including that position. We need to prevent leftward\\ninformation flow in the decoder to preserve the auto-regressive property. We implement this\\ninside of scaled dot-product attention by masking out (setting to ) all values in the input\\n_−∞_\\nof the softmax which correspond to illegal connections. See Figure 2.\\n\\n**3.3** **Position-wise Feed-Forward Networks**\\n\\nIn addition to attention sub-layers, each of the layers in our encoder and decoder contains a fully\\nconnected feed-forward network, which is applied to each position separately and identically. This\\nconsists of two linear transformations with a ReLU activation in between.\\n\\nFFN(x) = max(0, xW1 + b1)W2 + b2 (2)\\n\\nWhile the linear transformations are the same across different positions, they use different parameters\\nfrom layer to layer. Another way of describing this is as two convolutions with kernel size 1.\\nThe dimensionality of input and output is dmodel = 512, and the inner-layer has dimensionality\\n_dff = 2048._\\n\\n**3.4** **Embeddings and Softmax**\\n\\nSimilarly to other sequence transduction models, we use learned embeddings to convert the input\\ntokens and output tokens to vectors of dimension dmodel. We also use the usual learned linear transformation and softmax function to convert the decoder output to predicted next-token probabilities. In\\nour model, we share the same weight matrix between the two embedding layers and the pre-softmax\\nlinear transformation, similar to [30]. In the embedding layers, we multiply those weights by _[√]dmodel._\\n\\n\\n-----', mimetype='text/plain', start_char_idx=0, end_char_idx=3329, text_template='Metadata: {metadata_str}\\n-----\\nContent: {content}', metadata_template='{key}: {value}', metadata_seperator='\\n'), score=0.8199309663422424),\n",
              " NodeWithScore(node=ImageNode(id_='9e0b8f47-4313-4357-99da-a4c92ed5342e', embedding=None, metadata={'file_path': '/content/images/document.pdf-3-0.jpg', 'file_name': 'document.pdf-3-0.jpg', 'file_type': 'image/jpeg', 'file_size': 21547, 'creation_date': '2024-09-30', 'last_modified_date': '2024-09-30'}, excluded_embed_metadata_keys=['file_name', 'file_type', 'file_size', 'creation_date', 'last_modified_date', 'last_accessed_date'], excluded_llm_metadata_keys=['file_name', 'file_type', 'file_size', 'creation_date', 'last_modified_date', 'last_accessed_date'], relationships={<NodeRelationship.SOURCE: '1'>: RelatedNodeInfo(node_id='7b6dae63-ba6d-4469-86c3-bbfc80323449', node_type=<ObjectType.DOCUMENT: '4'>, metadata={'file_path': '/content/images/document.pdf-3-0.jpg', 'file_name': 'document.pdf-3-0.jpg', 'file_type': 'image/jpeg', 'file_size': 21547, 'creation_date': '2024-09-30', 'last_modified_date': '2024-09-30'}, hash='819a4d26027c292af9d778f2cf7ff9942fc7f254f6550db467cc58ebba9be038')}, text='', mimetype='text/plain', start_char_idx=0, end_char_idx=0, text_template='{metadata_str}\\n\\n{content}', metadata_template='{key}: {value}', metadata_seperator='\\n', image=None, image_path='/content/images/document.pdf-3-0.jpg', image_url=None, image_mimetype=None, text_embedding=None), score=0.2589032981592738)]"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "from PIL import Image\n",
        "\n",
        "def plot_images(image_paths):\n",
        "    images_shown = 0\n",
        "    plt.figure(figsize=(16, 9))\n",
        "    for img_path in image_paths:\n",
        "        if os.path.isfile(img_path):\n",
        "            image = Image.open(img_path)\n",
        "\n",
        "            plt.subplot(2, 3, images_shown + 1)\n",
        "            plt.imshow(image)\n",
        "            plt.xticks([])\n",
        "            plt.yticks([])\n",
        "\n",
        "            images_shown += 1\n",
        "            if images_shown >= 9:\n",
        "                break\n",
        "\n",
        "\n",
        "plot_images(image_path)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "id": "vMaDcwuT37Bi",
        "outputId": "5a820c75-8891-4499-c17a-5b36b1533d3f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1600x900 with 0 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "retrieved_image = []\n",
        "\n",
        "for res_node in retrieval_results:\n",
        "    if isinstance(res_node.node, ImageNode):\n",
        "        print(\"Highest Scored ImageNode\")\n",
        "        print(\"-----------------------\")\n",
        "        retrieved_image.append(res_node.node.metadata[\"file_path\"])\n",
        "    else:\n",
        "        print(\"Highest Scored TextNode\")\n",
        "        print(\"-----------------------\")\n",
        "\n",
        "        display_source_node(res_node, source_length=200)\n",
        "\n",
        "plot_images(retrieved_image)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 512
        },
        "id": "sErQhPTLL1t_",
        "outputId": "b61ddddd-d685-454f-8890-5c0cb528d080"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Highest Scored TextNode\n",
            "-----------------------\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.Markdown object>"
            ],
            "text/markdown": "**Node ID:** 70f127a0-5a71-4d31-abbe-b387efe87004<br>**Similarity:** 0.8199309663422424<br>**Text:** output values. These are concatenated and once again projected, resulting in the final values, as\ndepicted in Figure 2.\n\nMulti-head attention allows the model to jointly attend to information from ...<br>"
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Highest Scored ImageNode\n",
            "-----------------------\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1600x900 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    }
  ]
}