Script.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "id": "7e95804a",
      "metadata": {
        "id": "7e95804a"
      },
      "source": [
        "#  Image prossecing "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 28,
      "id": "637a10f0",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 347
        },
        "id": "637a10f0",
        "outputId": "9869abb4-eaf2-4bfe-cbfe-50c06d53151b"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(60000, 28, 28)\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'data = []\\nfor i in range(500):\\n  image=train_X[i]\\n  imageRow = [pixel for pixelRow in image for pixel in pixelRow]\\n  data.append(imageRow)'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 28
        }
      ],
      "source": [
        "#!pip install tensorflow\n",
        "from keras.datasets import mnist\n",
        "from matplotlib import pyplot\n",
        "\n",
        "#Reading the dataset 60 000 images\n",
        "(train_X, train_y), (test_X, test_y) = mnist.load_data()\n",
        "# Train_X is an array of 28*28 images and Train_Y is a list of the corresponding numbre \n",
        "# i.e Train_X[i] is an image of Train_Y[i] numbre\n",
        "# displaying 10 images from the data set \n",
        "for i in range(3):  \n",
        "    pyplot.subplot(330 + 1 + i)\n",
        "    pyplot.imshow(train_X[i], cmap=pyplot.get_cmap('gray'))\n",
        "    pyplot.show()\n",
        "print(train_X.shape)\n",
        "\"\"\"data = []\n",
        "for i in range(500):\n",
        "  image=train_X[i]\n",
        "  imageRow = [pixel for pixelRow in image for pixel in pixelRow]\n",
        "  data.append(imageRow)\"\"\""
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# insert images into mongoDB"
      ],
      "metadata": {
        "id": "kEMNnVaAY74r"
      },
      "id": "kEMNnVaAY74r"
    },
    {
      "cell_type": "code",
      "source": [
        "\"\"\"\n",
        "#instaling cnx packages\n",
        "!/usr/bin/python3 -m pip install \"pymongo[srv]\"\n",
        "!pip install dnspython\n",
        "\"\"\"\n",
        "\n",
        "# Create a connection using MongoClient. You can import MongoClient or use pymongo.MongoClient\n",
        "from pymongo import MongoClient\n",
        "import dns\n",
        "import numpy as np\n",
        "# Provide the mongodb atlas url to connect python to mongodb using pymongo\n",
        "CONNECTION_STRING = \"mongodb+srv://AMIRABELKIS:AMIRABELKIS@cluster0.jaykjmi.mongodb.net/?retryWrites=true&w=majority\"\n",
        "cluster = MongoClient(CONNECTION_STRING)\n",
        "database = cluster['TP-BDDA']\n",
        "collection = database['mnist-Photo']\n"
      ],
      "metadata": {
        "id": "GGoIbfKpp9Xv"
      },
      "id": "GGoIbfKpp9Xv",
      "execution_count": 25,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "\"\"\" \n",
        "Must clear all the posts insid database to avoid dublicate id err\n",
        "\"\"\"\n",
        "# Deleting all the posts in the collection\n",
        "myquery= {}\n",
        "database['mnist-Photo'].delete_many(myquery)\n",
        "database['mnist-Compressed'].delete_many(myquery)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "IgXJiR46pAQA",
        "outputId": "e8af9225-2e20-431d-afff-8e0fc6f7450c"
      },
      "id": "IgXJiR46pAQA",
      "execution_count": 83,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<pymongo.results.DeleteResult at 0x7fa9bd718590>"
            ]
          },
          "metadata": {},
          "execution_count": 83
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "for i in range(500):\n",
        "  image=train_X[i]\n",
        "  numbre=train_y[i]\n",
        "  imageRow = [int(pixel) for pixelRow in image for pixel in pixelRow]\n",
        "  print('image',type(image))\n",
        "  print('image row',type(imageRow))\n",
        "  print('image numbre',type(str(numbre)),str(numbre))\n",
        "  post = {\n",
        "      \"_id\": i,\n",
        "      \"numbre\":str(numbre), # numbre is <class 'numpy.uint8'> can't insert in mongoDB convert to string\n",
        "      \"image\": imageRow,\n",
        "  }\n",
        "  collection.insert_one(post)"
      ],
      "metadata": {
        "id": "fzwZgtuY-nIC"
      },
      "id": "fzwZgtuY-nIC",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Load data from MongoDB "
      ],
      "metadata": {
        "id": "2miebqEStO66"
      },
      "id": "2miebqEStO66"
    },
    {
      "cell_type": "code",
      "source": [
        "def getImg(numbre):\n",
        "  # the collection.find method return a <class 'pymongo.cursor.Cursor'> object\n",
        "  return collection.find({'numbre':numbre})\n",
        "numbre = '1'\n",
        "cursor = getImg(numbre)\n",
        "data = []\n",
        "imageIdList = []\n",
        "for post in cursor:\n",
        "  imageRow = post['image']\n",
        "  imageIdList.append(post['_id'])\n",
        "  data.append(imageRow)\n"
      ],
      "metadata": {
        "id": "Z6bqaNchtXfX"
      },
      "id": "Z6bqaNchtXfX",
      "execution_count": 68,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "id": "05cc55ea",
      "metadata": {
        "id": "05cc55ea"
      },
      "source": [
        "# Performing PCA "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 58,
      "id": "40fd7950",
      "metadata": {
        "id": "40fd7950",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "outputId": "eb6a0408-ab0e-45f8-e7ec-348430771182"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "import random\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import pandas as pd\n",
        "import sklearn as sk\n",
        "from sklearn.decomposition import PCA\n",
        "\n",
        "# scaling data ie (data - mean)^2/ standard deviation\n",
        "scaled_data = sk.preprocessing.scale(data)\n",
        "\n",
        "pca = PCA()\n",
        "pca.fit(scaled_data)\n",
        "pca_data = pca.transform(scaled_data)\n",
        "# claculate the % of variation of each PC \n",
        "per_var = np.round(pca.explained_variance_ratio_ *100 , decimals =2)\n",
        "# name each PC \n",
        "labels = ['PC' + str(i) for i in range(1, len(per_var) + 1)]\n",
        "# Create graph PC-%\n",
        "plt.bar(x=range(1,len(per_var)+1), height=per_var)#, tick_label=labels)\n",
        "plt.ylabel('Percentage')\n",
        "plt.xlabel('Principal Component')\n",
        "plt.title('Percentage of Explained Variance')\n",
        "plt.savefig('Explained variance percentage by PC.png')\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# % cummule\n",
        "plt.grid()\n",
        "plt.plot(np.cumsum(pca.explained_variance_ratio_ * 100))\n",
        "plt.xlabel('Number of components')\n",
        "plt.ylabel('Explained variance')\n",
        "plt.savefig('Explained variance percentage.png')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "id": "YmaY0M4M0IjK",
        "outputId": "51cc7277-2fb9-4cb2-bae7-e9ddf1cb2cf6"
      },
      "id": "YmaY0M4M0IjK",
      "execution_count": 59,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Taking only the PC that have cumulative percentage < per%\n",
        "def nbr_PC(per):\n",
        "  cumul_per_var = 0\n",
        "  for i in range(0 , len(per_var)):\n",
        "    if(cumul_per_var<= per):\n",
        "      cumul_per_var += per_var[i]\n",
        "    else:\n",
        "      return i  \n",
        "      break"
      ],
      "metadata": {
        "id": "hzL2u94Q7pTW"
      },
      "id": "hzL2u94Q7pTW",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Preserve 80% \n",
        "pca_80 = PCA(n_components=nbr_PC(80))\n",
        "mnist_pca_80_reduced = pca_80.fit_transform(data)\n",
        "mnist_pca_80_recovered = pca_80.inverse_transform(mnist_pca_80_reduced)\n",
        "\n",
        "image_pca_80 = mnist_pca_80_recovered[1,:].reshape([28,28])\n",
        "plt.imshow(image_pca_80, cmap='gray')\n",
        "plt.title('Compressed image with 80%', fontsize=15, pad=15)\n",
        "plt.savefig(\"image_pca_80.png\")\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 292
        },
        "id": "8POf2EoKKO0W",
        "outputId": "5df12314-af07-4c51-ad7a-03a41a4d5f57"
      },
      "id": "8POf2EoKKO0W",
      "execution_count": 55,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP0AAAETCAYAAAD9DAC6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAZp0lEQVR4nO3de7hdVX3u8e+PnftNEnIhIYEIYi1tj6C7HH0EjKUioILUAnI8Hqja2AqtfeqpUmyVPq2VIpdSa7GxUMC2KOdRKp5SFYMawUrZ0EDCHUIiCbmHQHYSQi6//jHGhpXJWmPsrL0uOxnv53nWs9eaY17Gmnu+c841x7yYuyMi5Tio2xUQkc5S6EUKo9CLFEahFymMQi9SGIVepDAdD72Zvd/M7jSzzWa2w8weN7OrzGxWp+tyoDGzCWbmZnZBop+5sZ/3dLBqw0a9729mnzKzeXX6dTO7qMnpfMDM7jezfjNbZWY3VZdxCy4xs2fMbLuZLTKzYyv9vMHM7jGz583s62Y2oVJ+Uhz/Xt1TOhp6M7sSuAVYBnwIOAW4GjgZ+HIn61Kw1cBbgbu6XZEuqff9PwXMa9UEzOwM4Gbgp8CZwKeBk4B/M7PazF0M/CnwV8B7gX7gB2Z2aE0/NwBPAucAxwCX1EznIOAa4I/dvX/QFXT3jrzil3Lgw3XKeoDTOlWXJuvfA4zqdj0ydZwQ5/EF3a7L/vQCNgCX1unuwEVNjO/rwH2VbmfE8f1i/DwGeB74bE0/44H1wF9U/p/T4udzgXtr+v8o8J+A7VP9Ojhj76zOiES/U4EbgY3ANuBHQG+ln+XAFXFtuTrOwCsBA04HHgK2AP8KTK4Zbl6ckacA/x/YCvwc+J3K+G8A+oD3xXHtBE6MZWfGsheBNcDlwMiaYWcT9mjWAduBp4A/ryn/JeC7wKY4/UeACyvTT04j9vN+4PE4jUVAby70wNzYz3taMC/HA38LPBb/T08T9tgmVaY5mRCErcCzhC3fFcDySn+Hx/42xfF9D/iFxHcZA+wA/ldNty/E73dGTbcvAXfX+/7xu3vlNa8m9J8A/pIQxnXx+43OLL/fBO6sdDspju+Y+PnX4uc3VPq7npgTYErsZ0L8/F7ggfh+Ulwu3rrPWexQ4EfGhffzg+z/rviFfit+0UVxoXtdZUFdCXwLOBX4TJxBVwP3Ab8BfBB4DvhKndA/E/+Z7wK+wquDcANhC/A48L+BXyeE+RxgN/B3hBXH7wKbgSsqK7j/IKww5gEfBi6vKV8G/BshUCcDHwcurikfzDTeBOwC/h9wGvBHcbzNhr6ZeTkNuBb4TeDtcT49AnyvMs1vE1bgHwXeAyyM8395TT9TCCvf/4rf/z1xOXgGGJtZVq6t+Xw3YSX4xZpuDwCXNQj9cXHe/gPwlviaVBP6n8dl4V1xHu8CPpVZft9N2Ej8H0I4X09YhhfW9PPxOK6eyrB/BGyt+fw08Mdx/nwHuC52/yLwT03lsUOhPzTOwI8Not9TY79vr+k2sNvz95UF9cnamUbY1dkFvLam2+XA2jqhX1CZ7h3Azyqhd+DYmm4GrAD+sTLsh+OCdkj83A+8t8H3mxrH+ysNygc7jVuAh6nZteOVsDYT+n2el3XGPQJ4Wxz/4bHbL8fPZ9f0N5awQq0N/Z8TVgxTarpNJux1XJiY5heApfH9wJb/bwf+l8DBhBXouxPfP7V7v6jS7V9rl5NEvT5I2NAN7D3cDRxc+V9trjPcR2P/o+Ln0+Ly5IQN0OHA6wgr4NnN5LHTR+99EP0cD6xz9x+/PJD7VsKu+AmVfn/k7rtrPj9JWJCernSbZmajKsPeWvn8LeDNZtZT022Vuy+u+fx6wky/xcxGDLwIW/YxhAUcYDHwBTO7wMwOr0xnE2Hr9RUzO9fMplfKBzuN44HbPC4ZNd+hWU3NSzP7kJn9l5n1E7ZuAwfIXh//9sa/3xkYxt23Az+oTP/XCSveF2q+8xbCnkYvjS0CjjGzKYStdD9h7+NNZjaOV5aZuxPjSPl+5fPDhD2+hszsHYS9x2uAdwAfIGypb60sX1nu/u/AdOAXCMcDfg5cBVzt7ivN7EIz+3l8fXww4+xU6DcS1sDVANQzk/DbqWotYcbV2lz5/FKDbgZUQ1+dxjrClmpqZZq1BspuJyzgA6+BYMyJf88l/B6/GlhhZovN7GQAd99D2GVfQ/j9tsbMfmJmx+3jNA5t8B2atc/z0szOAm4i/JQ5mxC6s2K/Y2rqucXdX6yMa33l81TCfNtZeb2DV75zPT8lbExOAE4khPthwh7CW2K3pe5e/S6DVW8ejKnXY40rCSvkT7v7j9z9G7zyU+/M2M9zwIQ6K4HJwDZ3f2mgg7tvc/fH3X23mb0TeCPwRTN7I2EP6ZT4+ryZ/Y/cFxqR66EV3H2nmd1N+F30J5neVxPWbFUzCFvJVqlOYzphd3ZDTbfqnsnA9OcTfntWPQ3g7quAC2KTyvHApcBtZna4u29090eB95vZSMJC+VeE5pzZg50GYaVR7zt00tnAPe7+8hbGzN5e6WcNMNHMxlSCP63S3ybgNsJCXLWlUQXc/Xkze5AwH48lHE9wM7srdjsJ+Mlgv1CLvIHQZFdbz8fMbDtwVOz0KKFF6HWEA6G1wz5ab6RxBXE14ZjC9nhuwZ1xecLMFhKOrTyYqlwnd+//Gug1s/OrBWZ2kJmdGj/eA0w3s5NqyscRDo60sm35rDqf76vs4lY9BqwC5rp7X53Xxtqe3X2Pu/8M+DNgHHBEpXynu99J2F2bSfj9Odhp3AucYWZWM8rf2Kc5MHRjCXtwtT5Y+dwX/54x0MHMxgLvrPS3kNCq8VCd7/wYaYsIR8PfGt8PdHsX8GbyoR/M1ntfrCAcaH2Zmf0iYX4tj51+CrxAWHEO9DOOcOD63xuM93eB5+Kew4BxNe/HE/bEkjqypQdw9++Y2VXAdWb2NsIR3X7Cmu13CDPju+7+PTP7KfANM7uY8NPg/xJm2BdbWKXTzOzzwI8JYXknr+x6NfoOe8zsk8DXzGwS4Z/zEnAkYfftNwktFd8j7PY+DowGPknY4j0Sd7+uAL5BONo+mdCE9YC7bwLITcPdtxH2Du4h/Pa/jvBb/yMtmC/74g7gy2b2mViXgdaIl7n7UjP7DnCtmU0kzIc/JDTJ7anp9SrC0f87zexLhBXfDMKW6y5332vLWfET4PcJy9P9Nd2uqnmf8ijwbjP7bhzHY+7ecO9iEL4CXG1mzxL+fzOAzxKW8dsB3P1FM7sM+FMzey7W4Q8JG+IvVUcYj1l8jrAiG7AIuNzMPkwI+68Rml3Tmjn6N5QXoW35h4TfXC8RgnEFcGhNP9MIoXmOcMT6x8CvVsaznJomrNjtBqCv0u0C9m7rnBc/vyv+Q7YRmqs+nhtXTdlphAVpK2FtvRj4C8JKdDTwVV5pu95AOAj5K3HY6cDXCIEfaIO/mXi0ezDTqOnnbMLBtRcJe0G/yhDa6ZuYlz3xf7cu1vGbwP+sM/4phJXcVsJxks/GebS4Mv5ZwD/GfnbEev0T8EuZZWpGnOb3a7r1EH4WLBvE938z8LNYP2fvdvqLKsNfCmzI1McIW+UH4zhXxe9/ZJ3+PhOXv+3x/31cg3H+DbG5rtL9E4SfxKuB3xtMBs33Ovh74Iu/g35ICOHSLlenSPHI/FLC8YBX/dyT9urY7r2Uy8zOJmzFlxBOVvlt4GjCySvSYQq9dMJWwtmVryPsdi8hnLz0n12tVaGK270XKZ1uoiFSGIVepDAKvUhhFHqRwij0IoVR6EUKo9CLFEahFymMQi9SGIVepDAdPfe+p6fHR4zQ6f4i7bJr1y52796dvJHGkBIY73ZzDeEiin9w98uSExsxgtmzk/cUFJEhWLlyZbafpnfv4/26vky42cMxwHlmdkyz4xORzhjKb/rjgSfdfZmHO3d+ncztpkSk+4YS+sMI928fsDJ224uZzTezPjPr27NnT7VYRDqs7Ufv3X2Bu/e6e+9BB6mxQKTbhpLCVez9EILZsZuIDGNDCf29wNFm9tr4mKMPEB5WICLDWNNNdu6+y8wuItzjvQe43t0falnNRKQthtRO7+63E2/eLyL7Bx1ZEymMQi9SGIVepDAKvUhhFHqRwij0IoVR6EUKo9CLFEahFymMQi9SGIVepDAKvUhhFHqRwuh+1DIkkyZNSpanbnm+ZcuW5LA7d+5sqk6Spi29SGEUepHCKPQihVHoRQqj0IsURqEXKYxCL1IYtdO3QLsf15V7MtCsWbMalp177rnJYadNm5Ysf+KJJ5LlixYtSpZv3LixYVl/f39y2FWr0s9OGTNmTLJ88uTJDcvGjh2bHPZApi29SGEUepHCKPQihVHoRQqj0IsURqEXKYxCL1IYtdPvB8wsWZ5qj547d25y2Fw7fa4t/IEHHkiWP/XUUw3Lcm3l06dPT5aPGzcuWb579+6GZbt27UoOmzs3Ilc+nA0p9Ga2HNgC7AZ2uXtvKyolIu3Tii39O9x9QwvGIyIdsP/uo4hIU4Yaege+b2b3mdn8ej2Y2Xwz6zOzvnafoy4ieUPdvT/B3VeZ2XTgDjN71N33ugLD3RcACwBGjx7tQ5yeiAzRkLb07r4q/l0H3Aoc34pKiUj7NB16MxtvZhMH3gOnAEtbVTERaY+h7N7PAG6NbcgjgH9x9++2pFbDUDePR6TamyFdt82bNyeHnTNnTrJ8xowZyfIXXnghWb5+/fqGZaNHj04O+5rXvCZZPnHixGR56r76Q22n3581HXp3Xwa8sYV1EZEOOHBXZyJSl0IvUhiFXqQwCr1IYRR6kcLo0tpBSjXhdPv04q1btzY97EsvvZQsd0+fRJlrsks1N+Yu2x05cmSyPFe31Hfbtm1bcthcc+D+3KS3/9ZcRJqi0IsURqEXKYxCL1IYhV6kMAq9SGEUepHCqJ2+BXJttu1ux0898jnXhj9hwoRkea4t/cQTT0yWpy7tzc233CXFuUdd79ixo2HZqFGjksPm/mdqpxeR/YZCL1IYhV6kMAq9SGEUepHCKPQihVHoRQqjdvphINcmnGpvBti5c2fDslx7dK6dPvc46ClTpjRdvnHjxuSwObl2/NT19rnHZB/ItKUXKYxCL1IYhV6kMAq9SGEUepHCKPQihVHoRQqjdvoOyD0WOdXODvn7u/f09OxznQbk2vFz98WfNWtWsnzSpEkNy3L3zM9dL5+r2/jx45Plpcpu6c3sejNbZ2ZLa7pNMbM7zOyJ+Hdye6spIq0ymN37G4BTK90uBha6+9HAwvhZRPYD2dC7+yJgU6XzmcCN8f2NwPtaXC8RaZNmf9PPcPfV8f0aYEajHs1sPjAfYMQIHUIQ6bYhH733cJSp4ZEmd1/g7r3u3rs/30xQ5EDRbArXmtlMgPh3XeuqJCLt1GzobwPOj+/PB77dmuqISLtlf2Sb2c3APGCqma0EPgdcBtxiZh8BVgDntLOSw4GZNSzLtaPnjmXk7i2fK09dE79q1arksLm28Nw5BJMnp1trU895z51fkGuHz/1c1M/J+rKhd/fzGhSd3OK6iEgHaFUoUhiFXqQwCr1IYRR6kcIo9CKF0Xmxg5S63XLuNtG5y1eHenrymjVrGpY99dRTyWFXrlyZLJ89e3ayPHfZcKo8dwvrnNx8lfq0pRcpjEIvUhiFXqQwCr1IYRR6kcIo9CKFUehFCqN2+ij3uOgXX3yxYVmurTp3aWzuEtCh1G3r1q3JYXOPwc6Vjx49Olk+lLb03HyT5mhLL1IYhV6kMAq9SGEUepHCKPQihVHoRQqj0IsURu30Ua4tPNXenLsePjfuXHlO6vbcqTIY+jkCuXsJjBw5smFZrg1/27ZtyXJpjrb0IoVR6EUKo9CLFEahFymMQi9SGIVepDAKvUhh1E4fDeXe87m28NyjrIcq9Tjp3PXwuUdR5+4VMGvWrGR5Su5R1NIe2S29mV1vZuvMbGlNt0vNbJWZLY6v09tbTRFplcHs3t8AnFqn+9Xufmx83d7aaolIu2RD7+6LgE0dqIuIdMBQDuRdZGYPxt3/yY16MrP5ZtZnZn1DPcdcRIau2dBfCxwFHAusBq5s1KO7L3D3XnfvzV3cISLt11QK3X2tu+929z3AV4HjW1stEWmXpkJvZjNrPp4FLG3Ur4gML9nGaTO7GZgHTDWzlcDngHlmdizgwHLgY22sY0fk2tq7KdeWnrrufMWKFclhN2/enCyfM2dOsjzXTn/ooYc2LMudA5A7BqSfi83Jht7dz6vT+bo21EVEOkCrSpHCKPQihVHoRQqj0IsURqEXKYwurR2kVJNeu08vTj2KGmD8+PENy9atW5cctr+/P1k+ceLEZPmECROS5UcccUTDsp6enuSwOUO9fXeptKUXKYxCL1IYhV6kMAq9SGEUepHCKPQihVHoRQqjdvood5vqdt7GOnfp7NixY5Plqfbq3bt3J4d99tlnk+W5y19zt9hOXXqbuyx32bJlyfJcO33qHILcuQ8H8u25taUXKYxCL1IYhV6kMAq9SGEUepHCKPQihVHoRQqjdvoWGOqjqnPDT506NVk+ZsyYhmW59ujnn38+WT55csMnlgH5a+IPO+ywpsoANm1KP0JxKNfj5661Vzu9iBwwFHqRwij0IoVR6EUKo9CLFEahFymMQi9SmME8qnoOcBMwg/Bo6gXufo2ZTQG+AcwlPK76HHd/rn1V7a7UtdupdnLIX5M+atSoZPm4ceOS5anzAHLt0UuWLEmWr127Nlmeu15/+/btDcuOOuqo5LDPPPNMsnzDhg3J8lTdcucvHMgGs6XfBXzS3Y8B3gJcaGbHABcDC939aGBh/Cwiw1w29O6+2t3vj++3AI8AhwFnAjfG3m4E3teuSopI6+zTb3ozmwscB9wDzHD31bFoDWH3X0SGuUGH3swmAN8E/sDdX6gt8/Cjsu4PSzObb2Z9ZtanZ4uJdN+gQm9mIwmB/2d3/1bsvNbMZsbymUDdJyW6+wJ373X33tyNDEWk/bIptHAJ2HXAI+5+VU3RbcD58f35wLdbXz0RabXBXFr7NuBDwBIzWxy7XQJcBtxiZh8BVgDntKeKw0Pqp0mu+Sd3ae0hhxySLB85cmSyfP369Q3Lnnsu3Yqae5T1woULk+UnnnhisnzNmjUNy7Zs2ZIcNnfr71x5qkkv14x6IO+VZkPv7ncBjS74Prm11RGRdjtwV2ciUpdCL1IYhV6kMAq9SGEUepHCKPQihdEtsKPcKcLbtm1rWJZrL861448fPz5Znrsdc+ry1c2bNyeHzZ0j0NfXlyw/8sgjk+Wp21TnbmF98MEHNz1ugP7+/qbKDnTa0osURqEXKYxCL1IYhV6kMAq9SGEUepHCKPQihVE7fZRrpx8xovGsyrX55h5FnbuNdO4W2KnzAHLTztmxY0eyPHcL7dQ5BLlr1nPXvD/99NPJ8tS5FSXTll6kMAq9SGEUepHCKPQihVHoRQqj0IsURqEXKYza6Vtg9OjRyfLc9fbTp09Plueut0+dQ5BrC89dq5+7nj5Xnjr/IdeOnnsE+FDuTX8g39c+p9xvLlIohV6kMAq9SGEUepHCKPQihVHoRQqj0IsUJttOb2ZzgJuAGYADC9z9GjO7FPhtYODh6Je4++3tqmi7pdq6B1Oekns+/YoVK5oe9/4sd58AaY/BLMm7gE+6+/1mNhG4z8zuiGVXu/sV7aueiLRaNvTuvhpYHd9vMbNHgMPaXTERaY99+k1vZnOB44B7YqeLzOxBM7vezCY3GGa+mfWZWV/ullQi0n6DDr2ZTQC+CfyBu78AXAscBRxL2BO4st5w7r7A3Xvdvbfk851FhotBpdDMRhIC/8/u/i0Ad1/r7rvdfQ/wVeD49lVTRFolG3oLt1O9DnjE3a+q6T6zprezgKWtr56ItNpgjt6/DfgQsMTMFsdulwDnmdmxhGa85cDH2lJDEWmpwRy9vwuod/P0/bZNXqRkOrImUhiFXqQwCr1IYRR6kcIo9CKFUehFCqPQixRGoRcpjEIvUhiFXqQwCr1IYRR6kcIo9CKFUehFCmO52zO3dGJm64Ha+z1PBTZ0rAL7ZrjWbbjWC1S3ZrWybke4+7RUDx0N/asmbtbn7r1dq0DCcK3bcK0XqG7N6nTdtHsvUhiFXqQw3Q79gi5PP2W41m241gtUt2Z1tG5d/U0vIp3X7S29iHSYQi9SmK6E3sxONbPHzOxJM7u4G3VoxMyWm9kSM1tsZn1drsv1ZrbOzJbWdJtiZneY2RPxb91nCHapbpea2ao47xab2eldqtscM/uhmT1sZg+Z2Sdi967Ou0S9OjrfOv6b3sx6gMeBdwIrgXuB89z94Y5WpAEzWw70unvXT+Qws5OAfuAmd//l2O1yYJO7XxZXmJPd/dPDpG6XAv3dfnx5fPrSzNrHqwPvAy6gi/MuUa9z6OB868aW/njgSXdf5u4vAV8HzuxCPYY9d18EbKp0PhO4Mb6/kbDQdFyDug0L7r7a3e+P77cAA49X7+q8S9Sro7oR+sOAZ2o+r2R4Pe/ege+b2X1mNr/blaljhruvju/XADO6WZk6so8v76TK49WHzbxr5rHvraIDea92gru/CTgNuDDuxg5LHn6bDac210E9vrxT6jxe/WXdnHfNPva9VboR+lXAnJrPs2O3YcHdV8W/64BbGX6P4F478MTg+Hddl+vzsuH0+PJ6j1dnGMy74fDY926E/l7gaDN7rZmNAj4A3NaFeryKmY2PB1gws/HAKQy/R3DfBpwf358PfLuLddnLcHl8eaPHq9PleTdsHvvu7h1/AacTjuA/BXymG3VoUK8jgQfi66Fu1w24mbC7t5Nw7OMjwCHAQuAJ4AfAlGFUt68BS4AHCQGb2aW6nUDYdX8QWBxfp3d73iXq1dH5ptNwRQqjA3kihVHoRQqj0IsURqEXKYxCL1IYhV6kMAq9SGH+G8xDuxPSXu63AAAAAElFTkSuQmCC\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Preserve 90%\n",
        "pca_90 = PCA(n_components=nbr_PC(90))\n",
        "mnist_pca_90_reduced = pca_90.fit_transform(data)\n",
        "mnist_pca_90_recovered = pca_90.inverse_transform(mnist_pca_90_reduced)\n",
        "\n",
        "image_pca_90 = mnist_pca_90_recovered[1,:].reshape([28,28])\n",
        "plt.imshow(image_pca_90, cmap='gray')\n",
        "plt.title('Compressed image with 90% ', fontsize=15, pad=15)\n",
        "plt.savefig(\"image_pca_90.png\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 292
        },
        "id": "Cx3A1CtaV-oe",
        "outputId": "97df26a7-8716-4712-f16f-bcedda462c97"
      },
      "id": "Cx3A1CtaV-oe",
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Preserve 95%\n",
        "pca_95 = PCA(n_components=nbr_PC(95))\n",
        "mnist_pca_95_reduced = pca_95.fit_transform(data)\n",
        "mnist_pca_95_recovered = pca_95.inverse_transform(mnist_pca_95_reduced)\n",
        "\n",
        "image_pca_95 = mnist_pca_95_recovered[1,:].reshape([28,28])\n",
        "plt.imshow(image_pca_95, cmap='gray')\n",
        "plt.title('Compressed image with 95%', fontsize=15, pad=15)\n",
        "plt.savefig(\"image_pca_95.png\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 292
        },
        "id": "UKaJu9bDV75D",
        "outputId": "d16fb6d1-ea25-4366-bedf-745ec88c295a"
      },
      "id": "UKaJu9bDV75D",
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Preserve 100%\n",
        "pca_100 = PCA(n_components=nbr_PC(100))\n",
        "mnist_pca_100_reduced = pca_100.fit_transform(data)\n",
        "mnist_pca_100_recovered = pca_100.inverse_transform(mnist_pca_100_reduced)\n",
        "\n",
        "image_pca_100 = mnist_pca_100_recovered[1,:].reshape([28,28])\n",
        "plt.imshow(image_pca_100, cmap='gray')\n",
        "plt.title('Compressed image with 100%', fontsize=15, pad=15)\n",
        "plt.savefig(\"image_pca_100.png\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 292
        },
        "outputId": "7939a613-c83c-4335-9a4b-bcb4620c7c87",
        "id": "1tMIfZ2szsf0"
      },
      "execution_count": 57,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "id": "1tMIfZ2szsf0"
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Store compressed images"
      ],
      "metadata": {
        "id": "cE1FTCcK1I_R"
      },
      "id": "cE1FTCcK1I_R"
    },
    {
      "cell_type": "code",
      "source": [
        "x,y =mnist_pca_80_recovered.shape\n",
        "comCollection = database['mnist-Compressed']\n",
        "for i in range(x):\n",
        "  image = mnist_pca_80_recovered[i,:].tolist()\n",
        "  comCollection.insert_one({ 'image':image, 'numbre':numbre, 'origin': imageIdList[i] })\n",
        "  "
      ],
      "metadata": {
        "id": "F8285oeP1RSk"
      },
      "id": "F8285oeP1RSk",
      "execution_count": 81,
      "outputs": []
    }
  ],
  "metadata": {
    "kernelspec": {
      "display_name": "Python 3 (ipykernel)",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.9.7"
    },
    "colab": {
      "name": "Script.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "7e95804a",
        "kEMNnVaAY74r"
      ]
    }
  },
  "nbformat": 4,
  "nbformat_minor": 5
}