predict_fx.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "predict_fx.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "mount_file_id": "1EPHdyyUrk-uETHR75nDbsy8eRBj0-4EM",
      "authorship_tag": "ABX9TyPrOuPeCPC2h4yMzkOWlmvl",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/tocom242242/simple_fx_predict_app/blob/master/predict_fx.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jWDMr0rYf6fd",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        },
        "outputId": "e55d17ed-1e4f-4f74-c594-e7a0164f9396"
      },
      "source": [
        "!pip install oandapyV20\n",
        "!pip install schedule"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting oandapyV20\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/0d/d5/f69c978f485db3048886c447398e39c6ec18041e689656e6d31697ea4e7f/oandapyV20-0.6.3.tar.gz (52kB)\n",
            "\r\u001b[K     |██████▏                         | 10kB 16.4MB/s eta 0:00:01\r\u001b[K     |████████████▍                   | 20kB 4.1MB/s eta 0:00:01\r\u001b[K     |██████████████████▋             | 30kB 5.5MB/s eta 0:00:01\r\u001b[K     |████████████████████████▉       | 40kB 6.5MB/s eta 0:00:01\r\u001b[K     |███████████████████████████████ | 51kB 6.3MB/s eta 0:00:01\r\u001b[K     |████████████████████████████████| 61kB 3.9MB/s \n",
            "\u001b[?25hBuilding wheels for collected packages: oandapyV20\n",
            "  Building wheel for oandapyV20 (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for oandapyV20: filename=oandapyV20-0.6.3-cp36-none-any.whl size=69742 sha256=5aeafd65bdd4c0d82adc0ce30fa4a6ec801966f17634d3d5dcbbbf72aac5aa45\n",
            "  Stored in directory: /root/.cache/pip/wheels/2c/b0/cd/49054320a6cc51db0ff4f1ee15d8840375f16cef57b81ebfc0\n",
            "Successfully built oandapyV20\n",
            "Installing collected packages: oandapyV20\n",
            "Successfully installed oandapyV20-0.6.3\n",
            "Collecting schedule\n",
            "  Downloading https://files.pythonhosted.org/packages/57/22/3a709462eb02412bd1145f6e53604f36bba191e3e4e397bea4a718fec38c/schedule-0.6.0-py2.py3-none-any.whl\n",
            "Installing collected packages: schedule\n",
            "Successfully installed schedule-0.6.0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gc5ccnQt-6z5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "ACCESS_TOKEN = \"********************************************\"  # あなたのOANDA API のアクセストークン\n",
        "COUNT = 5000        # 一度に取得するデータ数(max:5000)\n",
        "COUNT_TIMES = 1         # count * COUNT_TIMES 分データを取得\n",
        "GRANULARITY = \"M5\"   # 足の種類。今回は５分足\n",
        "INSTRUMENT = \"USD_JPY\"    # USD/JPY\n",
        "INPUT_LEN = 10            # 入力の長さ。現時点から10足前までのデータを入力とする。\n",
        "EPOCHS = 100              # ニューラルネットワークの学習EPOCH\n",
        "BATCH_SIZE = 32           # ニューラルネットワークのバッチサイズ\n",
        "PREDICT_INTERVAL = 1     # 予測の更新頻度 (分)\n",
        "\n",
        "# 学習結果のモデルの保存先\n",
        "MODEL_PATH = \"/content/drive/My Drive/simple_fx_predict_app/models/model.json\"\n",
        "WEIGHTS_PATH = \"/content/drive/My Drive/simple_fx_predict_app/models/weights.h5\"\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZZtOTL4chDXf",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 409
        },
        "outputId": "55aa6d69-dce7-4b25-86f3-4de28ce3fd24"
      },
      "source": [
        "from oandapyV20 import API\n",
        "import oandapyV20.endpoints.instruments as instruments\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# OANDA APIにアクセスする\n",
        "api = API(access_token=ACCESS_TOKEN, environment=\"practice\")\n",
        "\n",
        "def get_prices(granularity, count, count_times):\n",
        "    \"\"\"\n",
        "        終わり値を取得する\n",
        "    \"\"\"\n",
        "    params = {\n",
        "        \"granularity\": granularity,\n",
        "        \"count\": count,\n",
        "        \"price\": \"B\",\n",
        "    }\n",
        "\n",
        "    candles = None\n",
        "    for i in range(count_times):\n",
        "        instruments_candles = instruments.InstrumentsCandles(\n",
        "            instrument=INSTRUMENT, params=params)\n",
        "        # OANDA APIにリクエスト \n",
        "        api.request(instruments_candles)\n",
        "        response = instruments_candles.response\n",
        "        new_candles = pd.DataFrame(response[\"candles\"])\n",
        "        params[\"to\"] = new_candles[\"time\"].iloc[0]\n",
        "        # 取得したデータをつなげる \n",
        "        candles = pd.concat([new_candles, candles])\n",
        "\n",
        "    # 終値だけを抽出する\n",
        "    prices = np.array([x[\"c\"] for x in candles[\"bid\"].values])\n",
        "    prices = prices.astype(np.float64)\n",
        "    prices = prices.reshape(-1, 1)\n",
        "\n",
        "    return prices \n",
        "\n",
        "\n",
        "    \n",
        "prices = get_prices(GRANULARITY, COUNT, COUNT_TIMES)\n",
        "print(prices)\n",
        "\n",
        "# データのプロット\n",
        "fig, ax = plt.subplots()\n",
        "x = np.arange(len(prices))\n",
        "ax.plot(x, prices)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[105.462]\n",
            " [105.517]\n",
            " [105.506]\n",
            " ...\n",
            " [106.078]\n",
            " [106.099]\n",
            " [106.114]]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f0b23a3bdd8>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lCDi97CpicO1",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 197
        },
        "outputId": "e4a29d37-900f-49af-a9ba-292f5b8c1378"
      },
      "source": [
        "from sklearn.preprocessing import MinMaxScaler\n",
        "\n",
        "# データの取得\n",
        "prices = get_prices(GRANULARITY, COUNT, COUNT_TIMES)\n",
        "\n",
        "# データの正規化0〜１の間の数値に\n",
        "sc = MinMaxScaler(feature_range=(0, 1))\n",
        "prices = sc.fit_transform(prices)\n",
        "\n",
        "def create_data_for_model(prices):\n",
        "    \"\"\"\n",
        "        入力 x=[x1,x2,,,] -> 出力 y\n",
        "        といった形で学習・テスト用のデータの作成\n",
        "    \"\"\"\n",
        "\n",
        "    X, Y = [], []\n",
        "    prices_len = len(prices)\n",
        "    for i, _ in enumerate(prices):\n",
        "        if (i + INPUT_LEN + 1 >= prices_len):\n",
        "            break\n",
        "        X.append(prices[i:i + INPUT_LEN])\n",
        "        Y.append(prices[i + INPUT_LEN + 1])\n",
        "\n",
        "    X = np.array(X)\n",
        "    Y = np.array(Y)\n",
        "\n",
        "    # 学習用・テスト用に分割\n",
        "    split_index = int(len(prices) * 0.7)\n",
        "    X_train = X[:split_index]\n",
        "    Y_train = Y[:split_index]\n",
        "    X_test = X[split_index:]\n",
        "    Y_test = Y[split_index:]\n",
        "\n",
        "    return X_train, Y_train, X_test, Y_test\n",
        "\n",
        "# 学習・テスト用データの作成\n",
        "X_train, Y_train, X_test, Y_test = create_data_for_model(prices)\n",
        "\n",
        "# データの確認\n",
        "print(\"x:{} => y:{}\".format(X_train[0], Y_train[0]))\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "x:[[0.44342615]\n",
            " [0.46281283]\n",
            " [0.4589355 ]\n",
            " [0.44765597]\n",
            " [0.45928798]\n",
            " [0.46704265]\n",
            " [0.45541065]\n",
            " [0.43743391]\n",
            " [0.4282693 ]\n",
            " [0.40676771]] => y:[0.40148044]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HUIWknuz_clD",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "46473785-44e8-47ab-edda-9d58d2c59396"
      },
      "source": [
        "import tensorflow\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import Dense, Activation, Flatten\n",
        "import json\n",
        "\n",
        "\n",
        "# モデルの作成\n",
        "model = Sequential()\n",
        "model.add(Flatten(input_shape=[INPUT_LEN, 1]))\n",
        "model.add(Dense(32))\n",
        "model.add(Activation(\"relu\"))\n",
        "model.add(Dense(32))\n",
        "model.add(Activation(\"relu\"))\n",
        "model.add(Dense(1))\n",
        "model.add(Activation(\"linear\"))\n",
        "\n",
        "model.compile(loss=\"mse\",   # 誤差関数\n",
        "              optimizer=\"adam\",     # 最適化手法\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# モデルの学習\n",
        "model.fit(X_train, Y_train, batch_size=BATCH_SIZE, epochs=EPOCHS)\n",
        "\n",
        "# 学習結果の保存\n",
        "# モデル構造の保存\n",
        "model_json = model.to_json()\n",
        "with open(MODEL_PATH, \"w\") as f:\n",
        "    json.dump(model_json, f, indent=4)\n",
        "# モデルの重みの保存\n",
        "model.save_weights(WEIGHTS_PATH)\n",
        "\n",
        "# 評価\n",
        "# 予測値の計算と変換\n",
        "predicted_y = model.predict(X_test)\n",
        "predicted_y = sc.inverse_transform(predicted_y)\n",
        "# 正解データの変換\n",
        "Y_test = sc.inverse_transform(Y_test)\n",
        "\n",
        "# 以下結果のプロット\n",
        "fig, ax = plt.subplots()\n",
        "# 正解データのプロット\n",
        "ax.plot(np.arange(len(Y_test)), Y_test, label=\"real\")\n",
        "# 予測値のプロット\n",
        "ax.plot(np.arange(len(predicted_y)), predicted_y, label=\"predicted_value\")\n",
        "\n",
        "# 凡例の表示\n",
        "ax.legend()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 0.0141 - accuracy: 2.8571e-04\n",
            "Epoch 2/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.9827e-04 - accuracy: 2.8571e-04\n",
            "Epoch 3/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.7306e-04 - accuracy: 2.8571e-04\n",
            "Epoch 4/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.5821e-04 - accuracy: 2.8571e-04\n",
            "Epoch 5/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.6336e-04 - accuracy: 2.8571e-04\n",
            "Epoch 6/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.5246e-04 - accuracy: 2.8571e-04\n",
            "Epoch 7/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.4692e-04 - accuracy: 2.8571e-04\n",
            "Epoch 8/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.5106e-04 - accuracy: 2.8571e-04\n",
            "Epoch 9/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.6753e-04 - accuracy: 2.8571e-04\n",
            "Epoch 10/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.5377e-04 - accuracy: 2.8571e-04\n",
            "Epoch 11/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3538e-04 - accuracy: 2.8571e-04\n",
            "Epoch 12/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.4305e-04 - accuracy: 2.8571e-04\n",
            "Epoch 13/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.4094e-04 - accuracy: 2.8571e-04\n",
            "Epoch 14/100\n",
            "110/110 [==============================] - 0s 2ms/step - loss: 2.3729e-04 - accuracy: 2.8571e-04\n",
            "Epoch 15/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3044e-04 - accuracy: 2.8571e-04\n",
            "Epoch 16/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.4181e-04 - accuracy: 2.8571e-04\n",
            "Epoch 17/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3766e-04 - accuracy: 2.8571e-04\n",
            "Epoch 18/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3030e-04 - accuracy: 2.8571e-04\n",
            "Epoch 19/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3483e-04 - accuracy: 2.8571e-04\n",
            "Epoch 20/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3674e-04 - accuracy: 2.8571e-04\n",
            "Epoch 21/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3988e-04 - accuracy: 2.8571e-04\n",
            "Epoch 22/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3185e-04 - accuracy: 2.8571e-04\n",
            "Epoch 23/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3134e-04 - accuracy: 2.8571e-04\n",
            "Epoch 24/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3720e-04 - accuracy: 2.8571e-04\n",
            "Epoch 25/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3494e-04 - accuracy: 2.8571e-04\n",
            "Epoch 26/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3640e-04 - accuracy: 2.8571e-04\n",
            "Epoch 27/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3289e-04 - accuracy: 2.8571e-04\n",
            "Epoch 28/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3911e-04 - accuracy: 2.8571e-04\n",
            "Epoch 29/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3418e-04 - accuracy: 2.8571e-04\n",
            "Epoch 30/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2266e-04 - accuracy: 2.8571e-04\n",
            "Epoch 31/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2136e-04 - accuracy: 2.8571e-04\n",
            "Epoch 32/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3145e-04 - accuracy: 2.8571e-04\n",
            "Epoch 33/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.7422e-04 - accuracy: 2.8571e-04\n",
            "Epoch 34/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2855e-04 - accuracy: 2.8571e-04\n",
            "Epoch 35/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2596e-04 - accuracy: 2.8571e-04\n",
            "Epoch 36/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2571e-04 - accuracy: 2.8571e-04\n",
            "Epoch 37/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2544e-04 - accuracy: 2.8571e-04\n",
            "Epoch 38/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1693e-04 - accuracy: 2.8571e-04\n",
            "Epoch 39/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1873e-04 - accuracy: 2.8571e-04\n",
            "Epoch 40/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2814e-04 - accuracy: 2.8571e-04\n",
            "Epoch 41/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3218e-04 - accuracy: 2.8571e-04\n",
            "Epoch 42/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3665e-04 - accuracy: 2.8571e-04\n",
            "Epoch 43/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2358e-04 - accuracy: 2.8571e-04\n",
            "Epoch 44/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3313e-04 - accuracy: 2.8571e-04\n",
            "Epoch 45/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2084e-04 - accuracy: 2.8571e-04\n",
            "Epoch 46/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1944e-04 - accuracy: 2.8571e-04\n",
            "Epoch 47/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2528e-04 - accuracy: 2.8571e-04\n",
            "Epoch 48/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2753e-04 - accuracy: 2.8571e-04\n",
            "Epoch 49/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.5033e-04 - accuracy: 2.8571e-04\n",
            "Epoch 50/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3992e-04 - accuracy: 2.8571e-04\n",
            "Epoch 51/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2624e-04 - accuracy: 2.8571e-04\n",
            "Epoch 52/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2385e-04 - accuracy: 2.8571e-04\n",
            "Epoch 53/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2933e-04 - accuracy: 2.8571e-04\n",
            "Epoch 54/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1981e-04 - accuracy: 2.8571e-04\n",
            "Epoch 55/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2494e-04 - accuracy: 2.8571e-04\n",
            "Epoch 56/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1077e-04 - accuracy: 2.8571e-04\n",
            "Epoch 57/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1859e-04 - accuracy: 2.8571e-04\n",
            "Epoch 58/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2743e-04 - accuracy: 2.8571e-04\n",
            "Epoch 59/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3503e-04 - accuracy: 2.8571e-04\n",
            "Epoch 60/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2788e-04 - accuracy: 2.8571e-04\n",
            "Epoch 61/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2878e-04 - accuracy: 2.8571e-04\n",
            "Epoch 62/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.5811e-04 - accuracy: 2.8571e-04\n",
            "Epoch 63/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.4555e-04 - accuracy: 2.8571e-04\n",
            "Epoch 64/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1966e-04 - accuracy: 2.8571e-04\n",
            "Epoch 65/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1623e-04 - accuracy: 2.8571e-04\n",
            "Epoch 66/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2330e-04 - accuracy: 2.8571e-04\n",
            "Epoch 67/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2358e-04 - accuracy: 2.8571e-04\n",
            "Epoch 68/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2262e-04 - accuracy: 2.8571e-04\n",
            "Epoch 69/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2610e-04 - accuracy: 2.8571e-04\n",
            "Epoch 70/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3293e-04 - accuracy: 2.8571e-04\n",
            "Epoch 71/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2158e-04 - accuracy: 2.8571e-04\n",
            "Epoch 72/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1498e-04 - accuracy: 2.8571e-04\n",
            "Epoch 73/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.4305e-04 - accuracy: 2.8571e-04\n",
            "Epoch 74/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.7072e-04 - accuracy: 2.8571e-04\n",
            "Epoch 75/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1290e-04 - accuracy: 2.8571e-04\n",
            "Epoch 76/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2496e-04 - accuracy: 2.8571e-04\n",
            "Epoch 77/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1944e-04 - accuracy: 2.8571e-04\n",
            "Epoch 78/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.3824e-04 - accuracy: 2.8571e-04\n",
            "Epoch 79/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1838e-04 - accuracy: 2.8571e-04\n",
            "Epoch 80/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1333e-04 - accuracy: 2.8571e-04\n",
            "Epoch 81/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1569e-04 - accuracy: 2.8571e-04\n",
            "Epoch 82/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2314e-04 - accuracy: 2.8571e-04\n",
            "Epoch 83/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2116e-04 - accuracy: 2.8571e-04\n",
            "Epoch 84/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.4281e-04 - accuracy: 2.8571e-04\n",
            "Epoch 85/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.5186e-04 - accuracy: 2.8571e-04\n",
            "Epoch 86/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2426e-04 - accuracy: 2.8571e-04\n",
            "Epoch 87/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2952e-04 - accuracy: 2.8571e-04\n",
            "Epoch 88/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1779e-04 - accuracy: 2.8571e-04\n",
            "Epoch 89/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1169e-04 - accuracy: 2.8571e-04\n",
            "Epoch 90/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2930e-04 - accuracy: 2.8571e-04\n",
            "Epoch 91/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1674e-04 - accuracy: 2.8571e-04\n",
            "Epoch 92/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1868e-04 - accuracy: 2.8571e-04\n",
            "Epoch 93/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2066e-04 - accuracy: 2.8571e-04\n",
            "Epoch 94/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.2479e-04 - accuracy: 2.8571e-04\n",
            "Epoch 95/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1522e-04 - accuracy: 2.8571e-04\n",
            "Epoch 96/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1063e-04 - accuracy: 2.8571e-04\n",
            "Epoch 97/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.1886e-04 - accuracy: 2.8571e-04\n",
            "Epoch 98/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.0918e-04 - accuracy: 2.8571e-04\n",
            "Epoch 99/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.0985e-04 - accuracy: 2.8571e-04\n",
            "Epoch 100/100\n",
            "110/110 [==============================] - 0s 1ms/step - loss: 2.4221e-04 - accuracy: 2.8571e-04\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "error",
          "ename": "KeyboardInterrupt",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-8-4ca214ef2765>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     25\u001b[0m \u001b[0;31m# モデル構造の保存\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     26\u001b[0m \u001b[0mmodel_json\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mto_json\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 27\u001b[0;31m \u001b[0;32mwith\u001b[0m \u001b[0mopen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mMODEL_PATH\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"w\"\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     28\u001b[0m     \u001b[0mjson\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdump\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel_json\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mindent\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m4\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     29\u001b[0m \u001b[0;31m# モデルの重みの保存\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/lib/python3.6/_bootlocale.py\u001b[0m in \u001b[0;36mgetpreferredencoding\u001b[0;34m(do_setlocale)\u001b[0m\n\u001b[1;32m     21\u001b[0m             \u001b[0;32mreturn\u001b[0m \u001b[0mlocale\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgetpreferredencoding\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdo_setlocale\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     22\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 23\u001b[0;31m         \u001b[0;32mdef\u001b[0m \u001b[0mgetpreferredencoding\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdo_setlocale\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     24\u001b[0m             \u001b[0;32massert\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mdo_setlocale\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     25\u001b[0m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_locale\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnl_langinfo\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0m_locale\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mCODESET\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oGIxuWNP_CqO",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 722
        },
        "outputId": "85b261a5-28e0-45e3-bdc4-59f007cb57f1"
      },
      "source": [
        "import schedule\n",
        "import time\n",
        "from IPython.display import display, clear_output\n",
        "\n",
        "# 学習済みモデルの読み込み\n",
        "model_json = json.load(open(MODEL_PATH))\n",
        "model = tensorflow.keras.models.model_from_json(model_json)\n",
        "model.load_weights(WEIGHTS_PATH)\n",
        "\n",
        "plt.ion()\n",
        "fig, ax = plt.subplots()\n",
        "\n",
        "\n",
        "def predict():\n",
        "    prices = get_prices(GRANULARITY, 100, 1)\n",
        "    #sc = MinMaxScaler(feature_range=(0, 1))\n",
        "    prices_for_input = sc.fit_transform(prices)\n",
        "    # 実際の値と評価値を比較する\n",
        "    input_value = np.array([prices_for_input[-INPUT_LEN:]])\n",
        "    # 予想値\n",
        "    predicted_value = model.predict(input_value)\n",
        "    predicted_value = sc.inverse_transform(predicted_value)[0]\n",
        "\n",
        "    x = np.arange(len(prices))\n",
        "    # 実際の値のプロット\n",
        "    clear_output(wait=True)\n",
        "    ax.clear()\n",
        "    ax.plot(x, prices, label=\"real\")\n",
        "    ax.scatter([x[-1] + 1], predicted_value,\n",
        "               color=\"r\", label=\"predicted_value\")\n",
        "    ax.legend()\n",
        "    display(fig)\n",
        "    print(\"predicted_value:\", predicted_value)\n",
        "\n",
        "# 定期的に実行する時間とメソッドを設定\n",
        "schedule.every(PREDICT_INTERVAL).minutes.do(predict)\n",
        "\n",
        "# 定期的に実行\n",
        "while True:\n",
        "    schedule.run_pending()\n",
        "    time.sleep(1)\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "predicted_value: [105.44646]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "error",
          "ename": "KeyboardInterrupt",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-18-bfbc6f825a9d>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     39\u001b[0m \u001b[0;32mwhile\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     40\u001b[0m     \u001b[0mschedule\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrun_pending\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 41\u001b[0;31m     \u001b[0mtime\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msleep\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}