{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "k2PNBliRg61b"
      },
      "outputs": [],
      "source": [
        "\n",
        "# Install required packages\n",
        "\n",
        "!pip install --quiet yfinance joblib tensorflow huggingface_hub scikit-learn\n",
        "\n",
        "\n",
        "# Import Libraries\n",
        "\n",
        "from huggingface_hub import hf_hub_download\n",
        "import tensorflow as tf\n",
        "import joblib\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import yfinance as yf\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "\n",
        "\n",
        "# Hugging Face token\n",
        "\n",
        "HF_TOKEN = \"your_own_hf_token\"  # Replace with your token if private\n",
        "\n",
        "\n",
        "# Download LSTM model & scaler from Hugging Face Hub\n",
        "\n",
        "model_path = hf_hub_download(\n",
        "    repo_id=\"EsferSami/DataSynthis_ML_JobTask\",\n",
        "    filename=\"Apple-Stock-Price-Forecasting-LSTM-Model/apple_stock_lstm.h5\",\n",
        "    token=HF_TOKEN\n",
        ")\n",
        "scaler_path = hf_hub_download(\n",
        "    repo_id=\"EsferSami/DataSynthis_ML_JobTask\",\n",
        "    filename=\"Apple-Stock-Price-Forecasting-LSTM-Model/scaler.joblib\",\n",
        "    token=HF_TOKEN\n",
        ")\n",
        "\n",
        "\n",
        "# Load model & scaler\n",
        "\n",
        "model = tf.keras.models.load_model(model_path)\n",
        "scaler = joblib.load(scaler_path)\n",
        "print(\"LSTM model and scaler loaded successfully!\")\n",
        "\n",
        "\n",
        "# Download recent AAPL data (last 3 months)\n",
        "\n",
        "data = yf.download(\"AAPL\", period=\"3mo\", auto_adjust=False)\n",
        "recent_prices = data['Adj Close'].values.astype(float)\n",
        "print(f\"Downloaded {len(recent_prices)} recent price points\")\n",
        "\n",
        "\n",
        "# Prepare last 60 days for input\n",
        "\n",
        "last_60_days = recent_prices[-60:].reshape(-1, 1)\n",
        "last_60_scaled = scaler.transform(last_60_days)\n",
        "\n",
        "\n",
        "# 7-Day Forecast with daily change smoothing\n",
        "\n",
        "predictions = []\n",
        "current_seq = last_60_scaled.copy()\n",
        "last_price = last_60_days[-1][0]\n",
        "\n",
        "MAX_DAILY_CHANGE = 0.02  # ±2% daily change\n",
        "\n",
        "for day in range(7):\n",
        "    # Predict next day\n",
        "    input_data = current_seq.reshape(1, 60, 1)\n",
        "    pred_scaled = model.predict(input_data, verbose=0)\n",
        "    pred_price_raw = scaler.inverse_transform(pred_scaled)[0][0]\n",
        "\n",
        "    # Smooth prediction to prevent extreme jumps\n",
        "    change = pred_price_raw - last_price\n",
        "    change = np.clip(change, -MAX_DAILY_CHANGE*last_price, MAX_DAILY_CHANGE*last_price)\n",
        "    anchored_price = last_price + change\n",
        "    predictions.append(anchored_price)\n",
        "\n",
        "    # Update sequence for next step\n",
        "    pred_scaled_reshaped = scaler.transform(np.array([[anchored_price]]))\n",
        "    current_seq = np.append(current_seq[1:], pred_scaled_reshaped, axis=0)\n",
        "    last_price = anchored_price\n",
        "\n",
        "\n",
        "# Prepare results dataframe\n",
        "\n",
        "prediction_dates = pd.date_range(start=data.index[-1] + pd.Timedelta(days=1), periods=7)\n",
        "results_df = pd.DataFrame({\n",
        "    'Date': prediction_dates,\n",
        "    'Predicted_Price': np.round(predictions, 2)\n",
        "})\n",
        "\n",
        "\n",
        "# Display 7-Day Forecast\n",
        "\n",
        "print(\"\\nLSTM - 7-Day Forecast\")\n",
        "print(\"=\"*50)\n",
        "print(results_df.to_string(index=False))\n"
      ]
    }
  ]
}