UNetLIke_pianorolls.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "authorship_tag": "ABX9TyNKYE/YV1nOEqjdeHwmVPKn",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/hdperezn/MIDI_functions/blob/main/UNetLIke_pianorolls.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "P1QhyS-a7n7b"
      },
      "source": [
        "#setup"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qLcrvFqm6CUi",
        "outputId": "8bb7196c-d34f-41b3-c624-c64d054b7d4f"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ],
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install -U git+https://github.com/UN-GCPDS/python-gcpds.image_segmentation.git\n",
        "from gcpds.image_segmentation.losses import DiceCoefficient"
      ],
      "metadata": {
        "id": "ObrteJhg0_lF",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a75a8e83-11fb-41d5-cd2d-8b915c81d524"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Collecting git+https://github.com/UN-GCPDS/python-gcpds.image_segmentation.git\n",
            "  Cloning https://github.com/UN-GCPDS/python-gcpds.image_segmentation.git to /tmp/pip-req-build-7l6vgcn0\n",
            "  Running command git clone --filter=blob:none --quiet https://github.com/UN-GCPDS/python-gcpds.image_segmentation.git /tmp/pip-req-build-7l6vgcn0\n",
            "  Resolved https://github.com/UN-GCPDS/python-gcpds.image_segmentation.git to commit 537c809aed21aba452a28716370b343501633e3c\n",
            "  Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "Requirement already satisfied: scikit-image in /usr/local/lib/python3.9/dist-packages (from gcpds-image-segmentation==0.1a0) (0.19.3)\n",
            "Requirement already satisfied: matplotlib in /usr/local/lib/python3.9/dist-packages (from gcpds-image-segmentation==0.1a0) (3.7.1)\n",
            "Requirement already satisfied: gdown in /usr/local/lib/python3.9/dist-packages (from gcpds-image-segmentation==0.1a0) (4.4.0)\n",
            "Requirement already satisfied: opencv-python in /usr/local/lib/python3.9/dist-packages (from gcpds-image-segmentation==0.1a0) (4.6.0.66)\n",
            "Requirement already satisfied: scikit-learn in /usr/local/lib/python3.9/dist-packages (from gcpds-image-segmentation==0.1a0) (1.2.2)\n",
            "Requirement already satisfied: filelock in /usr/local/lib/python3.9/dist-packages (from gdown->gcpds-image-segmentation==0.1a0) (3.9.1)\n",
            "Requirement already satisfied: tqdm in /usr/local/lib/python3.9/dist-packages (from gdown->gcpds-image-segmentation==0.1a0) (4.65.0)\n",
            "Requirement already satisfied: six in /usr/local/lib/python3.9/dist-packages (from gdown->gcpds-image-segmentation==0.1a0) (1.15.0)\n",
            "Requirement already satisfied: requests[socks] in /usr/local/lib/python3.9/dist-packages (from gdown->gcpds-image-segmentation==0.1a0) (2.25.1)\n",
            "Requirement already satisfied: beautifulsoup4 in /usr/local/lib/python3.9/dist-packages (from gdown->gcpds-image-segmentation==0.1a0) (4.9.3)\n",
            "Requirement already satisfied: python-dateutil>=2.7 in /usr/local/lib/python3.9/dist-packages (from matplotlib->gcpds-image-segmentation==0.1a0) (2.8.2)\n",
            "Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.9/dist-packages (from matplotlib->gcpds-image-segmentation==0.1a0) (23.0)\n",
            "Requirement already satisfied: pyparsing>=2.3.1 in /usr/local/lib/python3.9/dist-packages (from matplotlib->gcpds-image-segmentation==0.1a0) (3.0.9)\n",
            "Requirement already satisfied: numpy>=1.20 in /usr/local/lib/python3.9/dist-packages (from matplotlib->gcpds-image-segmentation==0.1a0) (1.22.4)\n",
            "Requirement already satisfied: importlib-resources>=3.2.0 in /usr/local/lib/python3.9/dist-packages (from matplotlib->gcpds-image-segmentation==0.1a0) (5.12.0)\n",
            "Requirement already satisfied: pillow>=6.2.0 in /usr/local/lib/python3.9/dist-packages (from matplotlib->gcpds-image-segmentation==0.1a0) (8.4.0)\n",
            "Requirement already satisfied: fonttools>=4.22.0 in /usr/local/lib/python3.9/dist-packages (from matplotlib->gcpds-image-segmentation==0.1a0) (4.39.0)\n",
            "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.9/dist-packages (from matplotlib->gcpds-image-segmentation==0.1a0) (1.4.4)\n",
            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.9/dist-packages (from matplotlib->gcpds-image-segmentation==0.1a0) (0.11.0)\n",
            "Requirement already satisfied: contourpy>=1.0.1 in /usr/local/lib/python3.9/dist-packages (from matplotlib->gcpds-image-segmentation==0.1a0) (1.0.7)\n",
            "Requirement already satisfied: imageio>=2.4.1 in /usr/local/lib/python3.9/dist-packages (from scikit-image->gcpds-image-segmentation==0.1a0) (2.9.0)\n",
            "Requirement already satisfied: tifffile>=2019.7.26 in /usr/local/lib/python3.9/dist-packages (from scikit-image->gcpds-image-segmentation==0.1a0) (2023.2.28)\n",
            "Requirement already satisfied: networkx>=2.2 in /usr/local/lib/python3.9/dist-packages (from scikit-image->gcpds-image-segmentation==0.1a0) (3.0)\n",
            "Requirement already satisfied: scipy>=1.4.1 in /usr/local/lib/python3.9/dist-packages (from scikit-image->gcpds-image-segmentation==0.1a0) (1.10.1)\n",
            "Requirement already satisfied: PyWavelets>=1.1.1 in /usr/local/lib/python3.9/dist-packages (from scikit-image->gcpds-image-segmentation==0.1a0) (1.4.1)\n",
            "Requirement already satisfied: joblib>=1.1.1 in /usr/local/lib/python3.9/dist-packages (from scikit-learn->gcpds-image-segmentation==0.1a0) (1.1.1)\n",
            "Requirement already satisfied: threadpoolctl>=2.0.0 in /usr/local/lib/python3.9/dist-packages (from scikit-learn->gcpds-image-segmentation==0.1a0) (3.1.0)\n",
            "Requirement already satisfied: zipp>=3.1.0 in /usr/local/lib/python3.9/dist-packages (from importlib-resources>=3.2.0->matplotlib->gcpds-image-segmentation==0.1a0) (3.15.0)\n",
            "Requirement already satisfied: soupsieve>1.2 in /usr/local/lib/python3.9/dist-packages (from beautifulsoup4->gdown->gcpds-image-segmentation==0.1a0) (2.4)\n",
            "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.9/dist-packages (from requests[socks]->gdown->gcpds-image-segmentation==0.1a0) (2.10)\n",
            "Requirement already satisfied: urllib3<1.27,>=1.21.1 in /usr/local/lib/python3.9/dist-packages (from requests[socks]->gdown->gcpds-image-segmentation==0.1a0) (1.26.15)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.9/dist-packages (from requests[socks]->gdown->gcpds-image-segmentation==0.1a0) (2022.12.7)\n",
            "Requirement already satisfied: chardet<5,>=3.0.2 in /usr/local/lib/python3.9/dist-packages (from requests[socks]->gdown->gcpds-image-segmentation==0.1a0) (4.0.0)\n",
            "Requirement already satisfied: PySocks!=1.5.7,>=1.5.6 in /usr/local/lib/python3.9/dist-packages (from requests[socks]->gdown->gcpds-image-segmentation==0.1a0) (1.7.1)\n",
            "Building wheels for collected packages: gcpds-image-segmentation\n",
            "  Building wheel for gcpds-image-segmentation (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for gcpds-image-segmentation: filename=gcpds_image_segmentation-0.1a0-py3-none-any.whl size=33684 sha256=1d19e4ca9451222d7fd0c79ee44040fce9941c056d3a4db04bd2c3eb3fe1554d\n",
            "  Stored in directory: /tmp/pip-ephem-wheel-cache-mr5bp7sx/wheels/93/3d/96/2bc458dca20d46cfdf8f50c9871f921c847eb0324c71eb3f60\n",
            "Successfully built gcpds-image-segmentation\n",
            "Installing collected packages: gcpds-image-segmentation\n",
            "Successfully installed gcpds-image-segmentation-0.1a0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!git clone https://github.com/hdperezn/MIDI_functions.git\n",
        "!sudo apt install -y fluidsynth\n",
        "!pip install --upgrade pyfluidsynth\n",
        "!pip install pretty_midi"
      ],
      "metadata": {
        "id": "RuI5UccgKx29",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "926ba42a-8a87-4a85-e7db-7011f7675e61"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Cloning into 'MIDI_functions'...\n",
            "remote: Enumerating objects: 118, done.\u001b[K\n",
            "remote: Counting objects: 100% (118/118), done.\u001b[K\n",
            "remote: Compressing objects: 100% (103/103), done.\u001b[K\n",
            "remote: Total 118 (delta 33), reused 73 (delta 10), pack-reused 0\u001b[K\n",
            "Receiving objects: 100% (118/118), 6.54 MiB | 12.57 MiB/s, done.\n",
            "Resolving deltas: 100% (33/33), done.\n",
            "Reading package lists... Done\n",
            "Building dependency tree       \n",
            "Reading state information... Done\n",
            "The following additional packages will be installed:\n",
            "  fluid-soundfont-gm libfluidsynth2 libinstpatch-1.0-2 qsynth\n",
            "  timgm6mb-soundfont\n",
            "Suggested packages:\n",
            "  fluid-soundfont-gs timidity jackd musescore\n",
            "The following NEW packages will be installed:\n",
            "  fluid-soundfont-gm fluidsynth libfluidsynth2 libinstpatch-1.0-2 qsynth\n",
            "  timgm6mb-soundfont\n",
            "0 upgraded, 6 newly installed, 0 to remove and 23 not upgraded.\n",
            "Need to get 126 MB of archives.\n",
            "After this operation, 157 MB of additional disk space will be used.\n",
            "Get:1 http://archive.ubuntu.com/ubuntu focal/universe amd64 fluid-soundfont-gm all 3.1-5.1 [119 MB]\n",
            "Get:2 http://archive.ubuntu.com/ubuntu focal/universe amd64 libinstpatch-1.0-2 amd64 1.1.2-2build1 [238 kB]\n",
            "Get:3 http://archive.ubuntu.com/ubuntu focal/universe amd64 timgm6mb-soundfont all 1.3-3 [5,420 kB]\n",
            "Get:4 http://archive.ubuntu.com/ubuntu focal/universe amd64 libfluidsynth2 amd64 2.1.1-2 [198 kB]\n",
            "Get:5 http://archive.ubuntu.com/ubuntu focal/universe amd64 fluidsynth amd64 2.1.1-2 [25.6 kB]\n",
            "Get:6 http://archive.ubuntu.com/ubuntu focal/universe amd64 qsynth amd64 0.6.1-1build1 [245 kB]\n",
            "Fetched 126 MB in 6s (21.0 MB/s)\n",
            "debconf: unable to initialize frontend: Dialog\n",
            "debconf: (No usable dialog-like program is installed, so the dialog based frontend cannot be used. at /usr/share/perl5/Debconf/FrontEnd/Dialog.pm line 76, <> line 6.)\n",
            "debconf: falling back to frontend: Readline\n",
            "debconf: unable to initialize frontend: Readline\n",
            "debconf: (This frontend requires a controlling tty.)\n",
            "debconf: falling back to frontend: Teletype\n",
            "dpkg-preconfigure: unable to re-open stdin: \n",
            "Selecting previously unselected package fluid-soundfont-gm.\n",
            "(Reading database ... 128276 files and directories currently installed.)\n",
            "Preparing to unpack .../0-fluid-soundfont-gm_3.1-5.1_all.deb ...\n",
            "Unpacking fluid-soundfont-gm (3.1-5.1) ...\n",
            "Selecting previously unselected package libinstpatch-1.0-2:amd64.\n",
            "Preparing to unpack .../1-libinstpatch-1.0-2_1.1.2-2build1_amd64.deb ...\n",
            "Unpacking libinstpatch-1.0-2:amd64 (1.1.2-2build1) ...\n",
            "Selecting previously unselected package timgm6mb-soundfont.\n",
            "Preparing to unpack .../2-timgm6mb-soundfont_1.3-3_all.deb ...\n",
            "Unpacking timgm6mb-soundfont (1.3-3) ...\n",
            "Selecting previously unselected package libfluidsynth2:amd64.\n",
            "Preparing to unpack .../3-libfluidsynth2_2.1.1-2_amd64.deb ...\n",
            "Unpacking libfluidsynth2:amd64 (2.1.1-2) ...\n",
            "Selecting previously unselected package fluidsynth.\n",
            "Preparing to unpack .../4-fluidsynth_2.1.1-2_amd64.deb ...\n",
            "Unpacking fluidsynth (2.1.1-2) ...\n",
            "Selecting previously unselected package qsynth.\n",
            "Preparing to unpack .../5-qsynth_0.6.1-1build1_amd64.deb ...\n",
            "Unpacking qsynth (0.6.1-1build1) ...\n",
            "Setting up fluid-soundfont-gm (3.1-5.1) ...\n",
            "Setting up timgm6mb-soundfont (1.3-3) ...\n",
            "update-alternatives: using /usr/share/sounds/sf2/TimGM6mb.sf2 to provide /usr/share/sounds/sf2/default-GM.sf2 (default-GM.sf2) in auto mode\n",
            "update-alternatives: using /usr/share/sounds/sf2/TimGM6mb.sf2 to provide /usr/share/sounds/sf3/default-GM.sf3 (default-GM.sf3) in auto mode\n",
            "Setting up libinstpatch-1.0-2:amd64 (1.1.2-2build1) ...\n",
            "Setting up libfluidsynth2:amd64 (2.1.1-2) ...\n",
            "Setting up qsynth (0.6.1-1build1) ...\n",
            "Setting up fluidsynth (2.1.1-2) ...\n",
            "Created symlink /etc/systemd/user/multi-user.target.wants/fluidsynth.service → /usr/lib/systemd/user/fluidsynth.service.\n",
            "Processing triggers for hicolor-icon-theme (0.17-2) ...\n",
            "Processing triggers for libc-bin (2.31-0ubuntu9.9) ...\n",
            "Processing triggers for man-db (2.9.1-1) ...\n",
            "Processing triggers for mime-support (3.64ubuntu1) ...\n",
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Collecting pyfluidsynth\n",
            "  Downloading pyFluidSynth-1.3.2-py3-none-any.whl (19 kB)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.9/dist-packages (from pyfluidsynth) (1.22.4)\n",
            "Installing collected packages: pyfluidsynth\n",
            "Successfully installed pyfluidsynth-1.3.2\n",
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Collecting pretty_midi\n",
            "  Downloading pretty_midi-0.2.10.tar.gz (5.6 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m5.6/5.6 MB\u001b[0m \u001b[31m39.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "Requirement already satisfied: numpy>=1.7.0 in /usr/local/lib/python3.9/dist-packages (from pretty_midi) (1.22.4)\n",
            "Collecting mido>=1.1.16\n",
            "  Downloading mido-1.2.10-py2.py3-none-any.whl (51 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m51.1/51.1 KB\u001b[0m \u001b[31m5.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: six in /usr/local/lib/python3.9/dist-packages (from pretty_midi) (1.15.0)\n",
            "Building wheels for collected packages: pretty_midi\n",
            "  Building wheel for pretty_midi (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for pretty_midi: filename=pretty_midi-0.2.10-py3-none-any.whl size=5592303 sha256=85aa625a19b5e4b7b7da3f997f67c4b58786c90c1db4d6566a67438add3942e8\n",
            "  Stored in directory: /root/.cache/pip/wheels/75/ec/20/b8e937a5bcf1de547ea5ce465db7de7f6761e15e6f0a01e25f\n",
            "Successfully built pretty_midi\n",
            "Installing collected packages: mido, pretty_midi\n",
            "Successfully installed mido-1.2.10 pretty_midi-0.2.10\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "I6PzUnml8yaF"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "\n",
        "import fluidsynth\n",
        "import glob\n",
        "import pathlib\n",
        "import pretty_midi\n",
        "\n",
        "from IPython import display\n",
        "from MIDI_functions.functions.MIDI_functions import cut_midi_secTrial\n",
        "from MIDI_functions.functions.UNetLike_PianoRolls import UNet_Pianoroll\n",
        "\n",
        "# from tensorflow.keras import layers\n",
        "# from tensorflow.keras.regularizers import l1_l2\n",
        "#from sklearn.base import  BaseEstimator, TransformerMixin, ClassifierMixin\n"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Create Train test files "
      ],
      "metadata": {
        "id": "vIxDNt0qPS4W"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "paths from the MIDI data"
      ],
      "metadata": {
        "id": "DGI2mEuMhrcG"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "path_midi = '/content/MIDI_functions/deap_midis'\n",
        "data_dir = pathlib.Path(path_midi)\n",
        "filenames = glob.glob(str(data_dir/'*.mid*'))\n",
        "print('Number of files:', len(filenames))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9Ly_n-o0huDZ",
        "outputId": "1932587d-ba9f-4c20-f309-c2798b437d0a"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Number of files: 40\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Sampling rate (_SAMPLING_RATE) for audio playback and frequency (fProll) to create piano rolls\n"
      ],
      "metadata": {
        "id": "VVQPHmOmh0-v"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "_SAMPLING_RATE = 16000\n",
        "fProll = 20"
      ],
      "metadata": {
        "id": "Rn6yXUGCP0uf"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "windowed piano roll arrays"
      ],
      "metadata": {
        "id": "0g3E2k0Rh9Pz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "Prolls_windowed = []\n",
        "for i in range(len(filenames)):   \n",
        "  #load MIDI and create array\n",
        "  Proll = pretty_midi.PrettyMIDI(filenames[i]).get_piano_roll(fs=fProll) \n",
        "  if i in [0, 13, 16, 21, 39]:\n",
        "    Proll_cut = cut_midi_secTrial(Proll, 9)\n",
        "  elif i in [9, 11]:\n",
        "    Proll_cut = cut_midi_secTrial(Proll, 8)\n",
        "  else:\n",
        "    Proll_cut = cut_midi_secTrial(Proll, 10)     \n",
        "  Prolls_windowed.append(Proll_cut) "
      ],
      "metadata": {
        "id": "ak57Ao0NPxK1"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "calculate te mode to cut any sample at the same Length\n"
      ],
      "metadata": {
        "id": "F_faNUcMiISt"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "midi_lens = []\n",
        "for i in range(len(Prolls_windowed)):\n",
        "  midi_lens.append(Prolls_windowed[i].shape[2])\n",
        "from statistics import mode\n",
        "mode(midi_lens)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fMMxW107QtET",
        "outputId": "43c8f59a-d615-4700-e990-5794ab86570c"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "119"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Create array with equal dims 128 padding last samples"
      ],
      "metadata": {
        "id": "POU4XtDHiMtV"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "Prolls_cut = []\n",
        "Prolls_trial_len = 128 #64, 59\n",
        "for i in range(len(Prolls_windowed)):\n",
        "  if Prolls_windowed[i].shape[2] >= Prolls_trial_len:\n",
        "    x = Prolls_windowed[i][:,:,0:Prolls_trial_len]\n",
        "    Prolls_cut.append(x)\n",
        "  else:\n",
        "    dif = Prolls_trial_len - Prolls_windowed[i].shape[2]\n",
        "    x = np.concatenate( (Prolls_windowed[i][:,:,:],Prolls_windowed[i][:,:,-dif::] ), axis = 2)\n",
        "    Prolls_cut.append(x)\n"
      ],
      "metadata": {
        "id": "un_WcP0HQpoc"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "song = 2\n",
        "\n",
        "plt.figure()\n",
        "plt.pcolormesh(Prolls_windowed[song][0,:,:])\n",
        "plt.colorbar()\n",
        "plt.ylim(20,80)\n",
        "plt.title(\"Piano roll windowed\")\n",
        "plt.show()\n",
        "\n",
        "plt.figure()\n",
        "plt.pcolormesh(Prolls_cut[song][0,:,:])\n",
        "plt.colorbar()\n",
        "plt.ylim(20,80)\n",
        "plt.title(\"Piano roll windowed and padding\")\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "LdKFcKXYSfRF",
        "outputId": "a527d25f-d169-4660-d105-e56778916335"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Create a training array with dims (samples, time, notes), create raget mask   and spliting in traint and test sets\n"
      ],
      "metadata": {
        "id": "Mjo4PVQPim9z"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "Prolls_array = np.concatenate(np.asarray(Prolls_cut), axis = 0)\n",
        "Prolls_array_mask = np.where(Prolls_array>0,1,0)\n",
        "print('these are de dims of the piano roll array to use in the neural network: ', Prolls_array.shape)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8NWu8WW_VFck",
        "outputId": "8a9dd794-b0b6-4862-88e2-ce923e3cf3b4"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "these are de dims of the piano roll array to use in the neural network:  (391, 128, 128)\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "<ipython-input-11-96c1be7432f6>:1: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray.\n",
            "  Prolls_array = np.concatenate(np.asarray(Prolls_cut), axis = 0)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#train_test_data\n",
        "\n",
        "X_train = Prolls_array[0:313,8:88,:,np.newaxis].astype('float32').transpose(0,2,1,3)\n",
        "X_test = Prolls_array[313::,8:88,:, np.newaxis].astype('float32').transpose(0,2,1,3)\n",
        "\n",
        "\n",
        "y_train = Prolls_array_mask[0:313,8:88,:,np.newaxis].astype('float32').transpose(0,2,1,3)\n",
        "y_test = Prolls_array_mask[313::,8:88,:,np.newaxis].astype('float32').transpose(0,2,1,3)\n",
        "\n",
        "\n",
        "X_dims = X_train.shape\n",
        "print(X_dims)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zySTW6UBxqMv",
        "outputId": "71926e5e-2254-49df-aa76-75bd5de589d9"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(313, 128, 80, 1)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Load Model"
      ],
      "metadata": {
        "id": "W2yYsBZJjeGW"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "creating a model \"UNet_pianoroll\" and load the weigths"
      ],
      "metadata": {
        "id": "g6p416NDo65K"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "img_size = X_dims[-3::]\n",
        "num_classes = 1\n",
        "loss = DiceCoefficient()\n",
        "\n",
        "UNet_pianoroll = UNet_Pianoroll(img_size, loss, num_classes, epochs=250,batch_size=32,\n",
        "               learning_rate=1e-3, droprate = 0.6)"
      ],
      "metadata": {
        "id": "TO9O9yW5jiyf"
      },
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "UNet_pianoroll.get_model()\n",
        "\n",
        "UNet_pianoroll.model.load_weights('/content/MIDI_functions/Unet_piano_rolls_models/UnetLike_PianoRoll_3encoding.h5')\n",
        "UNet_pianoroll.model.summary()"
      ],
      "metadata": {
        "id": "wNhYC8PPl5_Q",
        "outputId": "e9cf6101-b343-425f-a892-52dcef04a774",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"UNet_MIDI\"\n",
            "__________________________________________________________________________________________________\n",
            " Layer (type)                   Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            " input_1 (InputLayer)           [(None, 128, 80, 1)  0           []                               \n",
            "                                ]                                                                 \n",
            "                                                                                                  \n",
            " encoder (Functional)           [(None, 16, 10, 128  46112       ['input_1[0][0]']                \n",
            "                                ),                                                                \n",
            "                                 (None, 16, 10, 128                                               \n",
            "                                )]                                                                \n",
            "                                                                                                  \n",
            " Decoder (Functional)           (None, 128, 80, 1)   462529      ['encoder[0][0]',                \n",
            "                                                                  'encoder[0][1]']                \n",
            "                                                                                                  \n",
            "==================================================================================================\n",
            "Total params: 508,641\n",
            "Trainable params: 506,913\n",
            "Non-trainable params: 1,728\n",
            "__________________________________________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "prediction and plots"
      ],
      "metadata": {
        "id": "ZZYrpBnNpDis"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X_hat_test = UNet_pianoroll.predict(X_test[:,:,:])"
      ],
      "metadata": {
        "id": "8tr6tqZtpNTl",
        "outputId": "63869c1e-b229-47a3-f7f1-f95844e89bc1",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "3/3 [==============================] - 6s 1s/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "sample = 30\n",
        "result_array = X_test[sample,:,:,0]\n",
        "plt.figure()\n",
        "plt.pcolormesh(result_array.T)\n",
        "plt.colorbar()\n",
        "plt.title(\"Orginal\")\n",
        "plt.show()\n",
        "\n",
        "\n",
        "result_array = X_hat_test[sample,:,:,0]\n",
        "plt.figure()\n",
        "plt.pcolormesh(result_array.T)\n",
        "plt.colorbar()\n",
        "plt.title(\"Prediction\")\n",
        "plt.show()\n"
      ],
      "metadata": {
        "id": "8B66spizpSUu",
        "outputId": "e18d6721-4cbf-4ca1-ec45-50bd6e30ae9b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 545
        }
      },
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}