{"cells":[{"metadata":{"_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","trusted":true},"cell_type":"code","source":"import tensorflow as tf\nimport librosa\nimport numpy as np\nimport pandas as pd\nfrom sklearn.model_selection import GroupKFold , StratifiedKFold","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"df = pd.read_csv('../input/rfcx-species-audio-detection/train_tp.csv')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"def _bytes_feature(value):\n  \"\"\"Returns a bytes_list from a string / byte.\"\"\"\n  if isinstance(value, type(tf.constant(0))):\n    value = value.numpy() # BytesList won't unpack a string from an EagerTensor.\n  return tf.train.Feature(bytes_list=tf.train.BytesList(value=[value]))\n\ndef _float_feature(value):\n  \"\"\"Returns a float_list from a float / double.\"\"\"\n  return tf.train.Feature(float_list=tf.train.FloatList(value=[value]))\n\ndef serialize_example(wav, recording_id, target, song_id, tmin,fmin, tmax, fmax):\n  feature = {\n      'wav': _bytes_feature(wav),\n      'recording_id': _bytes_feature(recording_id),\n      'target': _float_feature(target),\n      'song_id': _float_feature(song_id),\n      'tmin': _float_feature(tmin),\n      'fmin' : _float_feature(fmin),\n      'tmax': _float_feature(tmax),\n      'fmax' : _float_feature(fmax),\n  }\n  example_proto = tf.train.Example(features=tf.train.Features(feature=feature))\n  return example_proto.SerializeToString() ","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"tfrec_num = 0\nkfold = GroupKFold(n_splits=10)\nfor fold, (train_idx, test_idx) in enumerate(kfold.split(df['recording_id'], df['species_id'], df['recording_id'])):\n    x_train , y_train = df['recording_id'][test_idx] , df['species_id'][test_idx]\n   \n    with tf.io.TFRecordWriter('tp%.2i-%.2i.tfrec'%(tfrec_num, len(test_idx))) as writer:\n        print('Writing_tfrecords ',fold)\n        for recording_id , true_value in zip(x_train, y_train): \n            wav, _ = librosa.load(f'../input/rfcx-species-audio-detection/train/{recording_id}.flac', sr = None)\n            label_info = df.loc[df['recording_id'] == str(recording_id)].values[0]\n            wav = tf.audio.encode_wav(tf.reshape(wav,(wav.shape[0], 1)) ,sample_rate = 48000)\n            recording_id = label_info[0].encode()\n            target = label_info[1]\n            song_id = label_info[2]\n            tmin = label_info[3]\n            fmin = label_info[4]\n            tmax = label_info[5]\n            fmax = label_info[6]\n            example = serialize_example(wav, recording_id, target, song_id, tmin,fmin, tmax, fmax)\n            writer.write(example)\n    tfrec_num += 1","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"kfold.get_n_splits(df['recording_id'], df['species_id'], df['recording_id'])","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"","execution_count":null,"outputs":[]}],"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat":4,"nbformat_minor":4}