Egyptian fruit bats

Egyptian fruit bats#

egyptian fruit bat

© University of Georgia

Recordings from:

Prat, Y. et al. (2017). An annotated dataset of Egyptian fruit bat vocalizations across varying contexts and during ontogeny. Scientific Data. https://www.doi.org/10.1038/sdata.2017.143

Setup#

[1]:

# import packages
import chatter
from pathlib import Path

[2]:

# define config parameters
config = {
    # spectrogram parameters
    "sr": 250000,
    "n_mels": 64,
    "fmin": 1000,
    "fmax": 40000,
    # preprocessing parameters
    "high_pass": 1000,
    "low_pass": 40000,
    "threshold": 1.5,
    # pykanto segmentation parameters
    "pykanto_noise_floor": -65,
    "pykanto_top_dB": 65,
    "pykanto_max_dB": -30,
    "pykanto_dB_delta": 5,
    "pykanto_silence_threshold": 0.01,
    "pykanto_min_silence_length": 0.001,
    "pykanto_max_unit_length": 0.2,
    "pykanto_min_unit_length": 0.02,
    "pykanto_gauss_sigma": 10,
    # other parameters
    "plot_clip_duration": 5,
}
config = chatter.make_config(config)

# initialize the analyzer with the configuration
# using only 8 cores to avoid memory crash during segmentation
analyzer = chatter.Analyzer(config, n_jobs=8)
model = chatter.Trainer(config)

# set paths
input_dir = Path(
    "/Volumes/Expansion/data/chatter/examples/egyptian_fruit_bat/recordings/raw"
)
processed_dir = Path(
    "/Volumes/Expansion/data/chatter/examples/egyptian_fruit_bat/recordings/processed"
)
h5_path = Path(
    "/Volumes/Expansion/data/chatter/examples/egyptian_fruit_bat/spectrograms.h5"
)
csv_path = Path(
    "/Volumes/Expansion/data/chatter/examples/egyptian_fruit_bat/spectrograms.csv"
)
model_dir = Path("/Volumes/Expansion/data/chatter/examples/egyptian_fruit_bat/model")
output_csv_path = Path(
    "/Volumes/Expansion/data/chatter/examples/egyptian_fruit_bat/output.csv"
)

Using 8 cores for parallel processing
Initializing convolutional variational autoencoder
Using device: mps

Preprocessing#

[22]:

# demo the preprocessing pipeline
analyzer.demo_preprocessing(input_dir)

--- Demoing preprocessing for: 130527062206057583.WAV ---
   Segment: 0.00s - 5.00s
../_images/contents_egyptian_fruit_bat_7_1.png 
[24]:

# preprocess recordings
analyzer.preprocess_directory(input_dir=input_dir, processed_dir=processed_dir)

--- Found 10000 audio files to preprocess ---

Preprocessing audio: 100%|██████████| 10000/10000 [07:28<00:00, 22.30it/s]


--- Preprocessing complete. Standardized WAV audio saved to /Volumes/Expansion/data/chatter/examples/egyptian_fruit_bat/recordings/processed ---

Segmentation#

[13]:

# preview the segmentation pipeline
analyzer.demo_segmentation(input_dir=processed_dir, simple=True)
analyzer.demo_segmentation(input_dir=processed_dir, simple=True)
analyzer.demo_segmentation(input_dir=processed_dir, simple=True)
analyzer.demo_segmentation(input_dir=processed_dir, simple=True)

--- Demoing segmentation for: 130225053249075024.wav ---
   Segment: 0.00s - 5.00s
../_images/contents_egyptian_fruit_bat_10_1.png 
--- Demoing segmentation for: 130219215043723673.wav ---
   Segment: 0.00s - 5.00s
../_images/contents_egyptian_fruit_bat_10_3.png 
--- Demoing segmentation for: 130302044537445170.wav ---
   Segment: 0.00s - 5.00s
../_images/contents_egyptian_fruit_bat_10_5.png 
--- Demoing segmentation for: 130228174949622282.wav ---
   Segment: 0.00s - 5.00s
../_images/contents_egyptian_fruit_bat_10_7.png 
[3]:

# segment units and save spectrograms
unit_df = analyzer.segment_and_create_spectrograms(
    processed_dir=processed_dir, h5_path=h5_path, csv_path=csv_path, simple=False
)


--- Found 10000 files to segment using pykanto (image-based) method ---

Segmenting and saving spectrograms: 100%|██████████| 10000/10000 [02:16<00:00, 73.19it/s]


Warning: 858 files (8.6%) had no units matching the segmentation criteria.

--- Data preparation complete. Created records for 46949 units ---
Spectrograms saved to /Volumes/Expansion/data/chatter/examples/egyptian_fruit_bat/spectrograms.h5
Unit metadata saved to /Volumes/Expansion/data/chatter/examples/egyptian_fruit_bat/spectrograms.csv

Run model#

[14]:

# load segmented units
unit_df = analyzer.load_df(csv_path)

Attempting to load /Volumes/Expansion/data/chatter/examples/egyptian_fruit_bat/spectrograms.csv...
--- Successfully loaded /Volumes/Expansion/data/chatter/examples/egyptian_fruit_bat/spectrograms.csv ---

[15]:

# train ae
model.train_ae(unit_df=unit_df, h5_path=h5_path, model_dir=model_dir, subset=0.05)

--- Training on a random subset of 2347 units (5.0%) ---

Starting training for 100 epochs using 4 DataLoader workers...

Training model: 100%|██████████| 100/100 [09:15<00:00,  5.55s/it, loss=1591.1240]

--- Training complete. Model saved to /Volumes/Expansion/data/chatter/examples/egyptian_fruit_bat/model/model.pth ---
Loss history saved to /Volumes/Expansion/data/chatter/examples/egyptian_fruit_bat/model/loss.csv

[17]:

# load trained vae
model = chatter.Trainer.from_trained(config, model_dir)

Instantiating Trainer from pre-trained model at /Volumes/Expansion/data/chatter/examples/egyptian_fruit_bat/model...
Initializing convolutional variational autoencoder
Using device: mps
Successfully loaded pre-trained model from /Volumes/Expansion/data/chatter/examples/egyptian_fruit_bat/model/model.pth

[18]:

# assess the quality of reconstruction
model.plot_reconstructions(unit_df=unit_df, h5_path=h5_path)
../_images/contents_egyptian_fruit_bat_16_0.png 
[20]:

# export the latent features
output = model.extract_and_save_features(
    unit_df=unit_df,
    h5_path=h5_path,
    model_dir=model_dir,
    output_csv_path=output_csv_path,
)


--- Starting feature extraction ---
Successfully loaded pre-trained model from /Volumes/Expansion/data/chatter/examples/egyptian_fruit_bat/model/model.pth

Extracting features: 100%|██████████| 734/734 [00:11<00:00, 63.91it/s]


--- Pipeline complete. Exported data for 46949 units to /Volumes/Expansion/data/chatter/examples/egyptian_fruit_bat/output.csv ---

Postprocessing#

[21]:

# load in latent features
output = chatter.FeatureProcessor(analyzer.load_df(output_csv_path), config)

Attempting to load /Volumes/Expansion/data/chatter/examples/egyptian_fruit_bat/output.csv...
--- Successfully loaded /Volumes/Expansion/data/chatter/examples/egyptian_fruit_bat/output.csv ---

[22]:

# process output and save
output.run_pacmap()
output.df.to_csv(output_csv_path, index=False)

--- Running PaCMAP dimensionality reduction ---
--- PaCMAP complete ---

[28]:

# create interactive or static plot
output.static_embedding_plot(
    h5_path=h5_path,
    seed=111,
    focal_quantile=0.6,
    point_size=0.2,
    point_alpha=1,
    margin=0.01,
    zoom_padding=0.5,
    num_neighbors=5,
)

--- Automatically selecting focal points from quadrants with seed 111 ---
--- Finding nearest neighbors ---
--- Creating the plot ---
--- Plotting density background (using fast 2d histogram) ---
--- Calculating callout positions and adding spectrograms ---
--- Displaying plot ---
../_images/contents_egyptian_fruit_bat_21_1.png 
[28]:

<chatter.features.FeatureProcessor at 0x370753770>
Contents