Plot raw pitch data

Plot data as it is in the dataset, without any processing.
Raw pitch data is plotted with plotly, next steps are midi and
spectrogram.

related to: #7

penn/plot/__init__.py

@@ -1,3 +1,4 @@
 from . import density
 from . import logits
+from . import raw_data
 from . import threshold

penn/plot/raw_data/__init__.py

@@ -0,0 +1 @@
+from .core import *

penn/plot/raw_data/__main__.py

@@ -0,0 +1,29 @@
+import argparse
+from pathlib import Path
+
+import penn
+
+
+###############################################################################
+# Create figure
+###############################################################################
+
+
+def parse_args():
+    """Parse command-line arguments"""
+    parser = argparse.ArgumentParser(description='Create logits figure')
+    parser.add_argument(
+        '--data_dir',
+        type=Path,
+        required=True,
+        help='The audio file to plot the logits of')
+    parser.add_argument(
+        '--file_stem',
+        type=str,
+        required=True,
+        help='The jpg file to save the plot')
+    return parser.parse_known_args()[0]
+
+
+penn.plot.raw_data.from_data(**vars(parse_args()))
+

penn/plot/raw_data/core.py

@@ -0,0 +1,72 @@
+import penn
+import jams
+import scipy
+import numpy as np
+from os.path import isfile
+
+import plotly.express as px
+import plotly.graph_objects as go
+
+
+def extract_pitch(jams_track):
+    # Extract all of the pitch annotations
+    pitch_data_slices = jams_track.annotations[penn.data.preprocess.JAMS_PITCH_HZ]
+
+    pitch_dict = {}
+
+    for slice in pitch_data_slices:
+        string_indx = slice.annotation_metadata[penn.data.preprocess.JAMS_STRING_IDX]
+        pitch_slice_dict = pitch_dict[int(string_indx)] = {}
+
+        freq_list = []
+        time_list = []
+
+        # Loop through the pitch observations pertaining to this slice
+        for pitch in slice:
+            # Extract the pitch
+            freq = pitch.value['frequency']
+            time = pitch.time
+
+            if np.sum(freq) != 0 and pitch.value['voiced']:
+                freq_list.append(freq)
+                time_list.append(time)
+
+        pitch_slice_dict["pitch"] = np.array(freq_list)
+        pitch_slice_dict["time"] = np.array(time_list)
+
+    return pitch_dict
+
+def pitch_with_plotly(pitch_dict):
+    # fig = px.scatter(y=pitch_dict[0]["pitch"], x=pitch_dict[0]["time"])
+
+    fig = go.Figure()
+
+    for no_slice, pitch_slice in pitch_dict.items():
+        fig = fig.add_trace(go.Scatter(
+            name=f"String {no_slice}",
+            x = pitch_slice["time"],
+            y = pitch_slice["pitch"],
+            mode="markers",
+            marker=dict (
+                size=5
+                )
+            ))
+    fig.show()
+
+
+
+def from_data(data_dir, file_stem):
+    file_stem = file_stem.split('.')[0].split('mic')[0]
+
+    pitch_file = data_dir / 'annotations' / f"{file_stem}.jams"
+    audio_file = data_dir / 'audio-mono-mic' / f"{file_stem}_mic.wav"
+
+    print(audio_file)
+    print(pitch_file)
+
+    assert isfile(audio_file)
+    assert isfile(pitch_file)
+
+    jams_track = jams.load(str(pitch_file))
+    pitch_dict = extract_pitch(jams_track)
+    pitch_with_plotly(pitch_dict)