refactor + allow single directory in review tab

birdnet_analyzer/analyze.py

@@ -1,5 +1,4 @@
-"""Module to analyze audio samples.
-"""
+"""Module to analyze audio samples."""
 
 import argparse
 import datetime
@@ -25,7 +24,7 @@
 )
 CSV_HEADER = "Start (s),End (s),Scientific name,Common name,Confidence,File\n"
 SCRIPT_DIR = os.path.abspath(os.path.dirname(__file__))
-ASCII_LOGO = r'''                        
+ASCII_LOGO = r"""                        
                           .                                     
                        .-=-                                     
                     .:=++++.                                    
@@ -51,7 +50,7 @@
                                                  **=====        
                                                ***+==           
                                               ****+             
-'''
+"""
 
 
 def loadCodes():
@@ -101,7 +100,7 @@ def generate_raven_table(timestamps: list[str], result: dict[str, list], afile_p
         out_string += (
             f"{selection_id}\tSpectrogram 1\t1\t0\t3\t{low_freq}\t{high_freq}\tnocall\tnocall\t1.0\t{afile_path}\t0\n"
         )
-        
+
     utils.save_result_file(result_path, out_string)
 
 
@@ -255,12 +254,11 @@ def combine_raven_tables(saved_results: list[str]):
             if not rfile:
                 continue
             with open(rfile, "r", encoding="utf-8") as rf:
-
                 try:
                     lines = rf.readlines()
 
                     # make sure it's a selection table
-                    if not "Selection" in lines[0] or not "File Offset" in lines[0]:
+                    if "Selection" not in lines[0] or "File Offset" not in lines[0]:
                         continue
 
                     # skip header and add to file
@@ -270,7 +268,6 @@ def combine_raven_tables(saved_results: list[str]):
                     audiofiles.append(f_name)
 
                     for line in lines[1:]:
-
                         # empty line?
                         if not line.strip():
                             continue
@@ -312,12 +309,11 @@ def combine_rtable_files(saved_results: list[str]):
 
         for rfile in saved_results:
             with open(rfile, "r", encoding="utf-8") as rf:
-
                 try:
                     lines = rf.readlines()
 
                     # make sure it's a selection table
-                    if not "filepath" in lines[0] or not "model" in lines[0]:
+                    if "filepath" not in lines[0] or "model" not in lines[0]:
                         continue
 
                     # skip header and add to file
@@ -336,12 +332,11 @@ def combine_kaleidoscope_files(saved_results: list[str]):
 
         for rfile in saved_results:
             with open(rfile, "r", encoding="utf-8") as rf:
-
                 try:
                     lines = rf.readlines()
 
                     # make sure it's a selection table
-                    if not "INDIR" in lines[0] or not "sensitivity" in lines[0]:
+                    if "INDIR" not in lines[0] or "sensitivity" not in lines[0]:
                         continue
 
                     # skip header and add to file
@@ -360,12 +355,11 @@ def combine_csv_files(saved_results: list[str]):
 
         for rfile in saved_results:
             with open(rfile, "r", encoding="utf-8") as rf:
-
                 try:
                     lines = rf.readlines()
 
                     # make sure it's a selection table
-                    if not "Start (s)" in lines[0] or not "Confidence" in lines[0]:
+                    if "Start (s)" not in lines[0] or "Confidence" not in lines[0]:
                         continue
 
                     # skip header and add to file
@@ -378,7 +372,6 @@ def combine_csv_files(saved_results: list[str]):
 
 
 def combineResults(saved_results: list[dict[str, str]]):
-
     if "table" in cfg.RESULT_TYPES:
         combine_raven_tables([f["table"] for f in saved_results if f])
 
@@ -445,7 +438,6 @@ def predict(samples):
 
 
 def get_result_file_names(fpath: str):
-
     result_names = {}
 
     rpath = fpath.replace(cfg.INPUT_PATH, "")
@@ -588,7 +580,11 @@ def analyzeFile(item):
     freeze_support()
 
     # Parse arguments
-    parser = argparse.ArgumentParser(description=ASCII_LOGO, formatter_class=argparse.RawDescriptionHelpFormatter, usage="python -m birdnet_analyzer.analyze [options]")
+    parser = argparse.ArgumentParser(
+        description=ASCII_LOGO,
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+        usage="python -m birdnet_analyzer.analyze [options]",
+    )
     parser.add_argument("--i", default=os.path.join(SCRIPT_DIR, "example/"), help="Path to input file or folder.")
     parser.add_argument("--o", default=os.path.join(SCRIPT_DIR, "example/"), help="Path to output folder.")
     parser.add_argument("--lat", type=float, default=-1, help="Recording location latitude. Set -1 to ignore.")
@@ -720,7 +716,7 @@ def __call__(self, parser, args, values, option_string=None):
         cfg.TRANSLATED_LABELS_PATH, os.path.basename(cfg.LABELS_FILE).replace(".txt", "_{}.txt".format(args.locale))
     )
 
-    if not args.locale in ["en"] and os.path.isfile(lfile):
+    if args.locale not in ["en"] and os.path.isfile(lfile):
         cfg.TRANSLATED_LABELS = utils.readLines(lfile)
     else:
         cfg.TRANSLATED_LABELS = cfg.LABELS

birdnet_analyzer/audio.py

@@ -1,6 +1,10 @@
-"""Module containing audio helper functions.
-"""
+"""Module containing audio helper functions."""
+
+import librosa
 import numpy as np
+import soundfile as sf
+from scipy.signal import firwin, kaiserord, lfilter
+
 
 import birdnet_analyzer.config as cfg
 
@@ -22,26 +26,23 @@ def openAudioFile(path: str, sample_rate=48000, offset=0.0, duration=None, fmin=
         Returns the audio time series and the sampling rate.
     """
     # Open file with librosa (uses ffmpeg or libav)
-    import librosa
-
     sig, rate = librosa.load(path, sr=sample_rate, offset=offset, duration=duration, mono=True, res_type="kaiser_fast")
 
     # Bandpass filter
-    if fmin != None and fmax != None:
+    if fmin is not None and fmax is not None:
         sig = bandpass(sig, rate, fmin, fmax)
-        #sig = bandpassKaiserFIR(sig, rate, fmin, fmax)
+        # sig = bandpassKaiserFIR(sig, rate, fmin, fmax)
 
     return sig, rate
 
-def getAudioFileLength(path, sample_rate=48000):    
-    
+
+def getAudioFileLength(path, sample_rate=48000):
     # Open file with librosa (uses ffmpeg or libav)
-    import librosa
 
     return librosa.get_duration(filename=path, sr=sample_rate)
 
+
 def get_sample_rate(path: str):
-    import librosa
     return librosa.get_samplerate(path)
 
 
@@ -52,7 +53,6 @@ def saveSignal(sig, fname: str):
         sig: The signal to be saved.
         fname: The file path.
     """
-    import soundfile as sf
 
     sf.write(fname, sig, 48000, "PCM_16")
 
@@ -90,11 +90,11 @@ def pad(sig, seconds, srate, amount=None):
                 noise = np.zeros(noise_shape, dtype=sig.dtype)
 
         return np.concatenate((sig, noise))
-    
+
     return sig
 
 
-def splitSignal(sig, rate, seconds, overlap, minlen):
+def splitSignal(sig, rate, seconds, overlap, minlen, amount=None):
     """Split signal with overlap.
 
     Args:
@@ -103,7 +103,7 @@ def splitSignal(sig, rate, seconds, overlap, minlen):
         seconds: The duration of a segment.
         overlap: The overlapping seconds of segments.
         minlen: Minimum length of a split.
-    
+
     Returns:
         A list of splits.
     """
@@ -117,7 +117,7 @@ def splitSignal(sig, rate, seconds, overlap, minlen):
         overlap = cfg.SIG_OVERLAP
     if minlen is None or minlen <= 0 or minlen > seconds:
         minlen = cfg.SIG_MINLEN
-    
+
     # Make sure overlap is smaller then signal duration
     if overlap >= seconds:
         overlap = seconds - 0.01
@@ -153,7 +153,7 @@ def splitSignal(sig, rate, seconds, overlap, minlen):
     # Split signal with overlap
     sig_splits = []
     for i in range(0, 1 + lastchunkpos, stepsize):
-        sig_splits.append(data[i:i + chunksize])
+        sig_splits.append(data[i : i + chunksize])
 
     return sig_splits
 
@@ -177,76 +177,71 @@ def cropCenter(sig, rate, seconds):
 
     return sig
 
-def bandpass(sig, rate, fmin, fmax, order=5):
 
+def bandpass(sig, rate, fmin, fmax, order=5):
     # Check if we have to bandpass at all
     if fmin == cfg.SIG_FMIN and fmax == cfg.SIG_FMAX or fmin > fmax:
         return sig
 
     from scipy.signal import butter, lfilter
+
     nyquist = 0.5 * rate
 
     # Highpass?
-    if fmin > cfg.SIG_FMIN and fmax == cfg.SIG_FMAX:  
-
+    if fmin > cfg.SIG_FMIN and fmax == cfg.SIG_FMAX:
         low = fmin / nyquist
         b, a = butter(order, low, btype="high")
         sig = lfilter(b, a, sig)
 
     # Lowpass?
     elif fmin == cfg.SIG_FMIN and fmax < cfg.SIG_FMAX:
-
         high = fmax / nyquist
         b, a = butter(order, high, btype="low")
         sig = lfilter(b, a, sig)
 
     # Bandpass?
     elif fmin > cfg.SIG_FMIN and fmax < cfg.SIG_FMAX:
-
         low = fmin / nyquist
         high = fmax / nyquist
         b, a = butter(order, [low, high], btype="band")
         sig = lfilter(b, a, sig)
 
     return sig.astype("float32")
 
+
 # Raven is using Kaiser window FIR filter, so we try to emulate it.
-# Raven uses the Window method for FIR filter design. 
+# Raven uses the Window method for FIR filter design.
 # A Kaiser window is used with a default transition bandwidth of 0.02 times
-# the Nyquist frequency and a default stop band attenuation of 100 dB. 
-# For a complete description of this method, see Discrete-Time Signal Processing 
+# the Nyquist frequency and a default stop band attenuation of 100 dB.
+# For a complete description of this method, see Discrete-Time Signal Processing
 # (Second Edition), by Alan Oppenheim, Ronald Schafer, and John Buck, Prentice Hall 1998, pp. 474-476.
 def bandpassKaiserFIR(sig, rate, fmin, fmax, width=0.02, stopband_attenuation_db=100):
-
     # Check if we have to bandpass at all
     if fmin == cfg.SIG_FMIN and fmax == cfg.SIG_FMAX or fmin > fmax:
         return sig
 
-    from scipy.signal import kaiserord, firwin, lfilter
     nyquist = 0.5 * rate
 
     # Calculate the order and Kaiser parameter for the desired specifications.
     N, beta = kaiserord(stopband_attenuation_db, width)
 
     # Highpass?
-    if fmin > cfg.SIG_FMIN and fmax == cfg.SIG_FMAX: 
+    if fmin > cfg.SIG_FMIN and fmax == cfg.SIG_FMAX:
         low = fmin / nyquist
-        taps = firwin(N, low, window=('kaiser', beta), pass_zero=False)
+        taps = firwin(N, low, window=("kaiser", beta), pass_zero=False)
 
     # Lowpass?
     elif fmin == cfg.SIG_FMIN and fmax < cfg.SIG_FMAX:
         high = fmax / nyquist
-        taps = firwin(N, high, window=('kaiser', beta), pass_zero=True)
+        taps = firwin(N, high, window=("kaiser", beta), pass_zero=True)
 
     # Bandpass?
     elif fmin > cfg.SIG_FMIN and fmax < cfg.SIG_FMAX:
         low = fmin / nyquist
         high = fmax / nyquist
-        taps = firwin(N, [low, high], window=('kaiser', beta), pass_zero=False)
+        taps = firwin(N, [low, high], window=("kaiser", beta), pass_zero=False)
 
     # Apply the filter to the signal.
     sig = lfilter(taps, 1.0, sig)
 
     return sig.astype("float32")
-
-

birdnet_analyzer/client.py

@@ -1,5 +1,5 @@
-"""Client to send requests to the server.
-"""
+"""Client to send requests to the server."""
+
 import argparse
 import json
 import os
@@ -76,15 +76,20 @@ def saveResult(data, fpath):
         help="Week of the year when the recording was made. Values in [1, 48] (4 weeks per month). Set -1 for year-round species list.",
     )
     parser.add_argument(
-        "--overlap", type=float, default=0.0, help="Overlap of prediction segments. Values in [0.0, 2.9]. Defaults to 0.0."
+        "--overlap",
+        type=float,
+        default=0.0,
+        help="Overlap of prediction segments. Values in [0.0, 2.9]. Defaults to 0.0.",
     )
     parser.add_argument(
         "--sensitivity",
         type=float,
         default=1.0,
         help="Detection sensitivity; Higher values result in higher sensitivity. Values in [0.5, 1.5]. Defaults to 1.0.",
     )
-    parser.add_argument("--pmode", default="avg", help="Score pooling mode. Values in ['avg', 'max']. Defaults to 'avg'.")
+    parser.add_argument(
+        "--pmode", default="avg", help="Score pooling mode. Values in ['avg', 'max']. Defaults to 'avg'."
+    )
     parser.add_argument("--num_results", type=int, default=5, help="Number of results per request. Defaults to 5.")
     parser.add_argument(
         "--sf_thresh",

birdnet_analyzer/config.py

@@ -269,7 +269,6 @@ def getConfig():
 
 
 def setConfig(c):
-
     global RANDOM_SEED
     global MODEL_VERSION
     global PB_MODEL

birdnet_analyzer/embeddings.py

@@ -1,5 +1,4 @@
-"""Module used to extract embeddings for samples.
-"""
+"""Module used to extract embeddings for samples."""
 
 import argparse
 import datetime
@@ -107,7 +106,7 @@ def analyzeFile(item):
     # Save as embeddings file
     try:
         # We have to check if output path is a file or directory
-        if not cfg.OUTPUT_PATH.rsplit(".", 1)[-1].lower() in ["txt", "csv"]:
+        if cfg.OUTPUT_PATH.rsplit(".", 1)[-1].lower() not in ["txt", "csv"]:
             fpath = fpath.replace(cfg.INPUT_PATH, "")
             fpath = fpath[1:] if fpath[0] in ["/", "\\"] else fpath
 
@@ -136,10 +135,14 @@ def analyzeFile(item):
     # Parse arguments
     parser = argparse.ArgumentParser(description="Extract feature embeddings with BirdNET")
     parser.add_argument(
-        "--i", default=os.path.join(SCRIPT_DIR, "example/"), help="Path to input file or folder. If this is a file, --o needs to be a file too."
+        "--i",
+        default=os.path.join(SCRIPT_DIR, "example/"),
+        help="Path to input file or folder. If this is a file, --o needs to be a file too.",
     )
     parser.add_argument(
-        "--o", default=os.path.join(SCRIPT_DIR, "example/"), help="Path to output file or folder. If this is a file, --i needs to be a file too."
+        "--o",
+        default=os.path.join(SCRIPT_DIR, "example/"),
+        help="Path to output file or folder. If this is a file, --i needs to be a file too.",
     )
     parser.add_argument(
         "--overlap",