Embeddings with hoplite

birdnet_analyzer/analyze/utils.py

@@ -290,7 +290,7 @@ def combine_raven_tables(saved_results: list[str]):
 
                     # skip header and add to file
                     f_name = lines[1].split("\t")[10]
-                    f_duration = audio.get_audio_file_Length(f_name)
+                    f_duration = audio.get_audio_file_length(f_name)
 
                     audiofiles.append(f_name)
 
@@ -614,7 +614,7 @@ def analyze_file(item):
     print(f"Analyzing {fpath}", flush=True)
 
     try:
-        fileLengthSeconds = int(audio.get_audio_file_Length(fpath) / cfg.AUDIO_SPEED)
+        fileLengthSeconds = int(audio.get_audio_file_length(fpath) / cfg.AUDIO_SPEED)
     except Exception as ex:
         # Write error log
         print(f"Error: Cannot analyze audio file {fpath}. File corrupt?\n", flush=True)

birdnet_analyzer/audio.py

@@ -49,7 +49,7 @@ def open_audio_file(path: str, sample_rate=48000, offset=0.0, duration=None, fmi
     return sig, rate
 
 
-def get_audio_file_Length(path):
+def get_audio_file_length(path):
     """
     Get the length of an audio file in seconds.
 

birdnet_analyzer/cli.py

@@ -268,6 +268,22 @@ def bs_args():
 
     return p
 
+def db_args():
+    """
+    Creates an arguments parser for the database path.
+    Returns:
+        argparse.ArgumentParser: An argument parser with a database size argument.
+    The parser includes the following argument:
+        -db, --database: Path to the database folder.
+    """
+    p = argparse.ArgumentParser(add_help=False)
+    p.add_argument(
+        "-db",
+        "--database",
+        help="Path to the database folder.",
+    )
+
+    return p
 
 def analyzer_parser():
     """
@@ -360,21 +376,90 @@ def embeddings_parser():
     Creates and returns an argument parser for extracting feature embeddings with BirdNET.
 
     The parser includes arguments from the following parent parsers:
-    - io_args(): Handles input/output arguments.
+    - db_args(): Handles database arguments.
     - bandpass_args(): Handles bandpass filter arguments.
+    - audio_speed_args(): Handles audio speed arguments.
     - overlap_args(): Handles overlap arguments.
     - threads_args(): Handles threading arguments.
     - bs_args(): Handles batch size arguments.
 
     Returns:
         argparse.ArgumentParser: Configured argument parser for extracting feature embeddings.
     """
+
+    parents = [
+        db_args(),
+        bandpass_args(),
+        audio_speed_args(), 
+        overlap_args(), 
+        threads_args(), 
+        bs_args()
+    ]
+
     parser = argparse.ArgumentParser(
         formatter_class=argparse.ArgumentDefaultsHelpFormatter,
-        parents=[io_args(), bandpass_args(), overlap_args(), threads_args(), bs_args()],
+        parents=parents,
     )
+
+    parser.add_argument(
+        "-i",
+        "--input",
+        help="Path to input file or folder.",
+    )
+
     return parser
 
+def search_parser():
+    """
+    Creates and returns an argument parser for searching BirdNET embeddings.
+
+    The parser includes the following arguments:
+    - -q, --queryfile: Path to the query file.
+    - -o, --output: Path to the output folder.
+    - --n_results: Number of results to return.
+    - --score_function: Scoring function to use. Choose 'cosine', 'euclidean' or 'dot'. Defaults to 'cosine'.
+    - --crop_mode: Crop mode for the query sample. Can be 'center', 'first' or 'segments'.
+
+    The parser also includes arguments from the following parent parsers:
+    - overlap_args(): Handles overlap arguments if segments is selected as crop mode.
+    - db_args(): Handles database arguments.
+    """
+
+    parents = [overlap_args(), db_args()]
+
+    parser = argparse.ArgumentParser(
+        formatter_class=argparse.ArgumentDefaultsHelpFormatter,
+        parents=parents
+    )
+    parser.add_argument(
+        "-q",
+        "--queryfile",
+        help="Path to the query file."
+    )
+    parser.add_argument(
+        "-o",
+        "--output",
+        help="Path to the output folder."
+    )
+    parser.add_argument(
+        "--n_results",
+        default=10,
+        help="Number of results to return."
+    )
+
+    # TODO: use choice argument.
+    parser.add_argument(
+        "--score_function",
+        default="cosine",
+        help="Scoring function to use. Choose 'cosine', 'euclidean' or 'dot'. Defaults to 'cosine'."
+    )
+    parser.add_argument(
+        "--crop_mode",
+        default=cfg.SAMPLE_CROP_MODE,
+        help="Crop mode for the query sample. Can be 'center', 'first' or 'segments'.",
+    )
+
+    return parser
 
 def client_parser():
     """

birdnet_analyzer/embeddings/__init__.py

@@ -1,64 +1,8 @@
 def main():
-    import os
-    from multiprocessing import Pool
-
-    import birdnet_analyzer.config as cfg
-    import birdnet_analyzer.utils as utils
     import birdnet_analyzer.cli as cli
-
     parser = cli.embeddings_parser()
-
     args = parser.parse_args()
 
-    from birdnet_analyzer.embeddings.utils import analyze_file  # noqa: E402
-
-    ### Make sure to comment out appropriately if you are not using args. ###
-
-    # Set input and output path
-    cfg.INPUT_PATH = args.input
-
-    if not args.output:
-        if os.path.isfile(cfg.INPUT_PATH):
-            cfg.OUTPUT_PATH = os.path.dirname(cfg.INPUT_PATH)
-        else:
-            cfg.OUTPUT_PATH = cfg.INPUT_PATH
-    else:
-        cfg.OUTPUT_PATH = args.output
-
-    # Parse input files
-    if os.path.isdir(cfg.INPUT_PATH):
-        cfg.FILE_LIST = utils.collect_audio_files(cfg.INPUT_PATH)
-    else:
-        cfg.FILE_LIST = [cfg.INPUT_PATH]
-
-    # Set overlap
-    cfg.SIG_OVERLAP = args.overlap
-
-    # Set bandpass frequency range
-    cfg.BANDPASS_FMIN = args.fmin
-    cfg.BANDPASS_FMAX = args.fmax
-
-    # Set number of threads
-    if os.path.isdir(cfg.INPUT_PATH):
-        cfg.CPU_THREADS = args.threads
-        cfg.TFLITE_THREADS = 1
-    else:
-        cfg.CPU_THREADS = 1
-        cfg.TFLITE_THREADS = args.threads
-
-    # Set batch size
-    cfg.BATCH_SIZE = args.batchsize
-
-    # Add config items to each file list entry.
-    # We have to do this for Windows which does not
-    # support fork() and thus each process has to
-    # have its own config. USE LINUX!
-    flist = [(f, cfg.get_config()) for f in cfg.FILE_LIST]
+    from birdnet_analyzer.embeddings.utils import run  # noqa: E402
 
-    # Analyze files
-    if cfg.CPU_THREADS < 2:
-        for entry in flist:
-            analyze_file(entry)
-    else:
-        with Pool(cfg.CPU_THREADS) as p:
-            p.map(analyze_file, flist)
+    run(args.input, args.database, args.overlap, args.audio_speed, args.fmin, args.fmax, args.threads, args.batchsize)