From 0a331f04109c903174889cc436473749a26b164c Mon Sep 17 00:00:00 2001
From: Tilman Kamp <5991088+tilmankamp@users.noreply.github.com>
Date: Fri, 24 Apr 2020 17:26:32 +0200
Subject: [PATCH] Utility for comparing samples

---
 bin/compare_samples.py | 46 ++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 46 insertions(+)
 mode change 100644 => 100755 bin/compare_samples.py

diff --git a/bin/compare_samples.py b/bin/compare_samples.py
old mode 100644
new mode 100755
index e69de29bb2..2db6d14ba3
--- a/bin/compare_samples.py
+++ b/bin/compare_samples.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python
+"""
+Tool for comparing two wav samples
+"""
+import sys
+import argparse
+
+from deepspeech_training.util.audio import AUDIO_TYPE_NP, compute_dbfs
+from deepspeech_training.util.sample_collections import load_sample
+
+
+def fail(message):
+    print(message, file=sys.stderr, flush=True)
+    sys.exit(1)
+
+
+def compare_samples():
+    sample1 = load_sample(CLI_ARGS.sample1)
+    sample2 = load_sample(CLI_ARGS.sample2)
+    if sample1.audio_format != sample2.audio_format:
+        fail('Samples differ on: audio-format')
+    if sample1.duration != sample2.duration:
+        fail('Samples differ on: duration')
+    sample1.change_audio_type(AUDIO_TYPE_NP)
+    sample2.change_audio_type(AUDIO_TYPE_NP)
+    audio_diff = sample1.audio - sample2.audio
+    dbfs, _ = compute_dbfs(audio_diff)
+    if dbfs > CLI_ARGS.threshold:
+        fail('Samples differ on: sample data ({:0.2f} dB difference) '.format(dbfs))
+    print('Samples are considered equal ({:0.2f} dB difference)'.format(dbfs), file=sys.stderr, flush=True)
+
+
+def handle_args():
+    parser = argparse.ArgumentParser(
+        description="Tool for checking similarity of two samples"
+    )
+    parser.add_argument("sample1", help="Filename of sample 1 to compare")
+    parser.add_argument("sample2", help="Filename of sample 2 to compare")
+    parser.add_argument("--threshold", type=float, default=-50.0,
+                        help="Maximum dB difference of samples before they are considered different")
+    return parser.parse_args()
+
+
+if __name__ == "__main__":
+    CLI_ARGS = handle_args()
+    compare_samples()