Remove `/` prefix from the relative path (#2065)

TTS/tts/datasets/__init__.py

@@ -1,3 +1,4 @@
+import os
 import sys
 from collections import Counter
 from pathlib import Path
@@ -12,20 +13,16 @@ from TTS.tts.datasets.formatters import *
 def split_dataset(items, eval_split_max_size=None, eval_split_size=0.01):
     """Split a dataset into train and eval. Consider speaker distribution in multi-speaker training.
 
-        Args:
+    Args:
-    <<<<<<< HEAD
+        items (List[List]):
-            items (List[List]):
+            A list of samples. Each sample is a list of `[audio_path, text, speaker_id]`.
-                A list of samples. Each sample is a list of `[audio_path, text, speaker_id]`.
 
-            eval_split_max_size (int):
+        eval_split_max_size (int):
-                Number maximum of samples to be used for evaluation in proportion split. Defaults to None (Disabled).
+            Number maximum of samples to be used for evaluation in proportion split. Defaults to None (Disabled).
 
-            eval_split_size (float):
+        eval_split_size (float):
-                If between 0.0 and 1.0 represents the proportion of the dataset to include in the evaluation set.
+            If between 0.0 and 1.0 represents the proportion of the dataset to include in the evaluation set.
-                If > 1, represents the absolute number of evaluation samples. Defaults to 0.01 (1%).
+            If > 1, represents the absolute number of evaluation samples. Defaults to 0.01 (1%).
-    =======
-            items (List[List]): A list of samples. Each sample is a list of `[text, audio_path, speaker_id]`.
-    >>>>>>> Fix docstring
     """
     speakers = [item["speaker_name"] for item in items]
     is_multi_speaker = len(set(speakers)) > 1
@@ -64,10 +61,9 @@ def add_extra_keys(metadata, language, dataset_name):
         # add language name
         item["language"] = language
         # add unique audio name
-        relfilepath = os.path.splitext(item["audio_file"].replace(item["root_path"], ""))[0]
+        relfilepath = os.path.splitext(os.path.relpath(item["audio_file"], item["root_path"]))[0]
         audio_unique_name = f"{dataset_name}#{relfilepath}"
         item["audio_unique_name"] = audio_unique_name
-
     return metadata
 
 

tests/__init__.py

@@ -33,7 +33,9 @@ def get_tests_data_path():
 
 def get_tests_output_path():
     """Returns the path to the directory for test outputs."""
-    return os.path.join(get_tests_path(), "outputs")
+    path = os.path.join(get_tests_path(), "outputs")
+    os.makedirs(path, exist_ok=True)
+    return path
 
 
 def run_cli(command):

tests/data/ljspeech/speakers.json

@@ -1,5 +1,5 @@
 {
-    "#/wavs/LJ001-0001": {
+    "#wavs/LJ001-0001": {
         "name": "ljspeech-0",
         "embedding": [
             0.05539746582508087,
@@ -260,7 +260,7 @@
             -0.09469571709632874
         ]
     },
-    "#/wavs/LJ001-0002": {
+    "#wavs/LJ001-0002": {
         "name": "ljspeech-1",
         "embedding": [
             0.05539746582508087,
@@ -521,7 +521,7 @@
             -0.09469571709632874
         ]
     },
-    "#/wavs/LJ001-0003": {
+    "#wavs/LJ001-0003": {
         "name": "ljspeech-2",
         "embedding": [
             0.05539746582508087,
@@ -782,7 +782,7 @@
             -0.09469571709632874
         ]
     },
-    "#/wavs/LJ001-0004": {
+    "#wavs/LJ001-0004": {
         "name": "ljspeech-3",
         "embedding": [
             0.05539746582508087,
@@ -1043,7 +1043,7 @@
             -0.09469571709632874
         ]
     },
-    "#/wavs/LJ001-0005": {
+    "#wavs/LJ001-0005": {
         "name": "ljspeech-4",
         "embedding": [
             0.05539746582508087,
@@ -1304,7 +1304,7 @@
             -0.09469571709632874
         ]
     },
-    "#/wavs/LJ001-0006": {
+    "#wavs/LJ001-0006": {
         "name": "ljspeech-5",
         "embedding": [
             0.05539746582508087,
@@ -1565,7 +1565,7 @@
             -0.09469571709632874
         ]
     },
-    "#/wavs/LJ001-0007": {
+    "#wavs/LJ001-0007": {
         "name": "ljspeech-6",
         "embedding": [
             0.05539746582508087,
@@ -1826,7 +1826,7 @@
             -0.09469571709632874
         ]
     },
-    "#/wavs/LJ001-0008": {
+    "#wavs/LJ001-0008": {
         "name": "ljspeech-7",
         "embedding": [
             0.05539746582508087,
@@ -2087,7 +2087,7 @@
             -0.09469571709632874
         ]
     },
-    "#/wavs/LJ001-0009": {
+    "#wavs/LJ001-0009": {
         "name": "ljspeech-8",
         "embedding": [
             0.05539746582508087,
@@ -2348,7 +2348,7 @@
             -0.09469571709632874
         ]
     },
-    "#/wavs/LJ001-0010": {
+    "#wavs/LJ001-0010": {
         "name": "ljspeech-9",
         "embedding": [
             0.05539746582508087,

tests/inference_tests/test_synthesizer.py

@@ -1,7 +1,7 @@
 import os
 import unittest
 
-from tests import get_tests_output_path
+from tests import get_tests_input_path
 from TTS.config import load_config
 from TTS.tts.models import setup_model
 from TTS.utils.io import save_checkpoint
@@ -12,14 +12,14 @@ class SynthesizerTest(unittest.TestCase):
     # pylint: disable=R0201
     def _create_random_model(self):
         # pylint: disable=global-statement
-        config = load_config(os.path.join(get_tests_output_path(), "dummy_model_config.json"))
+        config = load_config(os.path.join(get_tests_input_path(), "dummy_model_config.json"))
         model = setup_model(config)
-        output_path = os.path.join(get_tests_output_path())
+        output_path = os.path.join(get_tests_input_path())
         save_checkpoint(config, model, None, None, 10, 1, output_path)
 
     def test_in_out(self):
         self._create_random_model()
-        tts_root_path = get_tests_output_path()
+        tts_root_path = get_tests_input_path()
         tts_checkpoint = os.path.join(tts_root_path, "checkpoint_10.pth")
         tts_config = os.path.join(tts_root_path, "dummy_model_config.json")
         synthesizer = Synthesizer(tts_checkpoint, tts_config, None, None)