Merge branch 'pr/agrinh/457-2' into dev

.github/workflows/main.yml

@@ -43,7 +43,7 @@ jobs:
         run: python3 -m pip install --upgrade pip
       - name: Install TTS
         run: |
-          python3 -m pip install .
+          python3 -m pip install .[all]
           python3 setup.py egg_info
       - name: Lint check
         run: |

Makefile

@@ -1,5 +1,5 @@
 .DEFAULT_GOAL := help
-.PHONY: test deps style lint install help
+.PHONY: test system-deps dev-deps deps style lint install help
 
 help:
 	@grep -E '^[a-zA-Z_-]+:.*?## .*$$' $(MAKEFILE_LIST) | sort | awk 'BEGIN {FS = ":.*?## "}; {printf "\033[36m%-30s\033[0m %s\n", $$1, $$2}'
@@ -10,6 +10,10 @@ system-deps:	## install linux system deps
 	sudo apt-get install -y espeak-ng
 	sudo apt-get install -y libsndfile1-dev
 
+dev-deps:  ## install development deps
+	pip install -r requirements.dev.txt
+	pip install -r requirements.tf.txt
+
 deps:	## install 🐸 requirements.
 	pip install -r requirements.txt
 
@@ -25,4 +29,4 @@ lint:	## run pylint linter.
 	pylint ${target_dirs}
 
 install:	## install 🐸 TTS for development.
-	pip install -e .
+	pip install -e .[all]

README.md

@@ -111,11 +111,17 @@ If you are only interested in [synthesizing speech](https://github.com/coqui-ai/
 pip install TTS
 ```
 
+By default this only installs the requirements for PyTorch. To install the tensorflow dependencies as well, use the `tf` extra.
+
+```bash
+pip install TTS[tf]
+```
+
 If you plan to code or train models, clone 🐸TTS and install it locally.
 
 ```bash
 git clone https://github.com/coqui-ai/TTS
-pip install -e .
+pip install -e .[all,dev,notebooks,tf]  # Select the relevant extras
 ```
 
 We use ```espeak-ng``` to convert graphemes to phonemes. You might need to install separately.

TTS/speaker_encoder/utils/prepare_voxceleb.py

@@ -27,10 +27,8 @@
 
 import pandas
 import soundfile as sf
-import tensorflow as tf
 from absl import logging
 
-gfile = tf.compat.v1.gfile
 
 SUBSETS = {
     "vox1_dev_wav": [
@@ -73,8 +71,7 @@ def download_and_extract(directory, subset, urls):
         subset: subset name of the corpus.
         urls: the list of urls to download the data file.
     """
-    if not gfile.Exists(directory):
-        gfile.MakeDirs(directory)
+    os.makedirs(directory, exist_ok=True)
 
     try:
         for url in urls:
@@ -107,7 +104,7 @@ def download_and_extract(directory, subset, urls):
             extract_path_ori = os.path.join(directory, zfile.infolist()[0].filename)
             subprocess.call("mv %s %s" % (extract_path_ori, extract_path), shell=True)
     finally:
-        # gfile.Remove(zip_filepath)
+        # os.remove(zip_filepath)
         pass
 
 
@@ -160,7 +157,7 @@ def convert_audio_and_make_label(input_dir, subset, output_dir, output_file):
 
     files = []
     # Convert all AAC file into WAV format. At the same time, generate the csv
-    for root, _, filenames in gfile.Walk(source_dir):
+    for root, _, filenames in os.walk(source_dir):
         for filename in filenames:
             name, ext = os.path.splitext(filename)
             if ext.lower() == ".wav":
@@ -172,7 +169,7 @@ def convert_audio_and_make_label(input_dir, subset, output_dir, output_file):
                 # Convert AAC to WAV.
                 aac_file = os.path.join(root, filename)
                 wav_file = aac_file + ".wav"
-                if not gfile.Exists(wav_file):
+                if not os.path.exists(wav_file):
                     if not decode_aac_with_ffmpeg(aac_file, wav_file):
                         raise RuntimeError("Audio decoding failed.")
             else:

requirements.dev.txt

@@ -0,0 +1,5 @@
+black
+coverage
+isort
+nose
+pylint==2.7.4

requirements.notebooks.txt

@@ -0,0 +1,2 @@
+bokeh==1.4.0
+numba==0.48

requirements.tf.txt

@@ -0,0 +1 @@
+tensorflow==2.3.1

requirements.txt

@@ -1,29 +1,20 @@
-torch>=1.7
-tensorflow==2.3.1
-numpy==1.18.5
-scipy>=0.19.0
-numba==0.48
+cython
+flask
+gdown
+inflect
+jieba
 librosa==0.7.2
+matplotlib
+numpy==1.18.5
+pandas
 phonemizer>=2.2.0
-unidecode==0.4.20
 pypinyin
-jieba
-tensorboardX
-matplotlib
-Pillow
-flask
-tqdm
-inflect
-bokeh==1.4.0
 pysbd
+pyyaml
+scipy>=0.19.0
 soundfile
-gdown
+tensorboardX
+torch>=1.7
+tqdm
 umap-learn==0.4.6
-cython
-pyyaml
-# quality and style
-nose
-coverage
-black
-isort
-pylint==2.7.4
+unidecode==0.4.20

setup.py

@@ -48,6 +48,14 @@ def pip_install(package_name):
 
 
 requirements = open(os.path.join(cwd, 'requirements.txt'), 'r').readlines()
+with open(os.path.join(cwd, 'requirements.notebooks.txt'), 'r') as f:
+    requirements_notebooks = f.readlines()
+with open(os.path.join(cwd, 'requirements.dev.txt'), 'r') as f:
+    requirements_dev = f.readlines()
+with open(os.path.join(cwd, 'requirements.tf.txt'), 'r') as f:
+    requirements_tf = f.readlines()
+requirements_all = requirements_dev + requirements_notebooks + requirements_tf
+
 with open('README.md', "r", encoding="utf-8") as readme_file:
     README = readme_file.read()
 
@@ -82,6 +90,12 @@ def pip_install(package_name):
         # 'build_ext': build_ext
     },
     install_requires=requirements,
+    extras_require={
+        "all": requirements_all,
+        "dev": requirements_dev,
+        "notebooks": requirements_notebooks,
+        "tf": requirements_tf,
+    },
     python_requires='>=3.6.0, <3.9',
     entry_points={
         'console_scripts': [