Merge pull request 'modify project structure and add download_models.py' (#3) from census_of_the_project_structure into master

Reviewed-on: #3

.gitignore

@@ -6,8 +6,8 @@ __pycache__/
 *$py.class
 
 # Custom
-model_small/
-model_large/
+data/model_small/
+data/model_large/
 
 # C extensions
 *.so

data/config.py

@@ -1,5 +1,5 @@
 VA_ALIAS = ('джарвис',)
 VA_TBR = ('скажи', 'покажи', 'ответь', 'произнеси', 'расскажи', 'сколько', 'слушай')
+MODEL_NAME = "vosk-model-small-ru-0.22"  # vosk-model-ru-0.42
 MICROPHONE_INDEX = -1
 PICOVOICE_TOKEN = "4xbwaZwZmSHeTiowFl5Rgqsc8CR4FKGV8YueJUlR4Zt2e1kB64IDcA=="
-OPENAI_TOKEN = "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"

main.py

@@ -1,4 +1,4 @@
-from Jarvis import Jarvis
+from modules.Jarvis import Jarvis
 
 
 def main():

modules/Jarvis.py

@@ -10,21 +10,22 @@ import yaml
 from fuzzywuzzy import fuzz
 from pvrecorder import PvRecorder
 
-import config
-from utils import execute_cmd, play
+from data import config
+from utils import download_models, execute_cmd, play
 
 
 class Jarvis:
     def __init__(self):
+        download_models.install_vosk_model()
         self.recorder = None
         self.CDIR = os.getcwd()
-        self.VA_CMD_LIST = yaml.safe_load(open('commands.yaml', encoding='utf8'))
+        self.VA_CMD_LIST = yaml.safe_load(open('data/commands.yaml', encoding='utf8'))
         self.porcupine = pvporcupine.create(
             access_key=config.PICOVOICE_TOKEN,
             keywords=['jarvis'],
             sensitivities=[1]
         )
-        self.kaldi_rec = vosk.KaldiRecognizer(vosk.Model("model_large"), 16000)
+        self.kaldi_rec = vosk.KaldiRecognizer(vosk.Model("data/model_small"), 16000)
 
     def main(self):
         self.recorder = PvRecorder(

utils/download_models.py

@@ -0,0 +1,26 @@
+import os
+import sys
+
+from data import config
+
+
+def install_vosk_model():
+    try:
+        open('data/model_small/README')
+    except Exception as e:
+        print(e)
+        if sys.platform == "linux" or sys.platform == "linux2":
+            os.system(f"wget https://alphacephei.com/vosk/models/{config.MODEL_NAME}.zip")
+            os.system(f"unzip {config.MODEL_NAME}.zip")
+            os.system(f"mv {config.MODEL_NAME} data/model_small")
+            os.system(f"rm -rf {config.MODEL_NAME}.zip")
+        elif sys.platform == "darwin":
+            os.system(f"curl https://alphacephei.com/vosk/models/{config.MODEL_NAME}.zip")
+            os.system(f"unzip {config.MODEL_NAME}.zip")
+            os.system(f"mv {config.MODEL_NAME} data/model_small")
+            os.system(f"rm -rf {config.MODEL_NAME}.zip")
+        elif sys.platform == "win32":
+            os.system(f"curl https://alphacephei.com/vosk/models/{config.MODEL_NAME}.zip --output 1.zip")
+            os.system('powershell -command "Expand-Archive 1.zip ./"')
+            os.system(f"rename {config.MODEL_NAME} data/model_small")
+            os.system("del /s /q 1.zip")

utils/play.py

@@ -15,7 +15,7 @@ def play(self, phrase, wait_done=True):
     if wait_done:
         self.recorder.stop()
     if filename:
-        wave_obj = sa.WaveObject.from_wave_file(f"{self.CDIR}/sound/{filename}")
+        wave_obj = sa.WaveObject.from_wave_file(f"{self.CDIR}/data/sound/{filename}")
         play_obj = wave_obj.play()
         if wait_done:
             play_obj.wait_done()

utils/tts.py

@@ -0,0 +1,31 @@
+import time
+
+import sounddevice as sd
+import torch
+
+language = 'ru'
+model_id = 'ru_v3'
+sample_rate = 48000  # 48000
+speaker = 'aidar'  # aidar, baya, kseniya, xenia, random
+put_accent = True
+put_yo = True
+device = torch.device('cpu')  # cpu или gpu
+text = "Хауди Хо, друзья!!!"
+
+model, _ = torch.hub.load(repo_or_dir='snakers4/silero-models',
+                          model='silero_tts',
+                          language=language,
+                          speaker=model_id)
+model.to(device)
+
+
+def va_speak(what: str):
+    audio = model.apply_tts(text=what + "..",
+                            speaker=speaker,
+                            sample_rate=sample_rate,
+                            put_accent=put_accent,
+                            put_yo=put_yo)
+
+    sd.play(audio, sample_rate * 1.05)
+    time.sleep((len(audio) / sample_rate) + 0.5)
+    sd.stop()