Merge pull request 'add_noisereduce' (#2) from add_noisereduce into master

Reviewed-on: #2

.gitignore

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

Jarvis.py

@@ -10,24 +10,21 @@ import yaml
 from fuzzywuzzy import fuzz
 from pvrecorder import PvRecorder
 
-from data import config
+import config
-from modules import HomeAssistant
+from utils import execute_cmd, play
-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('data/commands.yaml', encoding='utf8'))
+        self.VA_CMD_LIST = yaml.safe_load(open('commands.yaml', encoding='utf8'))
-        self.home_assistant = HomeAssistant.HomeAssistant()
         self.porcupine = pvporcupine.create(
             access_key=config.PICOVOICE_TOKEN,
             keywords=['jarvis'],
             sensitivities=[1]
         )
-        self.kaldi_rec = vosk.KaldiRecognizer(vosk.Model("data/model_small"), 16000)
+        self.kaldi_rec = vosk.KaldiRecognizer(vosk.Model("model_large"), 16000)
 
     def main(self):
         self.recorder = PvRecorder(
@@ -73,7 +70,6 @@ class Jarvis:
                 if vrt > rc['percent']:
                     rc['cmd'] = c
                     rc['percent'] = vrt
-                    rc['recognized_phrase'] = x
         if len(rc['cmd'].strip()) <= 0:
             return False
         elif rc['percent'] < 70 or rc['cmd'] not in self.VA_CMD_LIST.keys():
@@ -81,7 +77,7 @@ class Jarvis:
             time.sleep(1)
             return False
         else:
-            execute_cmd.execute_cmd(self, rc['cmd'], rc['recognized_phrase'], voice)
+            execute_cmd.execute_cmd(self, rc['cmd'])
             return True
 
     def play(self, phrase, wait_done=True):

commands.yaml

@@ -72,9 +72,3 @@ weather:
   - возможен дождь сегодня?
   - прогноз погоды на сегодня
   - погода
-home_assistant_execute:
-  - включи телевизор
-  - выключи телевизор
-  - начни уборку
-home_assistant_get:
-  - тест

config.py

@@ -0,0 +1,5 @@
+VA_ALIAS = ('джарвис',)
+VA_TBR = ('скажи', 'покажи', 'ответь', 'произнеси', 'расскажи', 'сколько', 'слушай')
+MICROPHONE_INDEX = -1
+PICOVOICE_TOKEN = "4xbwaZwZmSHeTiowFl5Rgqsc8CR4FKGV8YueJUlR4Zt2e1kB64IDcA=="
+OPENAI_TOKEN = "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"

data/config.py

@@ -1,10 +0,0 @@
-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=="
-
-
-# home assistant
-HOME_ASSISTANT_URL = "http://192.168.0.112:9999/api"
-HOME_ASSISTANT_TOKEN = "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpc3MiOiI5NjczNDZjYjc2YzI0YWQzODdhMmUwMmM2MjViZGVjZCIsImlhdCI6MTcxNDQ3MzkzNywiZXhwIjoyMDI5ODMzOTM3fQ.TATpIMXivJOioCtUI8PKg6gyTQYMG6bur6enm6NxjtY"

data/home_assistant_entities.yaml

@@ -1,4 +0,0 @@
-пылесос:
-  - entity_id:vacuum.roborock_vacuum_m1s
-  - state:находится в
-  - attributes.battery_level:а его уровень зарядки

main.py

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

modules/HomeAssistant.py

@@ -1,67 +0,0 @@
-import requests
-import yaml
-from fuzzywuzzy import process
-
-from data import config
-
-
-class HomeAssistant:
-    def __init__(self):
-        self.url = "http://192.168.0.112:9999/api"
-        self.token = config.HOME_ASSISTANT_TOKEN
-        self.HA_CMD_LIST = yaml.safe_load(open('data/home_assistant_entities.yaml', encoding='utf8'))
-
-    def get_info(self, state):
-        response = requests.get(
-            url=f"{self.url}/states",
-            headers={
-                "Authorization": "Bearer " + self.token
-            }
-        )
-        for entity in response.json():
-            if entity["entity_id"] == state:
-                return entity
-        return response
-
-    def send_process(self, command="выключи телевизор"):
-        response = requests.post(
-            url=f"{self.url}/services/conversation/process",
-            json={"text": command},
-            headers={
-                "Authorization": "Bearer " + self.token,
-                "content-type": "application/json"
-            },
-        )
-        if response.status_code == 200:
-            return True
-        return False
-
-    def voice_to_name(self, voice: str) -> str:
-        words = voice.lower().split()
-        best_match = None
-        highest_score = 0
-        for word in words:
-            result, score = process.extractOne(word, self.HA_CMD_LIST.keys())
-            if score > highest_score:
-                highest_score = score
-                best_match = result
-        return best_match
-
-    def validate_info(self, name: str):
-        answer = name
-        entity_config = self.HA_CMD_LIST.get(name)
-        if entity_config:
-            # Создание словаря, разделяя каждый элемент конфигурации на ключ и значение
-            entity_details = {item.split(':')[0]: item.split(':')[1] for item in entity_config}
-            entity_id = entity_details.pop("entity_id", None)
-            if entity_id:
-                responses = self.get_info(entity_id)
-                for attribute_path, label in entity_details.items():
-                    response = responses
-                    try:
-                        for attribute in attribute_path.split("."):
-                            response = response[attribute]
-                        answer += f" {label} {response}"
-                    except KeyError:
-                        continue
-        return answer

utils/download_models.py

@@ -1,26 +0,0 @@
-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/execute_cmd.py

@@ -1,4 +1,4 @@
-def execute_cmd(self, cmd: str, recognized_phrase: str, voice: str):
+def execute_cmd(self, cmd: str):
     if cmd == 'thanks':
         self.play("thanks")
     elif cmd == 'stupid':
@@ -7,9 +7,3 @@ def execute_cmd(self, cmd: str, recognized_phrase: str, voice: str):
         self.play("off", True)
         self.porcupine.delete()
         exit(0)
-    elif cmd == 'home_assistant_execute':
-        self.home_assistant.send_process(recognized_phrase)
-    elif cmd == 'home_assistant_get':
-        entity_name = self.home_assistant.voice_to_name(voice)
-        entity_info = self.home_assistant.validate_info(entity_name)
-        print(entity_info)

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}/data/sound/{filename}")
+        wave_obj = sa.WaveObject.from_wave_file(f"{self.CDIR}/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()