-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathkurrwa_led_weather_bulb.py
201 lines (156 loc) · 5.62 KB
/
kurrwa_led_weather_bulb.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
# Simple test program which listens to a wake word
# and after detecting it turns on LED
import RPi.GPIO as GPIO
import time
import pvporcupine as porcu
import sounddevice as sd
import numpy as np
import pyaudio
import speech_recognition as sr
import requests
import pyttsx3
from flux_led import WifiLedBulb
from bs4 import BeautifulSoup
# Set the pins
GPIO.setmode(GPIO.BCM)
pin_led = 14
GPIO.setup(pin_led, GPIO.OUT)
# Set the wake word to smth for now
pico_key_path = '/home/stetsenko/macode/pico_pico.txt'
with open(pico_key_path, "r") as file:
pico_key = file.read()
print('hey1')
# wake_word_path = "/home/stetsenko/.local/lib/python3.9/site-packages/pvporcupine/resources/keyword_files/raspberry-pi/porcupine_raspberry-pi.ppn"
model_path_pl = "/home/stetsenko/.local/lib/python3.9/site-packages/pvporcupine/lib/common/porcupine_params_pl.pv"
porcupine = porcu.create(keywords=['hej-kurwa'], access_key = pico_key, model_path = model_path_pl)
# Define audio settings
sample_rate = 16000 # Sample rate in Hz
frame_length = 512 # Frame length in number of samples
# Create an instance of PyAudio
audio_interface = pyaudio.PyAudio()
# Open the microphone stream
stream = audio_interface.open(
format=pyaudio.paInt16,
channels=1,
rate=sample_rate,
frames_per_buffer=frame_length,
input=True
)
def get_next_audio_frame():
audio_frame = stream.read(frame_length)
return np.frombuffer(audio_frame, dtype=np.int16)
def turn_on_led(some_led):
GPIO.output(some_led, GPIO.HIGH)
def turn_off_led(some_led):
GPIO.output(some_led, GPIO.LOW)
def kurrwa_led_greet(some_led):
turn_on_led(some_led)
sd.sleep(100)
turn_off_led(some_led)
sd.sleep(200)
def control_bulbs(command):
speak("Let the bulb shine")
if "bed" in command:
bulb_ip = '192.168.88.23'
else:
print("default is bed")
bulb_ip = '192.168.88.23'
bulb = WifeLedBulb(bulb_ip)
bulb.setRgb(1,0,0)
#flux_led bulb_ip -w 75 -1
turn_on_led(pin_led)
sd.sleep(3)
turn_off_led(pin_led)
# Function to fetch the current temperature
def get_current_temperature(location):
url = f"https://weather.com/en-IN/weather/today/l/{location}"
headers = {"User-Agent": "Mozilla/5.0"}
response = requests.get(url, headers=headers)
if response.status_code == 200:
soup = BeautifulSoup(response.content, "html.parser")
temperature = soup.find("span", {"data-testid": "TemperatureValue"}).text
return temperature
else:
return None
# Function to recognize speech after the wake word
def recognize_command():
recognizer = sr.Recognizer()
with sr.Microphone() as source:
print("Say something...")
audio = recognizer.listen(source)
try:
recognized_text = recognizer.recognize_google(audio, language="en-US")
print(recognized_text)
return recognized_text
except sr.UnknownValueError:
print("Speech could not be recognized.")
return ' '
# Function to speak out the given text
def speak(text):
engine = pyttsx3.init()
engine.say(text)
engine.runAndWait()
def speak_polish(text):
engine = pyttsx3.init()
engine.setProperty('voice', 'polish')
engine.say(text)
engine.runAndWait()
# Function to speak out the given text with a funny voice
def speak_funny(text):
engine = pyttsx3.init()
# Get a list of available voices
voices = engine.getProperty('voices')
# Select a funny voice (the voice index may vary depending on your system)
funny_voice_index = 1
if len(voices) > funny_voice_index:
# Set the selected voice
engine.setProperty('voice', voices[funny_voice_index].id)
# Set other voice properties for a funny effect (if available)
# For example, you can adjust the pitch and rate of speech
engine.setProperty('rate', 95) # Slower speech rate
engine.setProperty('pitch', 10) # Higher pitch for a funny voice
engine.say(text)
engine.runAndWait()
while True:
audio_frame = get_next_audio_frame()
keyword_index = porcupine.process(audio_frame)
if keyword_index == 0:
print("hej kurwa")
speak_polish("Ja pierdolę")
kurrwa_led_greet(pin_led)
command_detected = False
while not command_detected:
audio_frame = get_next_audio_frame()
command = recognize_command()
command_led_on = "turn on"
command_led_off = "turn off"
go_weather = "weather"
stop_listen = "goodbye"
location = "07029"
if command:
if command_led_on in command:
print("Command detected")
turn_on_led(pin_led)
audio_frame = np.array([], dtype=np.int16)
elif command_led_off in command:
print("Command detected")
turn_off_led(pin_led)
audio_frame = np.array([], dtype=np.int16)
elif go_weather in command:
print("Fetching the weather data")
temperature = get_current_temperature(location)
if temperature:
temp_info = f"The current temperature is {temperature}C."
print(temp_info)
# speak(temp_info)
speak_funny(temp_info)
else:
print("Failed to fetch weather data.")
elif stop_listen in command:
command_detected = True
speak_polish("jyake bydło jebane")
elif "shine" in command:
control_bulbs(command)
else:
speak_funny(command)
porcupine.delete()