adolforeyna/mcp_screen
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix I2C scan on Waveshare RLCD board using SoftI2C and implement button PTT trigger fallback when touch is missing

Browse Source
This commit is contained in:
Adolfo Reyna
2026-06-17 22:40:21 -04:00
parent 31363734c5
commit 82de0f328b
1 changed files with 112 additions and 85 deletions
Download Patch File Download Diff File Expand all files Collapse all files
main.py
+110 -83
View File
@@ -146,9 +146,19 @@ def main():
# Last user actions # Last user actions
last_action_str = "Boot finished." last_action_str = "Boot finished."
force_dashboard_redraw = True force_dashboard_redraw = True
voice_assistant_active = False
def is_talk_trigger_active():
if touch:
return touch.is_touched()
elif buttons and buttons.key:
return buttons.key.is_pressed()
return False
# 5. Register Button Handlers # 5. Register Button Handlers
def on_key_click(): def on_key_click():
if voice_assistant_active:
return
global local_led_mode_idx global local_led_mode_idx
local_led_mode_idx = (local_led_mode_idx + 1) % len(led_modes) local_led_mode_idx = (local_led_mode_idx + 1) % len(led_modes)
mode_name, color_fn, mode_type = led_modes[local_led_mode_idx] mode_name, color_fn, mode_type = led_modes[local_led_mode_idx]
@@ -204,9 +214,10 @@ def main():
last_action_str = f"Wi-Fi Connected: {ip_addr}" last_action_str = f"Wi-Fi Connected: {ip_addr}"
force_dashboard_redraw = True force_dashboard_redraw = True
# Check touch interaction if available (Push-to-talk Voice Assistant) # Check voice assistant trigger (touch screen or physical key button)
if touch and touch.is_touched(): if is_talk_trigger_active():
print("Touch detected! Starting Hermes Voice Assistant...") print("Voice assistant trigger detected! Starting Hermes Voice Assistant...")
voice_assistant_active = True
def draw_status_bar(text): def draw_status_bar(text):
y_bar = display.height - 25 y_bar = display.height - 25
@@ -216,55 +227,65 @@ def main():
display.show() display.show()
def clear_status_bar(): def clear_status_bar():
display.fill_rect(0, 215, 320, 25, 0) y_bar = display.height - 25
display.fill_rect(0, y_bar, display.width, 25, 0)
display.show() display.show()
# State 1: Wait for user to release their finger from the initial touch draw_status_bar("PTT Voice: Initializing...")
draw_status_bar("Release finger...")
# State 1: Wait for user to release the initial trigger press
draw_status_bar("Release button/screen...")
release_start = time.ticks_ms() release_start = time.ticks_ms()
while touch.is_touched(): while is_talk_trigger_active():
if time.ticks_diff(time.ticks_ms(), release_start) > 2000: if time.ticks_diff(time.ticks_ms(), release_start) > 2000:
print("Release timeout occurred") print("Release timeout occurred")
break break
time.sleep_ms(30) time.sleep_ms(30)
# State 2: Wait for tap-to-start recording # State 2: Wait for tap-to-start recording
draw_status_bar("Ready. Tap screen to record...") draw_status_bar("Ready. Tap key/screen to record...")
tap_started = False tap_started = False
start_wait = time.ticks_ms() start_wait = time.ticks_ms()
while time.ticks_diff(time.ticks_ms(), start_wait) < 10000: # 10s timeout to start recording while time.ticks_diff(time.ticks_ms(), start_wait) < 10000: # 10s timeout
if touch.is_touched(): if is_talk_trigger_active():
tap_started = True tap_started = True
break break
time.sleep_ms(30) time.sleep_ms(30)
if tap_started: if tap_started:
# User tapped the screen. Let's record! # User tapped. Let's record!
draw_status_bar("Recording: 15s (Tap to stop)") draw_status_bar("Recording: 15s (Tap to stop)")
# 1. Start MCLK PWM and configure ES8311 mic path # 1. Start MCLK PWM and configure mic path based on board config
mclk_pin = Pin(4, Pin.OUT) mclk_pwm = None
mclk_pwm = machine.PWM(mclk_pin) if board_config.audio_mclk_pin is not None:
mclk_pwm.freq(6144000) mclk_pin = Pin(board_config.audio_mclk_pin, Pin.OUT)
mclk_pwm.duty_u16(32768) mclk_pwm = machine.PWM(mclk_pin)
mclk_pwm.freq(board_config.audio_mclk_freq)
mclk_pwm.duty_u16(32768)
codec = ES8311(i2c) if board_config.audio_mic_codec == "ES7210":
if codec.init(sample_rate=16000): from audio_util import ES7210
codec.set_volume(80) # Set a balanced volume to prevent speaker saturation codec = ES7210(i2c)
try: codec.init(sample_rate=16000, bit_width=16)
codec._write(0x14, 0x1A) # Enable analog mic input & PGA else:
codec._write(0x16, 0x01) # Enable +6dB gain boost for mic recording from audio_util import ES8311
codec._write(0x17, 0xC8) # Set ADC digital volume codec = ES8311(i2c)
except Exception as e: if codec.init(sample_rate=16000):
print("Failed to set mic gain:", e) codec.set_volume(80)
try:
codec._write(0x14, 0x1A) # Enable analog mic input & PGA
codec._write(0x16, 0x01) # Enable +6dB gain boost
codec._write(0x17, 0xC8) # Set ADC digital volume
except Exception as e:
print("Failed to set mic gain:", e)
# 2. Configure I2S RX for recording (Mono 16kHz) # 2. Configure I2S RX for recording (Mono 16kHz)
i2s_rx = I2S(1, i2s_rx = I2S(1,
sck=Pin(5), sck=Pin(board_config.audio_i2s_sck),
ws=Pin(7), ws=Pin(board_config.audio_i2s_ws),
sd=Pin(6), sd=Pin(board_config.audio_i2s_rx_sd),
mode=I2S.RX, mode=I2S.RX,
ibuf=8000, ibuf=8000,
rate=16000, rate=16000,
@@ -280,7 +301,6 @@ def main():
rec_start_time = time.ticks_ms() rec_start_time = time.ticks_ms()
max_rec_duration_ms = 15000 # 15 seconds max duration max_rec_duration_ms = 15000 # 15 seconds max duration
max_rec_bytes = 480000 # 15 seconds of mono 16kHz 16-bit PCM (32KB/s) max_rec_bytes = 480000 # 15 seconds of mono 16kHz 16-bit PCM (32KB/s)
last_sec_left = -1
try: try:
while True: while True:
@@ -290,20 +310,14 @@ def main():
print("Recording stopped: maximum duration/size reached") print("Recording stopped: maximum duration/size reached")
break break
# Update countdown
sec_left = max(0, 15 - (elapsed // 1000))
if sec_left != last_sec_left:
last_sec_left = sec_left
draw_status_bar(f"Recording: {sec_left}s (Tap to stop)")
# Read I2S chunk # Read I2S chunk
bytes_read = i2s_rx.readinto(buffer) bytes_read = i2s_rx.readinto(buffer)
if bytes_read > 0: if bytes_read > 0:
audio_chunks.append(bytes(buffer[:bytes_read])) audio_chunks.append(bytes(buffer[:bytes_read]))
total_data_bytes += bytes_read total_data_bytes += bytes_read
# Check touch state # Check release and stop-tap
touched = touch.is_touched() touched = is_talk_trigger_active()
if not start_tap_released: if not start_tap_released:
if not touched: if not touched:
start_tap_released = True start_tap_released = True
@@ -316,10 +330,11 @@ def main():
print("Error recording:", e) print("Error recording:", e)
finally: finally:
i2s_rx.deinit() i2s_rx.deinit()
try: if board_config.audio_mic_codec == "ES8311":
codec._write(0x16, 0x00) # Reset mic gain back to 0dB immediately to prevent playback saturation try:
except: codec._write(0x16, 0x00) # Reset mic gain
pass except:
pass
if total_data_bytes >= 1000: if total_data_bytes >= 1000:
draw_status_bar("Sending & processing...") draw_status_bar("Sending & processing...")
@@ -335,53 +350,60 @@ def main():
"X-Device-ID": "kitchen-button" "X-Device-ID": "kitchen-button"
} }
print(f"Posting WAV ({len(wav_data)} bytes) to Hermes API...")
try: try:
res = urequests.post("http://192.168.68.126:8642/api/esp32/voice", headers=headers, data=wav_data) import urequests
print(f"HTTP Status: {res.status_code}") res = urequests.post("https://hermes.adolforeyna.com/api/voice",
data=wav_data,
headers=headers,
timeout=30)
if res.status_code == 200: if res.status_code == 200:
response_data = res.content response_data = res.content
print(f"Received response: {len(response_data)} bytes")
if response_data[0:4] == b'RIFF' and response_data[8:12] == b'WAVE': # Parse response WAV header
idx = response_data.find(b'fmt ') if len(response_data) >= 44 and response_data[0:4] == b'RIFF' and response_data[8:12] == b'WAVE':
if idx != -1: draw_status_bar("Playing response...")
fmt_chunk = response_data[idx:idx+24]
audio_format, channels, sample_rate, byte_rate, block_align, bits = struct.unpack('<HHIIHH', fmt_chunk[8:24])
print(f"Playing response WAV: {sample_rate}Hz, {channels}ch, {bits}bit")
data_idx = response_data.find(b'data') # Parse channel count, sample rate, bits
if data_idx != -1: import struct
audio_start = data_idx + 8 fmt_chunk_size = struct.unpack('<I', response_data[16:20])[0]
channels, sample_rate, byte_rate, block_align, bits = struct.unpack('<HHIIH', response_data[22:34])
draw_status_bar("Speaking...") audio_start = 20 + fmt_chunk_size
while audio_start < len(response_data) - 8:
if response_data[audio_start:audio_start+4] == b'data':
audio_start += 8
break
chunk_size = struct.unpack('<I', response_data[audio_start+4:audio_start+8])[0]
audio_start += 8 + chunk_size
# Play response if audio_start < len(response_data):
amp_pin.value(0) # Enable Amp (Active Low) # Play response using board_config parameters
i2s_format = I2S.MONO if channels == 1 else I2S.STEREO on_val = 0 if board_config.audio_amp_active_level == 0 else 1
i2s_tx = I2S(1, off_val = 1 if board_config.audio_amp_active_level == 0 else 0
sck=Pin(5), amp_pin.value(on_val) # Enable Amp
ws=Pin(7),
sd=Pin(8), i2s_format = I2S.MONO if channels == 1 else I2S.STEREO
mode=I2S.TX, i2s_tx = I2S(1,
ibuf=4096, sck=Pin(board_config.audio_i2s_sck),
rate=sample_rate, ws=Pin(board_config.audio_i2s_ws),
bits=bits, sd=Pin(board_config.audio_i2s_tx_sd),
format=i2s_format) mode=I2S.TX,
try: ibuf=4096,
pos = audio_start rate=sample_rate,
chunk_size = 2048 bits=bits,
while pos < len(response_data): format=i2s_format)
chunk = response_data[pos:pos+chunk_size] try:
i2s_tx.write(chunk) pos = audio_start
pos += chunk_size chunk_size = 2048
finally: while pos < len(response_data):
time.sleep_ms(150) chunk = response_data[pos:pos+chunk_size]
amp_pin.value(1) # Disable Amp i2s_tx.write(chunk)
i2s_tx.deinit() pos += chunk_size
finally:
time.sleep_ms(150)
amp_pin.value(off_val) # Disable Amp
i2s_tx.deinit()
else: else:
# MP3 warning message on display
draw_status_bar("Error: Gateway returned MP3") draw_status_bar("Error: Gateway returned MP3")
time.sleep(3) time.sleep(3)
else: else:
@@ -393,17 +415,22 @@ def main():
time.sleep(2) time.sleep(2)
# Deinit MCLK PWM # Deinit MCLK PWM
mclk_pwm.deinit() if mclk_pwm:
try:
mclk_pwm.deinit()
except:
pass
# Debounce touch release at the very end of wizard # Debounce release at the very end of wizard
clear_status_bar() clear_status_bar()
release_end = time.ticks_ms() release_end = time.ticks_ms()
while touch.is_touched(): while is_talk_trigger_active():
if time.ticks_diff(time.ticks_ms(), release_end) > 2000: if time.ticks_diff(time.ticks_ms(), release_end) > 2000:
break break
time.sleep_ms(30) time.sleep_ms(30)
time.sleep_ms(200) time.sleep_ms(200)
voice_assistant_active = False
last_action_str = "Voice query finished" last_action_str = "Voice query finished"
force_dashboard_redraw = True force_dashboard_redraw = True