krishashok/transcribe.py

## transcribe.py
# The Daily Transcriber will take voice notes created on the day the script is run
# from iPhone's voice memo app (via iCloud) and transcribe them using Whisper and email
# the transcriptions to a destination email of choice.
# The best way to set this script up is as a daily cron job

# Requirements
# You will need to enable "Less Secure App Access" on your Gmail account security settings to be able to send emails
# You will also need to install Whisper (follow instructions here - https://github.com/openai/whisper)

# Imports
import datetime as dt
import os
import smtplib
from email.mime.multipart import MIMEMultipart
from email.mime.text import MIMEText
import whisper

# Email Credentials and Server Details
MY_EMAIL = 'YOUR SENDER EMAIL HERE' # Try and avoid using your actual email ID. Create a dummy one for this
MY_PASSWORD = 'YOUR PASSWORD HERE'
EMAIL_SERVER = 'YOUR EMAIL SERVER'
EMAIL_PORT = 587 # YOUR SERVER PORT HERE

# Load the whisper model (Change to 'small' or 'base' if the script takes too long)
model = whisper.load_model('medium')

# Voice memo path (Change as appropriate)
VOICE_MEMOS_PATH = "PATH TO VOICE MEMOS ON YOUR MACHINE"

# Get today's recordings and process them with Whisper
today = dt.datetime.now().date()
transcripts = []
for file in os.listdir(VOICE_MEMOS_PATH):
    file_created_time = dt.datetime.fromtimestamp(os.path.getctime(VOICE_MEMOS_PATH + file))
    if (file[-3:] == 'm4a') and (file_created_time.date() == today):
        transcript = {}
        print(f"processing {file}")
        result = model.transcribe(VOICE_MEMOS_PATH + file, fp16=False)
        transcript["timestamp"] = file_created_time.strftime('%-d %B,%Y %H:%M')
        transcript["text"] = result["text"]
        transcripts.append(transcript)

# Construct email body content
msg_content = "<html></body>"
for item in transcripts:
    msg_content += "<p>"
    msg_content += item["timestamp"]
    msg_content += "<br>"
    msg_content += item["text"]
    msg_content += "</p>"
msg_content += "</html></body>"

# Construct email message
msg = MIMEMultipart('alternative')
msg['From'] = MY_EMAIL
msg['To'] = 'RECIPIENT EMAIL'
msg['Subject'] = f'Transcribed Voice Memos for {today.strftime("%d-%B-%Y")}'
txt_part = MIMEText(msg_content, 'plain', 'utf-8')
html_part = MIMEText(msg_content, 'html', 'utf-8')
msg.attach(txt_part)
msg.attach(html_part)
msg_str = msg.as_string()

# Send email
with smtplib.SMTP(EMAIL_SERVER, EMAIL_PORT) as connection:
    connection.starttls()
    connection.login(user=MY_EMAIL, password=MY_PASSWORD)
    connection.sendmail(
        from_addr=MY_EMAIL,
        to_addrs='YOUR RECIPIENT EMAIL',
        msg=msg_str
    )
	# The Daily Transcriber will take voice notes created on the day the script is run
	# from iPhone's voice memo app (via iCloud) and transcribe them using Whisper and email
	# the transcriptions to a destination email of choice.
	# The best way to set this script up is as a daily cron job

	# Requirements
	# You will need to enable "Less Secure App Access" on your Gmail account security settings to be able to send emails
	# You will also need to install Whisper (follow instructions here - https://github.com/openai/whisper)

	# Imports
	import datetime as dt
	import os
	import smtplib
	from email.mime.multipart import MIMEMultipart
	from email.mime.text import MIMEText
	import whisper

	# Email Credentials and Server Details
	MY_EMAIL = 'YOUR SENDER EMAIL HERE' # Try and avoid using your actual email ID. Create a dummy one for this
	MY_PASSWORD = 'YOUR PASSWORD HERE'
	EMAIL_SERVER = 'YOUR EMAIL SERVER'
	EMAIL_PORT = 587 # YOUR SERVER PORT HERE

	# Load the whisper model (Change to 'small' or 'base' if the script takes too long)
	model = whisper.load_model('medium')

	# Voice memo path (Change as appropriate)
	VOICE_MEMOS_PATH = "PATH TO VOICE MEMOS ON YOUR MACHINE"

	# Get today's recordings and process them with Whisper
	today = dt.datetime.now().date()
	transcripts = []
	for file in os.listdir(VOICE_MEMOS_PATH):
	file_created_time = dt.datetime.fromtimestamp(os.path.getctime(VOICE_MEMOS_PATH + file))
	if (file[-3:] == 'm4a') and (file_created_time.date() == today):
	transcript = {}
	print(f"processing {file}")
	result = model.transcribe(VOICE_MEMOS_PATH + file, fp16=False)
	transcript["timestamp"] = file_created_time.strftime('%-d %B,%Y %H:%M')
	transcript["text"] = result["text"]
	transcripts.append(transcript)

	# Construct email body content
	msg_content = "<html></body>"
	for item in transcripts:
	msg_content += "<p>"
	msg_content += item["timestamp"]
	msg_content += "<br>"
	msg_content += item["text"]
	msg_content += "</p>"
	msg_content += "</html></body>"

	# Construct email message
	msg = MIMEMultipart('alternative')
	msg['From'] = MY_EMAIL
	msg['To'] = 'RECIPIENT EMAIL'
	msg['Subject'] = f'Transcribed Voice Memos for {today.strftime("%d-%B-%Y")}'
	txt_part = MIMEText(msg_content, 'plain', 'utf-8')
	html_part = MIMEText(msg_content, 'html', 'utf-8')
	msg.attach(txt_part)
	msg.attach(html_part)
	msg_str = msg.as_string()

	# Send email
	with smtplib.SMTP(EMAIL_SERVER, EMAIL_PORT) as connection:
	connection.starttls()
	connection.login(user=MY_EMAIL, password=MY_PASSWORD)
	connection.sendmail(
	from_addr=MY_EMAIL,
	to_addrs='YOUR RECIPIENT EMAIL',
	msg=msg_str
	)