PiSnort: A Delayed Alarm System for Raspberry Pi

Introduction and Overview

import RPi.GPIO as gpio
import random                   # For randomizing sound files
import time                     # To delay sound playing

<<setup logging>>

<<sound-playing logic>>

<<process management>>

<<handle exit logic>>

if __name__ == '__main__':
    log.info('Execution started.')

    play_sound('system/welcome.wav')

    <<define state>>

    <<define pins>>

    <<define sounds>>

    log.info('Beginning main loop.')
    try:
        while True:
            <<run loop>>
            time.sleep(2)
    except KeyboardInterrupt:
        log.info('Process ended with C-c.')

    log.info('Program exit.')

Exit Logic

import atexit
atexit.register(kill_child_processes)
atexit.register(logging.shutdown)
atexit.register(gpio.cleanup)
atexit.register(lambda:subprocess.call(['mplayer',
                                        '--volume=100',
                                        'sounds/system/deactivate.wav'],
                                       stdout=open('/dev/null'),
                                       stderr=open('/dev/null')))

Process Management

import subprocess

global child_processes
child_processes = set()

def start_process(cmd, purpose=None):
    child = subprocess.Popen(cmd)
    child.snort_purpose = purpose
    child_processes.add(child)
    return child

def gc_processes():
    for child in list(child_processes):
        child.poll()
        if child.returncode is not None:
            log.debug('Removing process %d from the set.', child.pid)
            child_processes.remove(child)
            if child.returncode is not 0:
                log.error('Process %d ended abnormally; exit code %d; purpose: %s',
                          child.pid,
                          child.returncode,
                          child.snort_purpose)

def kill_child_processes():
    gc_processes()
    for child in child_processes:
        try:
            log.info('Forcing process %d to exit', child.pid)
            child.terminate()
        except:
            log.error('Unable to kill child process %d', pid)

The Meat

Choose a random sound file to play and then wait twenty seconds. Play the file, then wait another five minutes to avoid a double alarm.

<<choose a random sound file to play>>
log.debug('Playing the random sound file')
play_sound(state.current_recording)

Choosing a Random Sound to Play

log.debug('Choosing a random sound file')
choices = ['recordings/'+key for key in snorts if bool(snorts[key][0])]
state.current_recording = random.choice(choices)

Implement lambda filtering of recordings

I seem to remember a requirement being the ability to play certain sounds. I’m not sure of the details of this requirement, but lambdas can be used here (using the [i for i in [...] if f(i)] construction).

Define and Initialize Pins

These are all of the pins used in the circuitry. Explicit ground wires are not listed; if something is missing, just complete the circuit. Refer to the Fritzing diagram if you’re unsure.

log.debug('Setting up GPIO')
gpio.setmode(gpio.BOARD)

pin_led = 7
pin_pir = 12
pin_btn_no = 24
pin_btn_nc = 26

gpio.setup(pin_led   , gpio.OUT, initial=gpio.LOW)
gpio.setup(pin_pir   , gpio.IN)
gpio.setup(pin_btn_no, gpio.IN , gpio.PUD_DOWN)
gpio.setup(pin_btn_no, gpio.IN , gpio.PUD_DOWN)

for p in [pin_led, pin_pir, pin_btn_no, pin_btn_nc]:
    log.debug('Setup pin %d for function %s', p, gpio.gpio_function(p))

Define Recording Locations

Trusted sources say these are mixed squirrel and deer sounds. They should be organized as such. This is not something I am trained to do (and I neither recorded nor named these files), so I’m waiting on the project manager to do this. They all sound like bloody murdur to me.

snorts = {'record2snort2.wav':     ('some criterion'),
          'record3snort1.wav':     ('some criterion'),
          'record3snort2.wav':     ('some criterion'),
          'record4snort1.wav':     ('some criterion'),
          'record4snort2.wav':     ('some criterion'),
          'recorded snort1-3.wav': ('some criterion'),
          'recorded snort1.wav':   ('some criterion'),
          'recorded snort2.wav':   ('some criterion')}

The actual sound files aren’t included in the project as they are decently large and won’t benefit from version control.

The Main Loop

log.log(0, 'Executing loop')

First, we need to update the internal state so we know we have current information. This will intelligently update properties such as button_depressed and motion_detected so we can use them later.

<<update state>>

It should be noted that the device won’t actually turn off; it will merely stop looking for motion.

When testing a new wiring, it’s good to just get a few tests of the LEDs and the speakers in. (I’ll probably be adding more tests later, but that’s beside the point.) If we are not testing, then check to see if we’ve detected motion. If we have, choose a random sound file to play and then play it.

if state.testing:
    <<testing logic>>
    state.testing = False
elif gpio.input(pin_led):
    if state.motion_detected is True:
        state.motion_detected = time.time()
    elif 40 < time.time() - state.motion_detected and state.current_recording is not None:
        state.current_recording = None
        gc_processes()
    elif 20 < time.time() - state.motion_detected and state.current_recording is None:
        <<play a random sound file>>
elif not gpio.input(pin_led):
    gc_processes()

Testing Logic

Toggle the LED ten times, waiting a second between each.

log.info('Running diagnostic...')
for i in range(4):
    log.debug('Toggling LED on pin {}'.format(pin_led))
    gpio.output(pin_led, not gpio.input(pin_led))
    time.sleep(1)
play_sound('system/diagnostic.wav')
log.info('Running diagnostic... Done')

Playing a Sound

Actually playing the sound is a pretty simple matter. Don’t forget to ensure that your speaker is hooked up to the Pi over 3.5mm.

def play_sound(relative_path):
    log = logging.getLogger('sounds')
    log.info('Playing sound: "%s"', relative_path)
    return start_process(['mplayer',
                          '-msglevel',
                          'all=-1',
                          'sounds/{!s}'\
                              .format(relative_path)],
                         'Play sound {}'\
                             .format(relative_path))

Keeping State

class State: pass
state=State()
state.testing              = False
state.button_depressed     = False
state.button_depressed_old = True
state.motion_detected      = False
state.current_recording    = None

Reading a New State

log.log(0, 'Updating state...')

Check for changes in button state. This logic should allow the system to manage a toggle on/off button for the system state using the momentary latch provided by the button.

state.button_depressed = gpio.input(pin_btn_no)

if state.button_depressed is not state.button_depressed_old:
    state.button_depressed_old = state.button_depressed
    if state.button_depressed:
        <<toggle state>>

Check for any input from the PIR. If motion is detected, set the appropriate flag.

log.debug(state.motion_detected)
if gpio.input(pin_pir) and state.motion_detected is False:
    log.debug('Motion detected')
    state.motion_detected = True
if time.time() - state.motion_detected > 60:
    log.debug('Motion terminated')
    state.motion_detected = False

log.log(0, 'Updating state... Done.')

On/Off System Toggling

log.debug('Toggling state...')

if gpio.input(pin_led):
    log.debug('Turning LED OFF')
    play_sound('system/deactivate.wav')
else:
    log.debug('Turning LED ON')
    play_sound('system/activate.wav')

gpio.output(pin_led, not gpio.input(pin_led))

log.debug('Toggling state... Done.')

Logging

import logging

logging.basicConfig(
    filename='log',
    format='%(asctime)s (%(name)s) [%(funcName)s] %(levelname)s: %(message)s'
)
console = logging.StreamHandler()
console.setLevel(logging.DEBUG)
console.setFormatter(logging.Formatter('%(levelname)-8s: %(message)s'))

log = logging.getLogger('')
log.addHandler(console)
log.setLevel(logging.DEBUG)

Resources

• Pin Map
• Using a PIR

Licensing

Macauley Library

Resources

recordings under /sounds/recordings/

The recordings under /sounds/recordings/ were released by the Macauley Library of Cornell University and are released under the following terms. Refer to the website for the definitive license.

List of Recordings

/sounds/recordings/??

CommonName

Coyote

Genus

Canis

Species

latrans

VocType

barks

RecordingCount

125888

Comments

Waller, Sara

/sounds/recordings/??

CommonName

Eastern Gray Squirrel

Genus

Sciurus

Species

carolinensis

VocType

call

RecordingCount

127048

Comments

Little, Randolph S.

/sounds/recordings/??

CommonName

Woodchuck

Genus

Marmota

Species

monax

VocType

alarm call

RecordingCount

55370

Comments

Gunn, William W. H.

/sounds/recordings/??

CommonName

Coyote

Genus

Canis

Species

latrans

VocType

barks

RecordingCount

50261

Comments

Keller, Geoffrey A.

/sounds/recordings/??

CommonName

White-tailed deer

Genus

Odocoileus

Species

virginianus

VocType

snort

RecordingCount

131198

Comments

Keller, Geoffrey A.

/sounds/recordings/??

CommonName

White-tailed deer

Genus

Odocoileus

Species

virginianus

VocType

snort

RecordingCount

120460

Comments

Fischer, Martha J.

/sounds/recordings/??

CommonName

Gray Wolf

Genus

Canis

Species

lupus

VocType

call

RecordingCount

128377

Comments

MacDonald, Stewart D.

/sounds/recordings/??

CommonName

White-tailed deer

Genus

Odocoileus

Species

virginianus

VocType

snort

RecordingCount

41832

Comments

Evans, William R.

/sounds/recordings/??

CommonName

White-tailed deer

Genus

Odocoileus

Species

virginianus

VocType

snort

RecordingCount

52604

Comments

Langtimm, Catherine A.

/sounds/recordings/??

CommonName

Woodchuck

Genus

Marmota

Species

monax

VocType

alarm call

RecordingCount

172798

Comments

McGowan, Jay

/sounds/recordings/??

CommonName

Cougar

Genus

Puma

Species

concolor

VocType

call

RecordingCount

126382

Comments

Priori, Andrea L.

/sounds/recordings/??

CommonName

Eastern Gray Squirrel

Genus

Sciurus

Species

carolinensis

VocType

call

RecordingCount

94227

Comments

Hershberger, Wilbur L

Macaulay Library Website Terms of Use

By visiting, viewing and/or using this website, you are agreeing to these Terms of Use.

This website (herein the “Website”) and all materials displayed or accessible through it, including but not limited to, text, photographs, images, illustrations, audio and video, computer software and code (herein called the “Content”) are protected by copyright and are owned by Cornell University, which includes the Macaulay Library (ML) and the Cornell Lab of Ornithology, its licensors or the party credited as the content provider.

Subject to the provisions listed in this Terms of Use Agreement, visitors to the Website are permitted to:

• View Content online.
• Print Website pages for non-commercial, personal, educational, and research uses provided that ML is properly cited as the source.
• Retain copies of specimen record data in digital form for non-commercial, personal, educational and research purposes provided that ML is properly cited as the source.
• Link to and share Website pages from third-party websites for non-commercial, personal, educational and research purposes only provided that ML is properly cited as the source.
• Share Content for non-commercial, personal, educational and research purposes provided that ML is properly cited as the source.

Any other uses (including but not limited to commercial, promotional, or administrative uses), reproduction, alteration, modification, public performance or display, uploading or posting onto the internet, transmission, redistribution or any other exploitation of the Website or the Content, whether in whole or in part, are prohibited without prior written permission.

Specimen Database Content

Content from the specimen database accessible through the Website must not be regarded as definitive or published. Distributional, ecological, taxonomic and other such data should be verified in consultation with ML curatorial staff. Sensitive data (e.g. locality data for threatened or endangered species) and some data records may be restricted from public access through the Website. Access to these data records may be requested in writing from appropriate curatorial staff.

Links to the Website

You may create links to this website for non-commercial (personal, educational and research) purposes only. If you wish to solicit a business relationship with the ML or the Cornell Lab of Ornithology, you need explicit written permission. While ML encourages links to the Website [link to “how to cite/link” page], it reserves the right to prohibit or refuse to accept any link to the Website at any time. You agree to remove any link you may have to the Website upon the request of ML.

Links From the Website

This website contains links to third-party sites (Third-Party Websites). ML provides these links only as a service and convenience to our visitors. We take no responsibility for the content at Third-Party Websites including, without limitation, any representation or warranty regarding the legality, accuracy, reliability, completeness, timeliness, or suitability of any content on any Third-Party Website. The link to them in no way implies an endorsement or recommendation of the products, services or information found there. ML reserves the right to regularly review and re-evaluate any and all links originating from within the Website and reserves the right to terminate any link to Third-Party Websites without prior notification.

Sharing of the Content (embedding)

• By using the embed code to display audio and video Content from the Website the Embeddable Content on Third-Party Websites, you agree to be bound by the following terms:
• Subject to the Terms of Use for the Website, ML grants you a non-exclusive, non-transferable license to use the embeddable computer code to display the Embeddable Content on Third-Party Websites. You may not copy, re-publish, edit, alter, add to or use the Embeddable Content, embed code or embeddable player in any other way.
• All title, ownership rights and intellectual property rights in and to the Embeddable Content, and any code made available by ML to embed the Embeddable Content and the embeddable player shall remain the property of ML and/or its licensors.
• The Embeddable Content is subject to same terms and conditions described in the “Links to the Website” section above.
• The Embeddable Content is for personal use only and cannot be used in any commercial way. You may not charge visitors any fee for accessing the Embeddable Content, use the Embeddable Content as means to secure advertising, or commercialize the Embeddable Content or embeddable player in any other way.
• You must properly attribute and create a link back to the ML website on the pages of the Third-Party Websites where the Embeddable Content is displayed.
• You may not directly or indirectly suggest any endorsement or approval by ML of the Third-Party Sites displaying the Embeddable Content or any non-ML entity, product or content or any views expressed within Third-Party Sites without the ML’s prior written approval.
• You may not use the Embeddable Content in any way that could bring the ML into disrepute or otherwise cause any loss or damage to the ML.
• You acknowledge that ML has sole editorial control over the Embeddable Content at all times and it may change or restrict, suspend or terminate any or all Embeddable Content or your access to the Embeddable Content at any time at its sole discretion without liability.
• You acknowledge that the Embeddable Content is made available by ML on an “as is” and “as available” basis and ML gives no warranty of any kind in relation to the Embeddable Content, embed code or the embeddable player.

Disclaimer of warranties

• The Website and all Content are provided “as is” for informational purposes only. By accessing and using the Website you acknowledge and agree that use of the Website and Content is entirely at your own risk. Where use of the Content is critical to scientific investigation, scholarly publication, or policy decisions, you are responsible for verifying the Content against primary data sources. Although ML applies the highest standards to Website performance, security, and data quality, ML makes no representations or warranties regarding the Website and Content, including, without limitation, no representation or warranty that:
• the Website and/or content will be accurate, complete, reliable, suitable or timely;
• any the Content, software, product or service made available through the Website will be of suitable for any particular purpose;
• the operation of the Website will be uninterrupted or error free;
• defects or errors in the Website will be corrected;
• the Website will be free from viruses or harmful components; and that
• communications to or from the Website will be secure or not intercepted.

Modifications to Terms of Use

ML reserves the right to modify the content of these terms of use at any time and it is your responsibility to consult the terms of use on a regular basis to determine whether any modifications have been made. By using the Website, you agree to all terms and conditions. Use of the Website after such changes are posted will signify your agreement to these revised terms.

Indemnification

You agree to indemnify, defend, and hold harmless Cornell University; its officers, directors, employees, agents, licensors, and third party providers to the Website from and against all losses, expenses, damages, and costs, including reasonable attorneys’ fees, resulting from any violation of these Terms. Cornell University reserves the right to assume the exclusive defense and control of any matter subject to indemnification by you, in which event you will fully cooperate with Cornell University in asserting and available defenses.