Download all written notes and recorded audio from a LiveScribe Pulse or Echo pen. Convert the proprietary formats to open formats (pdf and ogg vorbis). Make the notes available online through a web app.
This program has four parts:
- dumpscribe: Downloads data from pen
- unmuddle: Converts data to open formats
- usb_watcher: Runs dumpscribe and unmuddle when smartpen connects
- dumpscribe-web: A web app for browsing the converted data
The dumpscribe command downloads and extracts all relevant info from the smartpen in the raw format as it exists on the pen. The audio is in aac format. The written notes are in a proprietary format called stf and the meta-data is in a combination of XML and undocumented binary files.
./dumpscribe [-d] output_dir -d: Enable debug output. -c: Delete files from pen after successful download.
Where output_dir is a directory where the downloaded data will be saved.
The unmuddle.py command converts written notes to PDF, converts audio data to Ogg Vorbis, parses meta-data and puts everything into a sane folder structure. It can optionally generate PNG thumbnails of the PDFs and combine all pages into a single PDF.
Usage: unmuddle.py [-h] [--aac] [--notebook] [--thumb] [--thumbsize THUMBSIZE] input_dir output_dir Convert file formats and organize output from dumpscribe. positional arguments: input_dir The directory generated by dumpscribe. output_dir Where to write the output from this program. optional arguments: -h, --help show this help message and exit --aac Don't convert aac audio files to ogg vorbis. (conversion required either ffmpeg or avconv) --notebook Additionally create one pdf per notebook with all notebook pages (requires pdftk). --thumb Generate png thumbnails of pdfs (requires either ImageMagick or GraphicsMagick). --thumbsize THUMBSIZE Set thumbnail maximum dimension.
The PDFs generated by the official LiveScribe software use background images to make them look like the actual notebook paper (with lines or grids). After running dumpscribe, you will have two background files for each of the types of notebooks you have written in so far. E.g:
To get unmuddle.py to use these backgrounds, make a backgrounds/ directory in the same directory as the unmuddle.py script and copy the pngs like so:
Unfortunately it is not clear how the official livescribe software figures out which backgrounds go with which notebooks, so for now this is manual. If you really need this feature then the best solution is to always use the same type of notebook.
The usb_watcher.py command waits for a smartpen to be connected to USB and then runs first dumpscribe, then unmuddle.py and optionally a user-supplied command at the end (e.g. to upload the resulting data to a server). It is meant to be run as a daemon.
Usage: usb_watcher.py [-h] [-d] [-c POST_COMMAND] dumpscribe_dir dumpscribe_output_dir organized_output_dir [cleanup_dir] Automatically run dumpscribe when LiveScribe pen is connected. positional arguments: dumpscribe_dir The full path to the directory where dumpscribe is installed. dumpscribe_output_dir Where dumpscribe should place its output. organized_output_dir Where unmuddle should place its output. cleanup_dir Optional directory to clean up if drive is more than 50 percent full. optional arguments: -h, --help show this help message and exit -d Daemonize this process -l Enable Beagle Bone Black LED control -c POST_COMMAND Command to run after running unmuddle.py
The -l argument only works on a Beagle Bone Black with the Adafruit Beagle Bone BPIO python library installed. If enabled, it will use led_control.py to indicate the current status using three LEDs as output. Modify led_control.py to change which GPIO pins to use and remember that the Beagle Bone Black GPIO pins cannot supply enough current to drive LEDs directly so you'll need to use transistors. In my setup led1 is green, led2 is purple and led3 is red.
If cleanup_dir is specified, files are deleted from that directory (by the cleanup.py script) until the usage is under 50% on the device where the directory resides or until there are no more files in the directory. If deleting a file causes a directory to become empty, the empty directory is also deleted.
Running as a daemon
To run usb_watcher.py as a daemon that automatically starts on boot, first copy the the init script:
sudo cp init_scripts/dumpscribe-usb-watcher /etc/init.d/
Then make the script start on boot:
sudo update-rc.d dumpscribe-usb-watcher defaults
A simple node.js web app that lets users browse the written and recorded notes.
Installing requirements for dumpscribe:
sudo apt-get install build-essential pkg-config libglib2.0-0 libglib2.0-dev libopenobex1 libopenobex1-dev libusb-1.0-0 libusb-1.0-0-dev libarchive13 libarchive-dev libxml2 libxml2-dev
Installing requirements for basic functionality for unmuddle.py:
sudo apt-get install python-cairosvg
If you want to use thumbnail generation, audio transcoding and generation of per-notebook pdfs, additionally you need:
sudo apt-get install graphicsmagick avconv pdftk
For usb_watcher.py the following packages are needed:
sudo apt-get install python-gobject python-gudev
Installing requirements for usb_watcher.py:
sudo apt-get install python-daemon
For the web app you need:
sudo apt-get install nodejs npm cd web/ npm install
The dumpscribe command is a C program and needs to be compiled. Just run:
The web app needs a settings file. Do:
cp web/settings.js.example web/settings.js
Then edit the web/settings.js file to at least change the admin password.
Running the web app as a daemon
If you want to run the web app as a daemon, e.g. on a VPS so everyone can access it, then follow the notes in this section.
Making the web app auto-start
This section will show how to make the dumpscribe web app auto-start on boot. This method will also make the web app automatically restart if it crashes.
This section assumes that you have already gotten the dumpscribe web app running on the server by downloading it, installing requirements and editing the config file. See previous sections for instructions.
This section assumes that dumpscribe is located in /opt/dumpscribe
First check if /usr/bin/node exists:
If it doesn't, make a symlink (otherwise the upstart script won't run):
cd /usr/bin sudo ln -s nodejs node
Install "psy" (a node.js app that automatically auto-starts your process when it crashes):
sudo npm -g install psy
Add a user that will be running dumpscribe (just hit enter when asked questions):
sudo adduser dumpscribe --disabled-password
Make the dumpscribe dir owned by the dumpscribe user:
sudo chown -R dumpscribe.dumpscribe /opt/dumpscribe
Now copy the web app's init script:
sudo cp init_script/dumpscribe-web /etc/init.d
Then, edit /etc/init.d/dumpscribe-web to set the correct user and paths for your system.
Start the dumpscribe web app with:
sudo /etc/init.d/dumpscribe-web start
Check if it's really working by visiting the web app in the browser at http://your-server.org:3000/ or seeing if the dumpscribe process is runnning:
To make the process auto-start on boot do:
sudo update-rc.d dumpscribe-web defaults
Using an apache reverse proxy
Having users type a URL ending in :3000 is not very nice (and besides, some corporate firewall will block port 3000). Setting up a reverse proxy is the way to go.
If you haven't already, install a apache2:
sudo apt-get install apache2
Enable some necessary apache2 modules:
sudo a2enmod proxy sudo a2enmod proxy_http sudo a2enmod alias
Now edit the file apache virtual hosts file for the host where you want dumpscribe-web to appear. If you just now installed apache2, edit the file /etc/apache2/sites-enabled/000-default.conf. Assuming your hostname is myhostname.org and you want the dumpscribe web app to appear at "http://myhostname.org/labnotes" add the following lines immediately before the line:
RedirectMatch /labnotes$ http://myhostname.org/labnotes/ ProxyRequests Off ProxyPass /labnotes/ http://127.0.0.1:3000/ ProxyPassReverse /labnotes/ http://127.0.0.1:3000/ ProxyPreserveHost on
Note that the trailing slashes are important.
Now restart apache:
sudo /etc/init.d/apache2 restart
Make sure the dumpscribe web app is running, open the url http://myhostname.org/labnotes in a browser to verify that everything is working.
During boot only LED1 (green) is on. When boot has completed, all LEDs should be off.
When the pen is docked, the following process occurs:
LED1 (green) is blinking while the data is downloading from the pen. When data download is complete, LED1 (green) stays on. LED2 (purple) then starts blinking as data is the data is converted to open formats. When conversion is complete, LED2 (purple) stays on. When the user command is running (data is uploading) LED3 (red) is blinking. When the user command completes (upload completed) LED3 (red) stays on for one second, followed by all LEDs turning off.
- If only LED1 (green) is on, then usb_watcher crashed or never started
- If LED1 (green) is blinking for a long time then downloading from the pen failed
- If LED1 (green) is on and LED2 (purple) is blinking for a long time, then converting to open formats failed.
- If LED1 (green) and LED2 (purple) are on and LED3 (red) is blinking for a long time, then running the user-specified command failed (the user-specified command is likely uploading the files to your server).
- Remove my ssh key from BBB and make BBB autogen new key in /root/.ssh/id_rsa* on first boot
- Get rid of dumpscribe memory leaks
- It looks like the obex downloads allocates memory that is only freed when the obex cleanup function is called (which disconnects).
- Use Valgrind to check for memory leaks: http://www.cprogramming.com/debugging/valgrind.html
- Get rid of dumpscribe compile warnings related to xmlChar vs. char
License and Copyright
This project is based on, and contains code from, the following projects:
This project is licensed under GPLv2. For more info see the COPYING file.
The code used in dumpscribe and unmuddle.py has had multiple contributors. Not all of them have identified themselves clearly.
- Copyright 2010 to 2011 Steven Walter (https://github.com/srwalter)
- Copyright 2010 Scott Hassan
- Copyright 2011 jhl (?)
- Copyright 2011 Nathanael Noblet (firstname.lastname@example.org)
- Copyright 2011 to 2013 Dylan M. Taylor (email@example.com)
- Copyright 2013 Yonathan Randolph (firstname.lastname@example.org)
- Copyright 2014 Robert Jordens (https://github.com/jordens)
- Copyright 2014 Ali Neishabouri (email@example.com)
- Copyright 2014-2015 Marc Juul (firstname.lastname@example.org)
- Copyright 2019 Eric Johnson (yut23 on github)