Flask-based EyeFi Server
Simple Flask-based Python3 EyeFi server
- Documentation: eyeflask.readthedocs.org
I use an Eye-Fi SD card in my portable scanner. Unfortunately, it used to upload directly to Evernote, but no longer supports uploading directly to any service that suits my needs. Additionally, they don’t provide a Linux version of their server software.
EyeFlask is a simple Flask-based Eye-Fi server written for Python >= 3.4. The Eye-Fi card can connect to it and will upload images to the folder specified in the config. EyeFlask attempts to verify the file integrity using the same security protocols used by Eye-Fi Server.
- Python >= 3.4
- See requirements.txt
Using a virtual env is highly recommended. Since EyeFlask is Python 3.4+ only, there’s no excuse not to use one!
- python3 -m venv venv on some systems may need to install venv for Python3.4 (e.g. sudo apt-get install python3.4-venv)
- venv/bin/pip3 install eyeflask
- Copy eyeflask/extras/eyeflask-sample.cfg to eyeflask.cfg, modify with your values, and put it in your instance folder
- Run: venv/bin/eyeflask
- Scan some stuff, see if it ends up in your uploads folder
- Clone the repo: git clone https://github.com/n8henrie/eyeflask && cd eyeflask
- Make a virtualenv: python3 -m venv venv
- Make an instance folder: mkdir -p instance
- Copy the config sample: cp eyeflask/extras/eyeflask-sample.cfg instance/eyeflask.cfg
- Edit the config to include your upload directory and upload_key (see below): vim instance/eyeflask.cfg
- Install with dev dependencies: venv/bin/pip install .[dev]
- Run: eyeflask (or venv/bin/python -m eyeflask.cli)
- Scan some stuff, see if it ends up in your uploads folder
- UPLOAD_KEY: See FAQ, below
- UPLOAD_FOLDER: Destination folder for uploads. May optionally be a strftime string to specify a subfolder format (e.g. `/path/to/eyeflask-uploads/%Y/%m-%d/’)
EyeFlask will help get the images uploaded and extracted to your server (e.g. a Raspberry Pi in my case), but what do to from there? If you’re running Raspbian Jessie (and using systemd), I’ve included in the extras folder a few files that may be of interested.
- upload_scans.service will run a given script when called (e.g. sudo systemctl start upload_scans.service)
- upload_scans.timer is an example systemd timer unit that will call upload_scans.service every 10 minutes
- upload_scans.path is an example systemd path unit that will call upload_scans.service every time a file changes in a watched directory
Put together, this makes it really easy to put together a script to upload new scans to a Dropbox folder whenever a new one is added, or whatever command you’d like to run on all your scans.
I’ve also included eyeflask/extras/eyeflask.service, which is a sample systemd service file to run EyeFlask at startup and restart it on errors.
Much of the code for EyeFlask came from or was inspired by the following projects / links. Many thanks to the authors for their work! If I’ve forgotten anyone, let me know.
Troubleshooting / FAQ
Where do I find my upload key?
You’ll need a supported platform (OS X or Windows) with Eye-Fi Center.app installed, and need to have uploaded photos to that computer at least once. This ensures everything is working, and generates the Settings.xml file, from which you need to copy the upload key into eyeflask.cfg.
Is it okay to be running this with the built-in Flask server?
It’s not perfect, but it seems to work okay for me and my single Eye-Fi card setup. You’d probably be better off running it behind gunicorn or a gunicorn / nginx setup, but I’m running it behind Flask alone for simplicity and because I haven’t had any issues so far.
If you want to give it a go with gunicorn / nginx, I’ve included an extremely simplified nginx configuration file: eyeflask/extras/nginx.conf. After installing gunicorn into your virtualenv, hopefully you’ll be able to get it running behind nginx without much trouble with something like:
venv/bin/pip install gunicorn venv/bin/gunicorn 'eyeflask:create_app("instance/eyeflask.cfg")'
For debugging you can also use the flags --log-file=- --log-level=debug.
NB: I do not plan on providing support for nginx / gunicorn setups, as I don’t know enough about it to be particularly helpful. I just verified that it seemed to work. (Just FYI, Gunicorn without nginx did not seem to work unless I used one of the async workers, kept getting timeouts.)
Why am I getting repeat or unreliable file uploads on my Raspberry Pi?
I’m not sure. I was getting excellent reliability when running EyeFlask on my Macbook Air and very poor reliability on my Raspberry Pi B+ with EyeFlask 0.1.0. It seemed like the file would be uploaded exctracted without issue, but the EyeFi card kept sending the same file over and over, leading me to believe that the confirmation response wasn’t getting received every time. I thought it might have something to do with slow response times, so I did a little code optimization with 0.1.1 which seems to have helped. I also gave up and put EyeFlask behind a gunicorn / nginx setup, and between the two of these changes I have much better upload reliability.
0.1.3 :: 20170310
- Add strftime based subfolders to config (see issue #2)
- Encourage use of virtualenv in the README
- Test docs creation in Travis
- Fix some flake8 compliants
0.1.2 :: 20170303
- Python 3.6 compatibility
- Stop using src/ subdirectory
- Try to fix some Pandoc exceptions
0.1.1 :: 20160318
- Use array.array instead of struct.iter_unpack for modest speed boost
- Rename start_session to handle_SOAP – because that’s what it does
- Extract the image from the tarfile data prior to writing to disk (eliminating the need to delete the tarfile afterwards)
0.1.0 :: 20160227
- Initial release to GitHub, PyPI
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
|Filename, size||File type||Python version||Upload date||Hashes|
|Filename, size eyeflask-0.1.3-py3-none-any.whl (20.6 kB)||File type Wheel||Python version py3||Upload date||Hashes View|
|Filename, size eyeflask-0.1.3.tar.gz (23.8 kB)||File type Source||Python version None||Upload date||Hashes View|