Pyframe is a Python-based digital photo frame application. It is capable of displaying photos and playing videos from local storage as well as WebDAV and rclone repositories.
Pyframe has been designed to run slideshows from image and video repositories with several thousand files. No conversion is required. Files remain in your repositories and fully under your control.
Files in slideshows can be dynamically arranged and filtered based on their metadata (EXIF and IPTC metadata supported). Slideshows can be run continuously or scheduled.
Pyframe supports reverse geocoding based on GPS data in the EXIF tag, using the geopy library and Photon geocoder (essentially OpenStreetMap).
Pyframe optionally integrates with Home Assistant via MQTT. Integration allows the display to be motion activated after coupling of the Pyframe device with a motion sensor.
Pyframe is being developed by Bernd Kalbfuss (aka langweiler) and is published under the General Public License version 3. The latest source code is available on GitHub.
Instructions for building your own digital photo frame can be found here.
Pyframe requires Python 3 to run. It has been developed with Python version 3.10 on Ubuntu Linux, but may run with earlier versions and on different operating systems.
Pyframe requires the following Python packages to be installed:
- exifread
- ffmpeg-python
- geopy
- IPTCInfo3
- Kivy
- paho-mqtt
- rclone-python
- schedule
- SQLAlchemy
- webdavclient3
- yaml
All packages are available on pypi.org and can be installed using the "pip install" (or "pip3 install") command. Where possible/available, packages should be installed using the distribution package manager (e.g "apt" on Debian/Ubuntu).
Pyframe further requires the following (non-Python) libraries to be installed:
- libxslt1.1
- libmtdev1
- libsqlite3-0
- libsdl2-2.0-0
- ffmpeg
Libraries should be installed using the distribution package manager.
Note that Pyframe requires the X windows system and a desktop environment to be installed. Pyframe will in principle also run under Wayland, but the display will not be turned off automatically since Wayland does fully implement the "xset" command.
Pyframe is still in early development. The easiest way to install the latest version is to clone the GitHub repository using the git client. After having installed the git client, installation of Pyframe becomes as simple as:
$ git clone git@github.com:kalbfuss/pyframe.gitThe command installs the Pyframe sources in the sub-directory "pyframe" within the current working directory. Pyframe can be updated to the latest version by changing into the "pyframe" directory and issuing the following command:
$ cd pyframe
$ git pull origin masterAt this stage of the project you should not expect the configuration syntax to be stable. Please, have a look at the documentation after each update and adjust the configuration as necessary.
The Pyframe application is configured via a single YAML configuration file. The file is named "config.yaml" and must be stored in the current (working) directory. The following sections provide examples for configuration and the documentation of all parameters.
A lot of effort has gone into configuration checks. The application should warn you in the event of invalid configurations immediately after startup. It is thus safe to explore the various configuration options. Under no circumstances is Pyframe going to modify any of your image or video files.
In this example, we want to continuously show all files stored in a local directory. For this purpose, we configure a single local repository ("Local storage"). Our files are stored under the relative path "./local/photos". We further define a single slideshow ("Favorites") containing all files from the repository. Files are shown in a random sequence for a duration of 60 s.
Per (application) default settings, the repository is indexed once after start of the application. The slideshow includes photos and videos. The slideshow starts playing after start of the application and the display is always on.
repositories:
# Local repository with our favorite photos and videos.
Local storage:
type: local
root: ./local/photos
slideshows:
# Slideshow with our favorite photos and videos.
Favorites:
repositories: Local storage
pause: 60
sequence: randomIn this example, we want to show our most recent photos stored in the cloud in the period from 8:00 to 10:00 and our favorite photos, which are stored locally, in the period from 18:00 to 20:00. Since we are not necessarily at home in the evening, we want the display to be motion activated during this time.
Firstly, we define two (enabled) repositories: A local repository ("Local storage") with files stored under the relative path "./local/photos" and a WebDAV repository ("Cloud storage") with files stored in the cloud. The third repository ("Test repository") used for testing has been disabled. Per the repository default settings, the index of the local repository is updated at the start of the application and every 24 hours. The index of the cloud repository is updated daily at 23:00.
Secondly, we define two slideshows: The first slideshow ("Favorites") includes files tagged as "favorites" from the local repository. Files are shown for a duration of 60 s. The second slideshow ("Recent") includes the 200 most recent files from the cloud repository, which are not tagged as "vacation" or "favorites". We further limit files to "images". Files are sorted by the creation date in ascending order. Per the slideshow defaults, images are shown for a duration of 180 s.
The slideshow defaults further ensure that files tagged as "private" are always excluded. Files are labeled with their description from the file metadata (if available) and labels are shown for a duration of 60 s at the start and the end of each file. Since the display is installed in vertical orientation, we rotate the content by -90° and limit content to files in portrait orientation.
Thirdly, we define a schedule to show the second slideshow ("Recent") in the time from 8:00 to 10:00 and the first slideshow ("Favorites") in the time from 18:00 to 20:00. In the first case, the display is always on. In the second case, the display is motion activated with a timeout interval of 300 s.
Finally, since we run Home Assistant and need the MQTT remote control for the motion activation feature, we configure an MQTT client connection. For the motion activation feature to function properly, we further have to link the touch button with a motion sensor in Home Assistant (see [motion activation](#Motion activation)).
repositories:
# Local repository with our favorite photos and videos.
Local storage:
type: local
root: ./local/photos
# WebDAV repository with the latest photos from our smartphone.
Cloud storage:
type: webdav
url: https://mycloud.mydomain.org
root: /remote.php/webdav/photos
user: pyframe
password: <password>
index_update_at: "23:00"
# Test repository, which has been disabled.
Test repository:
type: local
root: ./local/test
enabled: false
# Repository defaults
index_update_interval: 24
slideshows:
# Slideshow with our favorite photos and videos.
Favorites:
repositories: Local storage
pause: 60
tags: favorites
# Slideshow with most recent photos from our smartphone.
Recent:
repositories: Cloud storage
excluded_tags:
- vacation
- favorites
types: images
most_recent: 200
order: date
direction: descending
# Slideshow defaults
always_excluded_tags: private
label_content: description
label_mode: auto
label_duration: 30
orientation: portrait
pause: 180
rotation: -90
schedule:
# Play the slideshow "Recent" in the period from 8:00 to 10:00.
morning start:
time: "08:00"
slideshow: Recent
display_mode: static
play_state: playing
morning stop:
time: "10:00"
play_state: stopped
# Play the slideshow "Favorites" in the period from 18:00 to 20:00.
# Activate the display by motion.
evening start:
time: "18:00"
slideshow: Favorites
display_mode: motion
display_timeout: 300
play_state: playing
evening stop:
time: "20:00"
play_state: stopped
mqtt:
host: mqtt.local
user: pyframe
password: <my password>
device_name: My pyframe somwhere in the houseThe following parameters are used to configure the application.
| Parameter | Description |
|---|---|
| window_size | The size of the window provided as [width, height]. A value of "full" enables full screen mode. The default is "full". |
| display_mode | The following display modes are supported. The default is "static". - static: The display is always on if a slideshow is paused or playing and off if a slideshow is stopped. - motion: The display is turned on and the slideshow starts playing in the presence of motion (i.e. touch events). The slideshow is paused and the display turned off in the absence of motion after the display timeout interval. |
| display_timeout | The time in seconds after which the slideshow is paused and screen turned off in the absence of motion. The default is 300 seconds. |
Parameters in this section will likely not have to be modified by the majority of users.
| Parameter | Description |
|---|---|
| index | The index database file. The path may be absolute or relative to the current working directory. The default is "./index.sqlite". |
| cache | The directory in which files can be cached (used by WebDAV and rclone repositories). The directory path may be absolute or relative to the current working directory. The directory can be shared by multiple repositories. Do not use directory in which you store files as cache directory. The default is "./cache". |
| enable_exception_handler | Set to true in order to enable the generic exception handler. The generic exception handler prevents the application from exiting unexpectedly. Exceptions are logged, but the execution continues. The default is false. |
| enable_scheduler | Set to false in order to disable the scheduler. The scheduler is disabled even in the presence of a schedule configuration section. The default is true. |
| enable_mqtt | Set to false in order to disable the MQTT client. The client is disabled even in the presence of an mqtt configuration section. The default is true |
| enable_logging | Set to false in order to disable logging. The default is true. |
| log_level | The log level, which can be set to debug, info, warning, or error. The default is "warning". |
| log_dir | The directory to which log files are written. The directory path may be absolute or relative to the current working directory. The default is "./log". |
Pyframe supports the configuration of one or multiple file repositories. Repositories are configured in the repositories section of the configuration file. The section is required and must contain at least a single, valid repository definition. Repository parameter defaults may be provided as global parameters. The example below provides a typical repositories configuration section.
...
repositories:
# Local repository with our favorite photos and videos.
Local storage:
type: local
root: ./local/photos
# WebDAV repository with the latest photos from our smartphone.
Cloud storage:
type: webdav
url: https://mycloud.mydomain.org
root: /remote.php/webdav/photos
user: pyframe
password: <password>
# Test repository, which has been disabled.
Test repository:
type: local
root: ./local/test
enabled: false
# Repository defaults
index_update_interval: 24
...The following parameters are used to configure repositories.
| Parameter | Description |
|---|---|
| type | The following repository types are supported. A values must be provided. - local: Repository with files on the local file system. Note: Even if referred to as local, files may be stored on a network share as long as the network is mounted and integrated into the file system hierarchy (e.g. "/mnt/photos"). - rclone: Repository with files on an rclone remote. The remote must have been configured before using the "rclone config" command or directly in the rclone configuration file. - webdav: Repository with files on a WebDAV accessible site (e.g. ownCloud or NextCloud). |
| index_update_interval | Interval in hours at which the metadata index for the repository is updated. If zero, the index is only updated once after start of the application. The default ist 0. Do not use in combination with index_update_at. |
| index_update_at | The time at which the metadata index for the repository is updated. The index is updated once per day. Do not use in combination with index_update_interval. |
| enabled | Set to false in order to disable the repository. The default is true. |
Only a single parameter is required for the definition of local repositories.
| Parameter | Description |
|---|---|
| root | The repository root directory. Root directories may be absolute or relative to the current working directory. Files in sub-folders will be included in the repository. A value must be provided. |
Like for local repositories, only a single parameter is required for the definition of rclone repositories. However, the rclone remote must have been configured before. Pyframe currently does not provide any functionality to configure rclone remotes.
| Parameter | Description |
|---|---|
| root | The rclone remote and root directory (e.g. "mycloud:/photos/"). Files in sub-folders will be included in the repository. A value must be provided. |
As a minimum, the parameters url, user and password need to be specified for the definition of a WebDAV repository.
| Parameter | Description |
|---|---|
| url | The URL of the WebDAV server. Use "https://" protocol prefix for secure connections. A value must be provided. |
| user | Login name. A value must be provided. |
| password | Login password. A value must be provided. |
| root | The root directory relative to the URL. For ownCloud WebDAV access, the root directoy typically starts with "/remote.php/webdav". The default is /. |
Pyframe supports the configuration of one or multiple slideshows. Slideshows are configured in the slideshows section of the configuration file. The section is required and must contain at least a single, valid slideshow definition. The first slideshow is the default slideshow. Slideshow parameter defaults may be provided as global parameters. The example below provides a typical slideshows configuration section.
...
slideshows:
# Slideshow with our favorite photos and videos.
Favorites:
repositories: Local storage
pause: 60
tags: favorites
# Slideshow with most recent photos from our smartphone.
Recent:
repositories: Cloud storage
excluded_tags:
- vacation
- favorites
types: images
most_recent: 200
order: date
direction: descending
# Slideshow defaults
always_excluded_tags: private
label_content: description
label_mode: auto
label_duration: 30
orientation: portrait
pause: 180
rotation: -90
...The following parameters are used to configure slideshows.
| Parameter | Description |
|---|---|
| bg_color | The background color used to fill empty areas, provided as [r, g, b]. The default is [1, 1, 1] (white). |
| label_content | The following content based on file meta data is supported. The default is "full". - description: only image description - short: image description, location, and creation date - full: image description, location, creation date and tags, file name and repository |
| label_duration | Duration in seconds for which labels are shown. The default is 60. |
| label_font_size | The relative font size of labels, expressed as percentage of the shortest file dimension. The default is 0.08. |
| label_mode | The following label modes are supported. The default is "off". - auto: Labels are shown at the beginning and end of a file for the label_duration. - off: Labels are never shown. - on: Labels are always shown. |
| label_padding | The relative padding of labels, expressed as percentage of the shortest file dimension. The default is 0.03. |
| pause | The delay in seconds until the next file is shown. The default is 300. |
| resize | The following resize modes are supported. The default is "fill". - fit: The slideshow content is zoomed to fit the screen as good as possible. Empty areas are filled with the background color. - fill: The slideshow content is zoomed and cropped to completely fill the screen. Note that images which do not have the same orientation as the screen are not zoomed and cropped, but only fit to the screen. |
| rotation | The angle by which slideshow content is rotated clockwise. Useful for picture frames/screens, which are installed in non-standard orientation. The default is 0. |
The following parameters control the files included in a slideshow and the sequence in which they are shown. The default is to include all files from all repositories. Files are sorted by their name in ascending order.
| Parameter | Description |
|---|---|
| repositories | The repositories from which files shall be shown. The default is to show files from all repositories. |
| orientation | Valid orientations are portrait or landscape. The default is to include either orientation. |
| types | Supported file types are images and videos. May be a single value or list of values. The default is to include all file types. |
| tags | File tags, which shall be included. May be a single value or list of values. The default is to include all tags and untagged files. If set, untagged files are excluded. |
| excluded_tags | File tags, which shall be excluded. May be a single value or list of values. The default is not to exclude any tags. |
| always_excluded_tags | Same as excluded_tags, but not overwritten by an excluded_tags statement. Use in the slideshow default configuration to exclude certain tags in all slideshows (e.g. private content). |
| most_recent | Files in the slideshow are limited to the most_recent number of files based on the creation date after application of all other filter criteria. |
| order | The sort order in which files are shown. The default is "name". - date: Files are sorted by their creation date. - name: Files are sorted by their name. - random: Files are shown in a random sequence. - smart: A short sequence with random starting point, sorted by date in ascending order. |
| direction | Valid sort directions are ascending or descending. The default is "ascending". Ignored if random order is configured. |
| smart_limit | The (maximum) number of files in a smart sequence. If the smart_time criterion is not met, the sequence may be shorter. The default is 10. |
| smart_time | The maximum time allowed in-between subsequent files of a smart sequence in hours. If exceeded, the sequence is terminated early and a new sequence initiated. The default is 24. |
Pyframe supports the configuration of a schedule. The schedule allows to alter the application behavior at predefined points in time. The schedule is configured in the optional schedule section of the configuration file. The schedule may contain one or multiple events. The schedule is disabled if the configuration section is missing. The example below provides a typical schedule configuration section.
schedule:
# Play the slideshow "Recent" in the period from 8:00 to 10:00.
morning start:
time: "08:00"
slideshow: Recent
display_mode: static
play_state: playing
morning stop:
time: "10:00"
play_state: stopped
# Play the slideshow "Favorites" in the period from 18:00 to 20:00.
# Activate the display by motion.
evening start:
time: "18:00"
slideshow: Favorites
display_mode: motion
display_timeout: 30
play_state: playing
evening stop:
time: "20:00"
play_state: stoppedThe following parameters are used to configure events in the schedule.
| Parameter | Description |
|---|---|
| time | The time of the event. A value must be provided. Always specify in quotation marks. Note: Hours and minutes <10 must be preceded by a 0, i.e. "08:03" and never "8:3". |
| slideshow | Selected slideshow. If no slideshow is specified, the previous or default slideshow is assumed. |
| play_state | Valid play states are paused, playing and stopped. The play state remains unchanged if no value is provided. The default is "stopped". |
| display_mode | The following display modes are supported. The display mode remains unchanged if no value is provided. The default is "static". - static: The display is always on if a slideshow is paused or playing and off if a slideshow is stopped. - motion: The display is turned on and the slideshow starts playing in the presence of motion. The slideshow is paused and the display turned off in the absence of motion after the display timeout interval. |
| display_timeout | The time in seconds after which the slideshow is paused and screen turned off in the absence of motion. The display timeout remains unchanged if no value is provided. The default is 300. |
Pyframe implements an MQTT client, which registers the device with an MQTT broker. The MQTT configuration is provided in the optional mqtt section of the configuration file. MQTT support is disabled if the configuration section is missing. The example below provides a typical mqtt configuration section.
...
mqtt:
host: <hostname of MQTT broker>
user: <login name>
password: <my password>
device_name: My pyframe somwhere in the house
...The following parameters are used to configure the MQTT client.
| Parameter | Description |
|---|---|
| host | Hostname of the MQTT broker. A value must be specified. |
| port | Connection port of MQTT broker. The default is 8883 (standard for secure connections). |
| tls | The following values are supported. The default is true. - true: A TLS-encrypted secure connection is used. - false: A non-encrypted connection is used. |
| tls_insecure | The following values are supported. The default is false. - true: Insecure TLS connections with non-trusted certificates are permitted. - false: Only secure connections with trusted certificates are permitted. |
| user | Login name. A value must be provided. |
| password | Login password. A value must be provided. |
| device_id | The Pyframe device ID. The default is "pyframe". Note The device ID must be unique. A different value must be specified if multiple Pyframe instances connect to the same broker. |
| device_name | The human friendly device name. The default is to use the device_id. |
Once Pyframe has been configured, you can change into the Pyframe directory and start the application with the following command:
$ python3 pyframe.pyIn recent distributions you may have to use "python" instead of "python3". Unless configured otherwise, Pyframe is going to create an index database "index.sqlite" and directory "./log" for log files in the Pyframe directory. If WebDAV or rclone repositories are configured, Pyframe will further create a directory "./cache" for temporary storage of downloaded files.
For convenience you can install the following script, which will allow you to start the Pyframe application from anywhere (even SSH sessions). The placeholders and evidently need to be replaced with the proper values prior to running the script.
/usr/local/bin/start-pyframe
#!/bin/sh
USER=<your user>
SRC=<your pyframe directory>
# May be python or python3 depending on your distribution
PYTHON=/usr/bin/python3
# Set authority file and active display in case we are starting this script from an SSH session.
export XAUTHORITY=/home/$USER/.Xauthority
export DISPLAY=:0
# Change to pyframe source directory
cd $SRC
# Start pyframe
if [ $(/usr/bin/whoami) = 'root' ]; then
/usr/sbin/runuser -u langweiler -- $PYTHON pyframe.py
else
$PYTHON pyframe.py
fiIf you intend to run Pyframe as systemd service, you can optionally create a second script for clean up after termination. In this example, we turn off the screen (works only under X11, not Wayland).
/usr/local/bin/stop-pyframe
#!/bin/sh
USER=<your user>
# Set authority file and active display in case we are starting this script from an SSH session.
export XAUTHORITY=/home/$USER/.Xauthority
export DISPLAY=:0
# Turn backlight off.
/usr/bin/xset dpms force offBoth scripts should be owned by root.root and need to be executable (mode 750).
If you want to start Pyframe automatically during system boot, you can do so by configuring it in your desktop session manager. Alternatively, you can register a systemd service via a unit file. Below is an example for a unit file, which uses the two scripts we created before.
/etc/systemd/system/pyframe.service
[Unit]
Description=Pyframe digital photo frame
Wants=graphical.target
After=graphical.target
[Service]
Type=simple
ExecStart=/usr/local/bin/start-pyframe
ExecStop=/usr/local/bin/stop-pyframe
User=root
Group=root
Restart=always
[Install]
WantedBy=default.targetThe Wants and After statements make sure that we are in graphical mode and that the service is only started after the graphical system has been launched. The ExecStop script is optional as stated above. It is not required to stop the Pyframe service. The Restart statement ensures that Pyframe is restarted after unexpected exit. The WantedBy statement allows to start the service automatically at boot time.
Make sure the unit file belongs to root.root, is readable by the owner and group and writable by the owner only (mode 640). Afterwards you can start the service and verify the successful start via the following commands:
$ sudo systemctl start pyframe
$ sudo systemctl status pyframeTo enable automatic start of the service during boot time issue the following command:
$ sudo systemctl enable pyframeMake sure to additionally configure autologin for the user under which you intend to run Pyframe. Steps for configuration depend on the graphical system and Linux distribution. Under Armbian you can use the "armbian-config" tool. On Raspberry Pi OS, the "raspi-config" tool will do. For other systems/distributions consult the corresponding documentation.
Pyframe implements basic support for integration with the Home Assistant home automation system. Integration is achieved through the built-in Home Assistant MQTT integration. As an additional pre-requisite, an MQTT broker must be installed (e.g. Eclipse Mosquitto).
After the Pyframe MQTT client has been correctly configured and a connection to the MQTT broker established, Pyframe should automatically appear as a new device in Home Assistant. The device supports several push buttons and configuration selections, which allow you to control Pyframe from remote. The device further provides a file sensor, whose value is identical to the UUID of the currently displayed file.
In addition, the file sensor provides selected file metadata as sensor attributes.
For motion activation of the display, the touch button of the Pyframe device needs to be coupled to a motion sensor via an automation. Every time motion is detected, the touch button is pressed by the automation. Pressing the touch button activates the display and resets the display timeout counter.
