Asyncio UPnP Client library for Python/asyncio.
Written initially for use in Home Assistant to drive DLNA DMR-capable devices, but useful for other projects as well.
The UPnP Device Architecture document contains several sections describing different parts of the UPnP standard. These chapters/sections can mostly be mapped to the following modules:
- Chapter 1 Discovery
- Section 1.1 SSDP:
async_upnp_client.ssdp
- Section 1.2 Advertisement:
async_upnp_client.advertisement
provides basic functionality to receive advertisements. - Section 1.3 Search:
async_upnp_client.search
provides basic functionality to do search requests and gather the responses. async_upnp_client.ssdp_client
contains theSsdpListener
which combines advertisements and search to get the known devices and provides callbacks on changes. It is meant as something which runs continuously to provide useful information about the SSDP-active devices.
- Section 1.1 SSDP:
- Chapter 2 Description / Chapter 3 Control
async_upnp_client.client_factory
/async_upnp_client.client
provide a series of classes to get information about the device/services using the 'description', and interact with these devices.async_upnp_client.server
provides a series of classes to set up a UPnP server, including SSDP discovery/advertisements.
- Chapter 4 Eventing
async_upnp_client.event_handler
provides functionality to handle events received from the device.
There are several 'profiles' which a device can implement to provide a standard interface to talk to. Some of these profiles are added to this library. The following profiles are currently available:
- Internet Gateway Device (IGD)
async_upnp_client.profiles.igd
- Digital Living Network Alliance (DLNA)
async_upnp_client.profiles.dlna
- Printers
async_upnp_client.profiles.printer
For examples on how to use async_upnp_client
, see examples
/ .
Note that this library is most likely does not fully implement all functionality from the UPnP Device Architecture document and/or contains errors/bugs/mis-interpretations.
See CONTRIBUTING.rst
.
Development is done on the development
branch.
pre-commit
is used to run several checks before committing. You can install pre-commit
and the git-hook by doing:
$ pip install pre-commit
$ pre-commit --install
The Open Connectivity Foundation provides a bundle with all UPnP Specifications.
Changes are recorded using Towncier. Once a new release is created, towncrier is used to create the file CHANGES.rst
.
To create a new change run:
$ towncrier create <pr-number>.<change type>
A change type can be one of:
- feature: Signifying a new feature.
- bugfix: Signifying a bug fix.
- doc: Signifying a documentation improvement.
- removal: Signifying a deprecation or removal of public API.
- misc: A ticket has been closed, but it is not of interest to users.
A new file is then created in the changes
directory. Add a short description of the change to that file.
Steps for releasing:
- Switch to development:
git checkout development
- Do a pull:
git pull
- Run towncrier:
towncrier build --version <version>
- Commit towncrier results:
git commit -m "Towncrier"
- Run bump2version:
bump2version --tag major/minor/patch
- Note that this creates a new commit + tag.
- Run bump2version:
- Checkout tag:
git checkout <version>
- Build:
rm -rf build dist && python setup.py build sdist
- Upload using twine:
twine upload dist/*
- Switch to master:
git checkout master
- Merge tag to master:
git merge <version>
- Switch to development:
git checkout development
- Push to github:
git push --all && git push --tags
To do profiling it is recommended to install pytest-profiling. Then run a test with profiling enabled, and write the results to a graph:
# Run tests with profiling and svg-output enabled. This will generate prof/*.prof files, and a svg file.
$ pytest --profile-svg -k test_case_insensitive_dict_profile
...
# Open generated SVG file.
$ xdg-open prof/combined.svg
Alternatively, you can generate a profiling data file, use pyprof2calltree to convert the data and open kcachegrind. For example:
# Run tests with profiling enabled, this will generate prof/*.prof files.
$ pytest --profile -k test_case_insensitive_dict_profile
...
$ pyprof2calltree -i prof/combined.prof -k
launching kcachegrind
A command line interface is provided via the upnp-client
script. This script can be used to:
- call an action
- subscribe to services and listen for events
- show UPnP traffic (--debug-traffic) from and to the device
- show pretty printed JSON (--pprint) for human readability
- search for devices
- listen for advertisements
The output of the script is a single line of JSON for each action-call or subscription-event. See the programs help for more information.
An example of calling an action:
$ upnp-client --pprint call-action http://192.168.178.10:49152/description.xml RC/GetVolume InstanceID=0 Channel=Master
{
"timestamp": 1531482271.5603056,
"service_id": "urn:upnp-org:serviceId:RenderingControl",
"service_type": "urn:schemas-upnp-org:service:RenderingControl:1",
"action": "GetVolume",
"in_parameters": {
"InstanceID": 0,
"Channel": "Master"
},
"out_parameters": {
"CurrentVolume": 70
}
}
An example of subscribing to all services, note that the program stays running until you stop it (ctrl-c):
$ upnp-client --pprint subscribe http://192.168.178.10:49152/description.xml \*
{
"timestamp": 1531482518.3663802,
"service_id": "urn:upnp-org:serviceId:RenderingControl",
"service_type": "urn:schemas-upnp-org:service:RenderingControl:1",
"state_variables": {
"LastChange": "<Event xmlns=\"urn:schemas-upnp-org:metadata-1-0/AVT_RCS\">\n<InstanceID val=\"0\">\n<Mute channel=\"Master\" val=\"0\"/>\n<Volume channel=\"Master\" val=\"70\"/>\n</InstanceID>\n</Event>\n"
}
}
{
"timestamp": 1531482518.366804,
"service_id": "urn:upnp-org:serviceId:RenderingControl",
"service_type": "urn:schemas-upnp-org:service:RenderingControl:1",
"state_variables": {
"Mute": false,
"Volume": 70
}
}
...
You can subscribe to list of services by providing these names or abbreviated names, such as:
$ upnp-client --pprint subscribe http://192.168.178.10:49152/description.xml RC AVTransport
An example of searching for devices:
$ upnp-client --pprint search
{
"Cache-Control": "max-age=3600",
"Date": "Sat, 27 Oct 2018 10:43:42 GMT",
"EXT": "",
"Location": "http://192.168.178.1:49152/description.xml",
"OPT": "\"http://schemas.upnp.org/upnp/1/0/\"; ns=01",
"01-NLS": "906ad736-cfc4-11e8-9c22-8bb67c653324",
"Server": "Linux/4.14.26+, UPnP/1.0, Portable SDK for UPnP devices/1.6.20.jfd5",
"X-User-Agent": "redsonic",
"ST": "upnp:rootdevice",
"USN": "uuid:e3a17dd5-9d85-3131-3c34-b827eb498d72::upnp:rootdevice",
"_timestamp": "2018-10-27 12:43:09.125408",
"_host": "192.168.178.1",
"_port": 49152
"_udn": "uuid:e3a17dd5-9d85-3131-3c34-b827eb498d72",
"_source": "search"
}
An example of listening for advertisements, note that the program stays running until you stop it (ctrl-c):
$ upnp-client --pprint advertisements
{
"Host": "239.255.255.250:1900",
"Cache-Control": "max-age=30",
"Location": "http://192.168.178.1:1900/WFADevice.xml",
"NTS": "ssdp:alive",
"Server": "POSIX, UPnP/1.0 UPnP Stack/2013.4.3.0",
"NT": "urn:schemas-wifialliance-org:device:WFADevice:1",
"USN": "uuid:99cb221c-1f15-c620-dc29-395f415623c6::urn:schemas-wifialliance-org:device:WFADevice:1",
"_timestamp": "2018-12-23 11:22:47.154293",
"_host": "192.168.178.1",
"_port": 1900
"_udn": "uuid:99cb221c-1f15-c620-dc29-395f415623c6",
"_source": "advertisement"
}
IPv6 is supported for the UPnP client functionality as well as the SSDP functionality. Please do note that multicast over IPv6 does require a scope_id
/interface ID. The scope_id
is used to specify which interface should be used.
There are several ways to get the scope_id
. Via Python this can be done via the ifaddr library. From the (Linux) command line the scope_id
can be found via the ip command:
$ ip address
...
6: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000
link/ether 00:15:5d:38:97:cf brd ff:ff:ff:ff:ff:ff
inet 192.168.1.2/24 brd 192.168.1.255 scope global eth0
valid_lft forever preferred_lft forever
inet6 fe80::215:5dff:fe38:97cf/64 scope link
valid_lft forever preferred_lft forever
In this case, the interface index is 6 (start of the line) and thus the scope_id
is 6
.
Or on Windows using the ipconfig
command:
C:\> ipconfig /all
...
Ethernet adapter Ethernet:
...
Link-local IPv6 Address . . . . . : fe80::e530:c739:24d7:c8c7%8(Preferred)
...
The scope_id
is 8
in this example, as shown after the %
character at the end of the IPv6 address.
Or on macOS using the ifconfig
command:
% ifconfig
...
en0: flags=8863<UP,BROADCAST,SMART,RUNNING,SIMPLEX,MULTICAST> mtu 1500
options=50b<RXCSUM,TXCSUM,VLAN_HWTAGGING,AV,CHANNEL_IO>
ether 38:c9:86:30:fe:be
inet6 fe80::215:5dff:fe38:97cf%en0 prefixlen 64 secured scopeid 0x4
...
The scope_id
is 4
in this example, as shown by scopeid 0x4
. Note that this is a hexadecimal value.
Be aware that Python <3.9
does not support the IPv6Address.scope_id
attribute. As such, a AddressTupleVXType
is used to specify the source
- and target
-addresses. In case of IPv4, AddressTupleV4Type
is a 2-tuple with address
, port
. AddressTupleV6Type
is used for IPv6 and is a 4-tuple with address
, port
, flowinfo
, scope_id
. More information can be found in the Python socket
module documentation.
All functionality regarding SSDP uses AddressTupleVXType
the specify addresses.
For consistency, the AiohttpNotifyServer
also uses a tuple the specify the source
(the address and port the notify server listens on.)