I looked at a few libraries. The LGWebOSRemote repository by klattimer is definitely a good library, but it has a few problems:
- Meant to be used with Python 2.x or Python 3.X
- Assumes all the users of the library would like to save the credentials to ~/.lgtv.json.
- Assumes only a single command will be fired and waited on at any given time (ctrl+F for
self.__waiting_callback
) - Mouse/Keyboard not supported.
This SDK is a tiny attempt at overcoming some of the above problems.
At the moment, I haven't been able to do any kind of extensive testing. No unit test cases too!
Current status: Works for quite a few people! :) Also, accepting PRs for bug fixes or if you feel some of the controls could be added or enhanced.
$ pip install pywebostv
If you see errors such as opensslv.h file not found, it most likely indicates that you don't have a working installation of OpenSSL. In such a case, we would recommend that you install it using the following steps (assuming you've brew as your package manager):
$ brew install openssl
$ env LDFLAGS="-L$(brew --prefix openssl)/lib" CFLAGS="-I$(brew --prefix openssl)/include" pip3 install cryptography
$ pip3 install pywebostv
from pywebostv.discovery import * # Because I'm lazy, don't do this.
from pywebostv.connection import *
from pywebostv.controls import *
# 1. For the first run, pass in an empty dictionary object. Empty store leads to an Authentication prompt on TV.
# 2. Go through the registration process. `store` gets populated in the process.
# 3. Persist the `store` state to disk.
# 4. For later runs, read your storage and restore the value of `store`.
if your_custom_storage_is_empty():
store = {}
else:
store = load_from_your_custom_storage()
# Scans the current network to discover TV. Avoid [0] in real code. If you already know the IP,
# you could skip the slow scan and # instead simply say:
# client = WebOSClient("<IP Address of TV>")
client = WebOSClient.discover()[0]
client.connect()
for status in client.register(store):
if status == WebOSClient.PROMPTED:
print("Please accept the connect on the TV!")
elif status == WebOSClient.REGISTERED:
print("Registration successful!")
# Keep the 'store' object because it contains now the access token
# and use it next time you want to register on the TV.
print(store) # {'client_key': 'ACCESS_TOKEN_FROM_TV'}
persist_to_your_custom_storage(store)
NOTE: If you're seeing repeated prompts on the TV to re-authenticate, there's a good chance you're not using the store
correctly. Read the FAQs section for more.
The client
instance represents the main channel of communication with the TV. All *Control
instances (MediaControl
, ApplicationControl
etc) share the same underlying connection. All
available APIs are grouped into separate classes (for cleanliness) like MediaControl
,
SystemControl
etc.
Most *Control
classes behave in a very similar way and are super extensible. This is because most
of the heavy lifting is done in the base class -- incorporating a new API that isn't currently
supported by this library should be very easy. Read the extension section for more on this.
Things to note:
- Most APIs support
block=
argument. IfTrue
the call blocks for the response to arrive. IfFalse
, it is a good idea to provide acallback=
argument. If you don't care about the response at all, simply call the API withblock=False
. - Some APIs support subscribing for changes. Provide a callback and you will be notified when the
event happens. It is an error to subscribe more than once on the same underlying connection. To
subscribe, the function you'd call is
control.subscribe_api_name()
assuming the regular API is calledapi_name
. To unsubscribe, just call:control.unsubscribe_api_name()
.
The general pattern is:
control = SomeControl(client)
# Blocking call
api_response = control.some_api()
# Blocking call, with parameters (the table below lists API & arguments)
api_response = control.some_other_api(arg1, arg2)
# Blocking call can throw as error:
try:
control.good_api(bad_argument1)
except ...:
print("Something went wrong.")
# non-blocking call with callback
def my_function(status_of_call, payload):
if status_of_call:
# Successful response from TV.
# payload is a dict or an object (see API details)
print(payload) # Successful response from TV
else:
# payload is the error string.
print("Error message: ", payload)
control.async_api(arg1, arg2, callback=my_function)
# Subscription (if the API supports it, that is).
control.subscribe_api(my_function).
# Unsubscribe
control.unsubscribe_api() # After this point, you can resubscribe.
Please note that all the examples below use the blocking calls. Their return values and structure are documented in the comments. They throw python exceptions when unsuccessful. To make non-blocking calls, refer to the section above.
media = MediaControl(client)
media.volume_up() # Increase the volume by 1 unit. Doesn't return anything
media.volume_down() # Decrease the volume by 1 unit. Doesn't return anything
media.get_volume() # Get volume status. Returns something like:
# {'scenario': 'mastervolume_tv_speaker', 'volume': 9, 'muted': False}
media.set_volume(<int>) # The argument is an integer from 1 to 100. Doesn't return anything.
media.mute(status) # status=True mutes the TV. status=Fale unmutes it.
media.play()
media.pause()
media.stop()
media.rewind()
media.fast_forward()
cur_media_output_source = media.get_audio_output() # Returns the currently used audio output source as AudioOutputSource instance.
audio_outputs = media.list_audio_output_sources() # Returns a list of AudioOutputSource instances.
media.set_audio_output(audio_outputs[0]) # .set_audio_output(..) accepts an AudioOutputSource instance.
get_volume
supports subscription. To subscribe to volume changes, say something like:
def on_volume_change(status, payload):
if status:
print(payload)
else:
print("Something went wrong.")
media.subscribe_get_volume(on_volume_change) # on_volume_change(..) will now be called when the
# volume/mute status etc changes.
system = SystemControl(client)
system.notify("This is a notification message!", # Show a notification message on the TV.
icon_bytes=data, # optional: the icon to be displayed,
# e.g.: requests.get(url).content
icon_ext="png") # optional: specify icon type if icon is specified above
system.power_off() # Turns off the TV. There is no way to turn it
# back on programmically unless you use
# something like Wake-on-LAN.
system.info() # Returns a dict with keys such as product_name,
# model_name, # major_ver, minor_ver etc.
system.screen_off() # Energy Saving: Turns off the screen.
system.screen_on() # Energy Saving: Turns the screen back on.
app = ApplicationControl(client)
apps = app.list_apps() # Returns a list of `Application` instances.
# Let's launch YouTube!
yt = [x for x in apps if "youtube" in x["title"].lower()][0]
# Search for YouTube & launch it (Of course, don't
# be this lazy. Check for errors). Also, Try
# searching similarly for "amazon", "netflix" etc.
launch_info = app.launch(yt) # Launches YouTube and shows the main page.
launch_info = app.launch(yt, content_id="dQw4w9WgXcQ")
# Or you could even launch a video directly!
app.close(launch_info) # Close what we just launched.
# Let's get the icon of the foreground app.
app_id = app.get_current() # Returns the application ID (string) of the
# foreground app.
foreground_app = [x for x in apps if app_id == x["id"]][0]
# Application app["id"] == app.data["id"].
icon_url = foreground_app["icon"] # This returns an HTTP URL hosted by the TV.
.get_current()
supports subscription. To subscribe, call app.subscribe_get_current(callback)
in
the same way as .subscribe_get_volume(..)
above.
inp = InputControl(client)
inp.type("This sends keyboard input!") # This sends keystrokes, but needs the keyboard to
# be displayed on the screen.
inp.enter() # Return key.
inp.delete(10) # Backspace 10 chars
The above APIs behave much like the other APIs above. The ones below are a little different. WebOS
requires that we open a different connection and uses a different message structure. You must call
inp.connect_input()
to create this connection and inp.disconnect_input()
to close it. All the
APIs below should be called between connect and disconnect.
inp.connect_input()
inp.move(10, 10) # Moves mouse
inp.click() # Click where the mouse pointer is. It sometimes also acts as the center "OK"
# button on the remote.
inp.ok()
inp.up()
inp.down()
inp.left()
inp.right()
inp.home()
inp.back()
inp.dash() # The right side menu that appears with Live button
inp.info()
inp.num_1() # Number keys...
inp.num_2()
inp.num_3()
inp.num_4()
inp.num_5()
inp.num_6()
inp.num_7()
inp.num_8()
inp.num_9()
inp.num_0()
inp.asterisk() # Literally just an "*"
inp.cc() # Closed captioning
inp.exit()
inp.red() # Colored buttons
inp.green()
inp.yellow()
inp.blue()
inp.menu() # the menu for adjusting settings for the television
inp.mute() # The remaining commands are also available in either MediaControl or TvControl
inp.volume_up()
inp.volume_down()
inp.channel_up()
inp.channel_down()
inp.play()
inp.pause()
inp.stop()
inp.fastforward()
inp.rewind()
inp.disconnect_input()
tv_control = TvControl(client)
tv_control.channel_down()
tv_control.channel_up()
tv_control.channel_list()
tv_control.get_current_channel()
tv_control.get_current_program() # Returns the current channel and the EPG data
tv_control.set_channel_with_id(channelId) # channelId can be found in channel_list(), get_current_channel() or get_current_program()
source_control = SourceControl(client)
sources = source_control.list_sources() # Returns a list of InputSource instances.
source_control.set_source(sources[0]) # .set_source(..) accepts an InputSource instance.
# To get the current current source being used, please use the API that retrieves the foreground
# app.
- How do I turn on the TV?
This library helps to establish connection with the TV and control various aspects exposed over the WebSocket API (once it discovers the IP address). If the TV is off, there's no active connection and obviously, this library can't do much. However, the TV still supports WakeOnLAN (WOL). This involves getting the MAC address of the TV and sending a magic packet to it. We would highly encourage you to perhaps use the discovered IP address, translate that to the MAC address (there are tons of libraries that do this for you, or look for it in your TV settings or get it from your router) and then send the magic WOL packet to the TV (again, tons of libraries that do this). The WOL must be enabled on the TV. As far as including the feature in this library goes -- not depending on other libraries, such as WOL, prevents client projects from having to pull in this transitive dependency which might not be always needed. Further, waking up a sleeping TV requires us sending a magic packet outside of the WebSocket layer that this library tries to limit itself to.
- Why am I getting repeated prompts to turn on the TV?
There's a good chance you're using the store
incorrectly. The very first time, you run with an unregistered TV, the store
you pass in would be empty. Once the registration is successful, remember to persist the store
object to database, config file or any other storage. The next time you your program connects to the TV, you should restore the state of store
from your storage. If you don't do this, and pass in an empty object for the second time, you're bound to get the prompt to re-authenticate. In other words, if store
is empty, you get the prompt. If you successfully restore the value from the last registration, you won't.
A tiny side note: What client.register(..)
expects is an object with __getitem__(..)
and __setitem__(..)
. So, you could potentially pass an object that persists the key+value pair on __setitem__(..)
and reads values from your storage in __getitem__(..)
.
- klattimer for his library! Since WebOS team decided against providing any sort of documentation, his repository was extremely useful for an initial implementation
- As far as input controls are concerned, they are based on the Java package written by Connect-SDK folks!
- All individual contributors to this repository.