fossasia / susi_api_wrapper Goto Github PK

Parameters regarding adding a new device on Server have changed and are still left in the code of this repo

Current Behaviour
Right now, the API endpoint has been changed in the server which allowed the Smart Speaker to be shown under the devices tab of chat.susi.ai

Expected Behaviour
To update the API call and make it functional again

Would you like to work on it?
Yes

All response decoders are currently in a single file main.py . Their number are expected to grow as well. It will be best to refactor them a to separate file for cleaner code.

Currently, Susi Python module make an API request to Susi Server when we need to get an answer for a query.
But when we connect Susi Hardware to Susi Webclient ( Issue #26 (susi_hardware) ) , we get JSON from Susi Web client directly and no call to server is made by Hardware device running Susi, thus we need to get answers from JSON directly in that case and this Module must expose a method for that.

Currently, login methods are there but the token is not included in chat message request and login information (access token) is not persisted. This needs improvement.

This issue is dependent on this issue fossasia/susi_linux#309

The need is the be able to parse resume/restart/pause actions.

Installation failing due to presence of fuzzywuzzy and pafy in the requirements file

While talking to SUSI, if we get an error of server down , we should have the ability to change the server in between

In line 19,
print('Server''s Message:') should be written as print("Server's Message')

To add Tizonia along with the Linux package to auto download

I would like to work on this issue

Currently while other map functions have zoom parameter as default without the user not to be bothered about the zoom value, here we have to supply the zoom parameter.

Sample code for authentication is needed for sample usage purposes.

I am working on this

Zoom for OpenStreetMap is by default set to 13 under location of susi server system skill files.
So we may set it to default 13 value unless users specifies the zoom value specifically.

Actual Behaviour
Currently, Smart Speaker is not shown in the web client

Expected Behaviour
To show it in the web client(chat.susi.ai)

Would you like to work on it?
Yes

Please add travis.

Now , Since we are able to deploy the local server on the pi itself . I plan to add a function in app.py to choose between local and online server.

I would like to work on it

Working on it.

What is this repository about and what can people get here?

Please add all relevant project info to the Readme.md.

Description
The youtube parser is running in an endless loop and we have to manually break to go to next conversation.

Expected Behaviour
To stop after the video is over

Would you like to work on this issue?
Yes

It would be nice to have a versioned release which an initial PyPI upload and distribution packaging could start with.

Currently, we are using a Custom Decoder based SUSI Json Response parsing. While this works on many systems, it fails due on Ubuntu 16.04 due to simplejson / json library conflict since the module depends on native json library while simplejson is supplied in Ubuntu while import json.

Here is the log:

Traceback (most recent call last):
  File "main.py", line 114, in <module>
    start_speech_recognition()
  File "main.py", line 95, in start_speech_recognition
    ask_susi(value)
  File "main.py", line 62, in ask_susi
    reply = susi.ask(input_query)
  File "/home/betterclever/Github/susi_hardware/app/susi_python/main.py", line 46, in ask
    response = query(query_string)
  File "/home/betterclever/Github/susi_hardware/app/susi_python/main.py", line 22, in query
    return api_response.json(cls=SusiResponseDecoder)
  File "/usr/local/lib/python3.5/dist-packages/requests/models.py", line 894, in json
    return complexjson.loads(self.text, **kwargs)
  File "/usr/lib/python3/dist-packages/simplejson/__init__.py", line 533, in loads
    return cls(encoding=encoding, **kw).decode(s)
TypeError: __init__() got an unexpected keyword argument 'encoding'

This conflict can be resolved by first parsing the response to a json directly and getting the information from it to make the Response Objects.

Add repr methods for subclasses for proper string representation of objects.

Working on it.

Country code and country name are also required to get country wise usage analytics.
Refer to this PR : fossasia/susi_server#811

The type of device ie, "Smart Speaker" is required to get device wise usage analytics.
Refer to this PR :fossasia/susi_server#825

Steps to reproduce it

NA

LogCat for the issue

https://api.susi.ai/susi/chat.json?q=hello&language=en&access_token=PGbzC5zYQBQUuCrYrarqIuRe4lvjfn&latitude=30.1667&longitude=77.3

Screenshots of the issue

NA

Would you like to work on the issue?

Yes

Is it possible for Susi to consider the timezone of the person talking to it? It seems to show only Central European Summer Time (CEST) when asked: "What time is it?".

I'm using sample.py for the chat.

The issue is also reflected here. Susi's response on this web page seems to be very slow too.

Actual Behaviour
Right now SUSI first plays the video and then gives the answer.

Expected Behaviour
To fix the order

Yes , I would like to work on this issue

UTF-8 encoding must be used explicitly for parsing output of all requests made to SUSI Server. Currently, Python 3.4 (max available on Raspberry Pi) uses ASCII Encoding by default which causes issues while parsing some responses.

RSS Action Type is functional now. Server is sending RSS ActionType instead of Websearch.
Library must be modified accordingly.

Please convert this project so that it can be installed by python3 setup.py install, and hence by standard tool like pip.

Count Data which specifies the number of items to be displayed / spoken is not being added to parsed SUSI reply.
This is needed in fossasia/susi_linux#68

Actual Behaviour

Right now SUSI cannot change volume according to the user

Expected Behaviour

To add a VolumeAction class

Would you like to work on the issue?

Yes

Description of the bug:

Response of random text is not taken care of. Response JSON of random text returns an empty array of 'answers' key, as below :

This results in the following error :

It is expected that the app does not stop abruptly and returns a sorry response, like 'Could not get you' or any appropriate message.

I would like to work on this issue.

Add option to add custom Susi API endpoint like local deployments while configuring module.

I want to work on JS API Wrapper just as the Python API Wrapper.

I will start working on it just as I have a go from you guys

@mariobehling @stealthanthrax

Map Support can be added in the module itself to provide ability to create a map image from the URL provided in Susi Response and provide it as a byte array.

Currently, response containing table is ignored. A correct way to represent those responses must be implemented.

Current Behavior
Stop action only works for SUSI_API_Wrapper repo

Expected Behaviour
It should work for SUSI_Linux repo too

Would you like to work on it
Yes

Needed for fossasia/susi_linux#7

timezoneOffset and location params are not being sent to the SUSI Server. Support for it is needed.

Susi Server has added ability to get previous messages by a user across sessions. Support for using it in the Python Module should be added.