Responses not working for me about botfront

[znat]

@melnichukme
Thank you for submitting bugs!
Would it be ok to post the other bug as a separate issue?

[znat]

Hi @melnichukme , the logs seem correct as utter_hi is predicted. What does the bit actually answer?

[melnichukme]

@znat I expect utter_hi to be replaced with my rules defined in responses. I expect the text "simple text message".

[znat]

And what do you have instead?

[melnichukme]

@znat I expect utter_hi to be replaced with my message
http://joxi.ru/DmB0wNhJze5EAP

[znat]

Oh I see. Are you running Botfront with the Botfront CLI? Can you verify that the API is running (http://localhost:8080)?

[melnichukme]

@znat Maybe I'm doing something wrong. botfront is installed on the virtual machine through docker. Yes, I use botfront cli.

1. Launch - sudo botfront up
2. By the ip address of the virtual machine, I have access to port 8080. http://joxi.ru/Dr8OD6hozelgA6

p.s. Here are the settings
http://joxi.ru/ZrJ0bXhwMEZB2j
http://joxi.ru/LmG0bEhwJVd62l
http://joxi.ru/YmE0bnhwJp6pr6
http://joxi.ru/J2bak1t0GQEjm6
http://joxi.ru/V2VgWBHdK6L0Av

.env
http://joxi.ru/D2P0lDhqk34q23
.botfront.ym
http://joxi.ru/krD0bOhKJNO4rp

I also noticed an error in the settings section (google chrome console devtool)

Error: Cannot find module './docker-compose/Settings.docker-compose'
at y (6273a5d45ad43771add2ac9731419c64b7f4c3c7.js?meteor_js_resource=true:7)
at d.resolve (6273a5d45ad43771add2ac9731419c64b7f4c3c7.js?meteor_js_resource=true:7)
at 6273a5d45ad43771add2ac9731419c64b7f4c3c7.js?meteor_js_resource=true:7
at new Promise ()
at d.prefetch (6273a5d45ad43771add2ac9731419c64b7f4c3c7.js?meteor_js_resource=true:7)
at d.c.dynamicImport (6273a5d45ad43771add2ac9731419c64b7f4c3c7.js?meteor_js_resource=true:25)
at g.componentDidMount (6273a5d45ad43771add2ac9731419c64b7f4c3c7.js?meteor_js_resource=true:149)

[znat]

Can you check in the logs? There must be an error while fetching the response from the API. botfront logs

[melnichukme]

@znat
After starting docker, I see such a log.
https://pastebin.com/bcJL572L
After sending the message 'hi' to the chat log, the following:
https://pastebin.com/fv24eKkx

The only error I noticed in the log:

botfront-rasa | 2019-08-05 23:06:07 DEBUG rasa.core.nlg.callback - Requesting NLG for utter_hi from http: // botfront-api: 8080 / project / bf / nlg.
botfront-rasa | 2019-08-05 23:06:07 ERROR rasa.core.nlg.callback - NLG web endpoint returned an invalid response: 404, Not Found, body = 'b' {"code": 404, "error": "not_found "} ''

[melnichukme]

@znat
I still haven't solved my problem .. I installed ubuntu 19, with python 3.7 but response still doesn't work. I tried to write you in scectrum, but you are not responding. How can I contact you to help identify a problem? if you want it ...

Is there any information on my problem? I found another bug. How to reproduce:

1. Open chat
2. Select French
3. Write any word

Result:
There is no response from the server

Logs:

botfront-rasa | 2019-08-06 22:43:05 ERROR asyncio - Task exception was never retrieved
botfront-rasa | future: <Task finished coro = <AsyncServer._handle_event_internal () done, defined at /usr/local/lib/python3.6/site-packages/socketio/asyncio_server.py:426> exception = AttributeError ("'NoneType' object has no attribute 'parse' ",) created at /usr/local/lib/python3.6/site-packages/engineio/asyncio_server.py:309>
botfront-rasa | source_traceback: Object created at (most recent call last):
botfront-rasa | File "/ usr / local / bin / rasa", line 10, in
botfront-rasa | sys.exit (main ())
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/rasa/main.py", line 76, in main
botfront-rasa | cmdline_arguments.func (cmdline_arguments)
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/rasa/cli/run.py", line 92, in run
botfront-rasa | rasa.run (** vars (args))
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/rasa/run.py", line 54, in run
botfront-rasa | ** kwargs
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/rasa/core/run.py", line 179, in serve_application
botfront-rasa | app.run (host = "0.0.0.0", port = port)
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/sanic/app.py", line 1096, in run
botfront-rasa | serve (** server_settings)
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/sanic/server.py", line 770, in serve
botfront-rasa | loop.run_forever ()
botfront-rasa | File "/usr/local/lib/python3.6/asyncio/coroutines.py", line 126, in send
botfront-rasa | return self.gen.send (value)
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/spf/framework.py", line 476, in _handle_request
botfront-rasa | stream_callback)
botfront-rasa | File "/usr/local/lib/python3.6/asyncio/coroutines.py", line 110, in next
botfront-rasa | return self.gen.send (None)
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/sanic/app.py", line 917, in handle_request
botfront-rasa | response = await response
botfront-rasa | File "/usr/local/lib/python3.6/asyncio/coroutines.py", line 110, in next
botfront-rasa | return self.gen.send (None)
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/engineio/asyncio_server.py", line 238, in handle_request
botfront-rasa | packets = await socket.handle_get_request (environ)
botfront-rasa | File "/usr/local/lib/python3.6/asyncio/coroutines.py", line 110, in next
botfront-rasa | return self.gen.send (None)
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/engineio/asyncio_socket.py", line 89, in handle_get_request
botfront-rasa | return await getattr (self, 'upgrade' + transport) (environ)
botfront-rasa | File "/usr/local/lib/python3.6/asyncio/coroutines.py", line 110, in next
botfront-rasa | return self.gen.send (None)
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/engineio/asyncio_socket.py", line 128, in _upgrade_websocket
botfront-rasa | return await ws (environ)
botfront-rasa | File "/usr/local/lib/python3.6/asyncio/coroutines.py", line 110, in next
botfront-rasa | return self.gen.send (None)
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/engineio/async_drivers/sanic.py", line 122, in call
botfront-rasa | await self.handler (self)
botfront-rasa | File "/usr/local/lib/python3.6/asyncio/coroutines.py", line 110, in next
botfront-rasa | return self.gen.send (None)
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/engineio/asyncio_socket.py", line 222, in _websocket_handler
botfront-rasa | await self.receive (pkt)
botfront-rasa | File "/usr/local/lib/python3.6/asyncio/coroutines.py", line 110, in next
botfront-rasa | return self.gen.send (None)
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/engineio/asyncio_socket.py", line 42, in receive
botfront-rasa | run_async = self.server.async_handlers)
botfront-rasa | File "/usr/local/lib/python3.6/asyncio/coroutines.py", line 110, in next
botfront-rasa | return self.gen.send (None)
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/engineio/asyncio_server.py", line 407, in _trigger_event
botfront-rasa | ret = await self.handlers [event] (* args)
botfront-rasa | File "/usr/local/lib/python3.6/asyncio/coroutines.py", line 110, in next
botfront-rasa | return self.gen.send (None)
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/socketio/asyncio_server.py", line 494, in _handle_eio_message
botfront-rasa | await self._handle_event (sid, pkt.namespace, pkt.id, pkt.data)
botfront-rasa | File "/ usr / local / lib / p

[znat]

Can you try this url in your browser and paste the output here?

http://localhost:8080/project/bf/responses

[znat]

Thanks @melnichukme
If you can just paste the ouput of curl http://localhost:8080/project/bf/responses that would help understanding your issue

[melnichukme]

@znat
I have 2 projects, one in Russian, the second chitchat in English and French. I checked your request and it returned the following answer:
{"responses":[{"values":[{"sequence":[{"text":"simple text message"}],"lang":"ru"}],"key":"utter_hi"}],"timestamp":1564963627849}

Very strange, but in json I only see Russian. I thought that he was somewhere defined as the default language and tried to make a response for him, but the chat still does not respond to me with messages

[melnichukme]

@znat
Hooray, I think I understand what the problem is!
When installing botfront, you are prompted to create your own project. And this project, it seems to me, is the default project and all the data will be related to it, even if you switch to another project:
http://localhost:8080/project/bf/responses always returns your project data. If you switch to chitchat and create response messages for it, they will not work, and even the search for intents will be carried out from your project and not from chitchat ...
http://joxi.ru/L21dN0sRg5p3mX

I understand that the Russian language does not work in botfront? Tell me how can I make it work?

[znat]

There is no language limitation in Botfront. However we're currently fixing bugs related to the model not being trained correctly the first time. Fix very soon, but in the meantime you can try these workarounds:

• make sure you have at least 2 intents per language
• add/remove an NLU example
• restart rasa explicitly botfront restart rasa

[melnichukme]

@znat
How do I add Russian? If you add the Russian language through the settings, rasa cannot recognize it and highlight intent ... Also, the chit chat tab in the NLU section tells me that the language is not supported.

How do I remove the NLU example? where does it lie on the file drive?

The product is interesting. But I found other errors and would like to tell you more about them, not through github

[znat]

There is no chitchat data in Russian, that does not mean that Russian is not supported, just that you'll have to build your own chitchat data (and ideally contribute it :) )
You're saying that you can't train Russian? Have you seen something about it in the Rasa forums?

[melnichukme]

@znat
On the forums I read that there are ready-made language files. But where to put them, I do not know. Where is rasa located after installation?

Is it possible to create more than 1 project in botfront? I did not find such an opportunity. For example, I need to make 5 bots, but within one server I can only create one ..

[znat]

No, it's one project at a time. Yo can create several projects by running botfront init in several folders. We're working on a simpler way to deploy projects so you don't have to deploy the whole stack for each project

[znat]

Re Russian, you should not need language files as Botfront comes with the tensorflow pipeline which is supposed to be language agnostic

[melnichukme]

@znat
Bug 1
Added 3 entries to training data, defined intent. Clicked training. In the text box "Users says ..." entered one of 3 phrases and do not get the definition of intent .... Only 1 intent was created

botfront-rasa | 2019-08-08 22:52:45 ERROR rasa.nlu.classifiers.embedding_intent_classifier - There is no trained tf.session: component is either not trained or didn't receive enough training data

When I created 2 intent it worked, maybe it’s worth blocking the button in the UI until more than 1 intention is created or is this an error on the server?

Bug 2 (russian language)
http://joxi.ru/Vm6RGQS4jBzJmZ
error after write message in input 'Users says..."

What is the reason for this is not clear. But after I created more training phrases and intents, it worked

botfront-rasa | 2019-08-08 23:01:43 DEBUG rasa.server - Traceback (most recent call last):
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/rasa/server.py", line 829, in parse
botfront-rasa | request.json.get("lang")).parse(
botfront-rasa | AttributeError: 'NoneType' object has no attribute 'parse'
botfront-rasa |
botfront-rasa | 2019-08-08 23:01:43 DEBUG rasa.server - Traceback (most recent call last):
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/rasa/server.py", line 829, in parse
botfront-rasa | request.json.get("lang")).parse(
botfront-rasa | AttributeError: 'NoneType' object has no attribute 'parse'
botfront-rasa |
botfront-rasa | During handling of the above exception, another exception occurred:
botfront-rasa |
botfront-rasa | Traceback (most recent call last):
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/rasa/server.py", line 838, in parse
botfront-rasa | "An unexpected error occurred. Error: {}".format(e),
botfront-rasa | rasa.server.ErrorResponse: (400, 'ParsingError', "An unexpected error occurred. Error: 'NoneType' object has no attribute 'parse'")
botfront-rasa |
botfront-rasa | 2019-08-08 23:01:44 DEBUG rasa.server - Traceback (most recent call last):
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/rasa/server.py", line 829, in parse
botfront-rasa | request.json.get("lang")).parse(
botfront-rasa | AttributeError: 'NoneType' object has no attribute 'parse'
botfront-rasa |
botfront-rasa | 2019-08-08 23:01:44 DEBUG rasa.server - Traceback (most recent call last):
botfront-rasa | File "/usr/local/lib/python3.6/site-packages/rasa/server.py", line 829, in parse
botfront-rasa | request.json.get("lang")).parse(
botfront-rasa | AttributeError: 'NoneType' object has no attribute 'parse'
botfront-rasa |
botfront-rasa | During handling of the above exception, another exception occurred:

Bug 3 (after start evaluation)
http://joxi.ru/Q2KBbpsLwzdZmj

---

As I understand it, the main problems are related to the UI. My personal opinion is to block the buttons, thereby making it clear to the user when you can train (when there is already enough data).
http://joxi.ru/52aZxkhEb6DjA0

I eventually created a new project and created at least 2 intents. "Users says" defined my intents. As I understand it, the main problems are related to the UI. My personal opinion is to block the buttons, thereby making it clear to the user when you can train (when there is already enough data).

I was glad to help you

[znat]

Hi @melnichukme , the latest release fixes a bunch of issues relating to training.
I tried with Russian (using Google translate) and I could train 2 intents and get the model to interpret utterances correctly. It is true that the app doesn't let you know you need at least two intents with 2 examples each.
Thanks for reporting issues, it's extremely helpful!

[melnichukme]

@znat Installed a new version ... and it does not work
http://joxi.ru/n2YVlksbe0ZJA6
After installation, I get to a white screen .. (section "stories"). Changed the configuration from localhost to my real ip adress - nothing has changed.

Also the error is not fixed:
http://joxi.ru/DmB0wNhJ46zGAP

Chat does not open .. in the developer's console errors:

Error: Cannot find module './docker-compose/Settings.docker-compose'
    at y (1231d1e8b92370f5b6e0173f6da2b88156e4b9c6.js? meteor_js_resource = true: 7)
    at d.resolve (1231d1e8b92370f5b6e0173f6da2b88156e4b9c6.js? meteor_js_resource = true: 7)
    at 1231d1e8b92370f5b6e0173f6da2b88156e4b9c6.js? meteor_js_resource = true: 7
    at new Promise ()
    at d.prefetch (1231d1e8b92370f5b6e0173f6da2b88156e4b9c6.js? meteor_js_resource = true: 7)
    at d.o.dynamicImport (1231d1e8b92370f5b6e0173f6da2b88156e4b9c6.js? meteor_js_resource = true: 25)
    at g.componentDidMount (1231d1e8b92370f5b6e0173f6da2b88156e4b9c6.js? meteor_js_resource = true: 149)

And also (there is no localhost in the settings):

Failed to load resource: net :: ERR_CONNECTION_REFUSED
http: // localhost: 8888 / __ meteor __ / dynamic-import / fetch

For two weeks, I could not use your product))

[znat]

Regarding your first screenshot:

1. Why do you need sudo
2. It seems that two versions are mixed up. The latest version doesn't need a default model but the cli still tries find it. And it's not something I can reproduce. Can you try to completely uninstall and reinstall botfront?

npm uninstall -g botfront
npm install -g botfront

Also, are you changing stuff? Can you paste the content of you botfront.yml(in .botfront)

[znat]

Evaluation is indeed broken. Just created an issue, Thanks for reporting

[znat]

Should be fixed. Feel free to re-open if the issue persists