I have not yet gone to intent prediction level sir. I am trying to bring up Rasa and get a clean. setup.
Jitesh
am I missing any setup over here ?. probably, there must be some ENV variable to be set for Rasa to read from the trained model place rather than from regular temp space.
You wonāt get to know whether itās working fine or not unless you have tested it sir 
I am using windows system so I didnāt do any environment path variable for the temp space.
Got it. I will check and get back to you. I am new to Rasa. will try to see . but if you have any findings, then let me know
Sure sir 
Jitesh
I have built a separate virtual image under Vmware fusion. built a ubuntu and gone through this fully. too many things I have to fix. finally for this.but still our issue still perists. logged a bug with Rasa, see
Heloooooo again @JiteshGaikwad!
Iām trying to use your modifications, I was able to run them, even debug the socketChannel.py you made. The customData arrive there with no problem but isnāt sent to the āaction_session_startā event on the action server (as metadata).
File: socketChannel.py - Line: await on_new_message(message)
Iām pretty new to RASA and Python (1 week working which each), Iām trying to find why the customData isnāt sent to āaction_session_startā (as metadata), but my chances arenāt very good based on my current level of expertise at RASA and Python.
Can you help me?
I even had to change your method extract_metadata_from_tracker because of the events dictionary is empty and added some logs and the return is always empty:
tracker.current_state() {āsender_idā: ā79630f5a353f499fbec162302a30fb12ā, āslotsā: {}, ālatest_messageā: {āintentā: {}, āentitiesā: [ ], ātextā: None, āmessage_idā: None, āmetadataā: {}}, ālatest_event_timeā: None, āpausedā: False, āeventsā: , 'latest_input_channel ': None, āactive_formā: {}, ālatest_action_nameā: None}
tracker.current_state()[āeventsā]
2020-04-08 10:40:21.050844 MetaData:
File: actions.py
def extract_metadata_from_tracker(tracker: Tracker):
events = tracker.current_state()['events']
print('tracker.current_state()', tracker.current_state())
print('tracker.current_state()["events"]', tracker.current_state()['events'])
user_events = []
for e in events:
print(e)
if e['event'] == 'user':
user_events.append(e)
if len(user_events) > 0:
return user_events[-1]['metadata']
return ''
class ActionSessionStart(Action):
def name(self) -> Text:
return 'action_session_start'
async def run(self,
dispatcher: CollectingDispatcher,
tracker: Tracker,
domain: Dict[Text, Any]) -> List[Dict]:
metadata = extract_metadata_from_tracker(tracker)
print(datetime.now(), 'MetaData: ', metadata)
UPDATE:
I found that the new session always send a blank tracker, so now Iām looking for other ways to pass the customData to āaction_session_startā
I have clarification how to run the rasa server on rasa.ai.com (provided static IP and domain name) instead of localhost:5005, my environment is windows and offline environment (there is no access to internet)
I was able to pass the custom data to the action server, I had to modify your socketChannel.py (Iāll try to update the Github) and the processor.py from RASA.
My doubt here is how to make the changes o processor permanent? Just commit and push to RASA repo?
hey @lluchini, sorry I was busy with something else so didnāt got time to reply you 
Glad that you were able to solve it 
you can make a pull request, is that what you are trying to say?
No prob bro, no need to apologize for, everyone does what they can! 
I hope someday to be good at RASA/Python to help as much you do.
Yes, I never used Github (I use git), I just need to create a pull request for yours/RASA repo with the change? like on Bitbucket from Atlassian?
And I would like to know if you can just check what I did, and be sure it wonāt break anything else on RASA, can you?
Thanks and Ya sure definitely you will 
I too recently started learning about the PR, no problem you can raise the issue in the GitHub repo and tell me the changes there, I will integrate those changes, does that works for you?
1 Like
I donāt know if is this what you meant, but I forked RASA on my Github and pushed a branch with the adjusts, the link is below.
Hey @lluchini, sorry but thatās too complicated for meš
Your socketChannel thatās too complicated for meā¦ 
No problem, Iāll submit the pull request, letās see what happenā¦
Just everybody knows, it already had an issue opened for my problem Issues #5574.
I linked the open issue with my pull requestā¦
Hi,
Iām replying here because my issue seems similar.
I have worked before with the early rasa_nlu and I have one computer (Ubuntu 18.04) with working Rasa 1.3.1 installation. Now I wanted to install new Rasa to my new server, create a new bot and connect it to Webchat. This works fine in that old computer, Webchat also.
I have the latest Rasa version in Ubuntu 18.04 (Python 3.6) server. First I tried to install it to venv but it kept failing. Maybe I got it wrong somehow. But it is now installed in the server. āRasa initā worked fine.
When I run ārasa run āenable-apiā ācors ā*ā ādebugā, Rasa server startsā. The same error āFailed to load modelā as was discussed earlier. However the model gets extracted later.
In my web server Webchat tries to GET http://localhost:5005/socket.io/ā¦ but there is ERR_CONNECTION_REFUSED.
If I curl localhost:5005, I get āHello from RASAā.
Credentials.yml (socketio) should be configured fine, as I have this working in the other computer.
Any suggestions what to try next?
Hello @Kejender,
can you start rasa with debug option post the logs from rasa here so we can check with more details whatās happening?
Ok here it is:
2020-04-21 07:45:03,802 [DEBUG] Available web server routes:
/conversations/<conversation_id>/messages POST add_message
/conversations/<conversation_id>/tracker/events POST append_events
/webhooks/rasa GET custom_webhook_RasaChatInput.health
/webhooks/rasa/webhook POST custom_webhook_RasaChatInput.receive
/webhooks/rest GET custom_webhook_RestInput.health
/webhooks/rest/webhook POST custom_webhook_RestInput.receive
/model/test/intents POST evaluate_intents
/model/test/stories POST evaluate_stories
/conversations/<conversation_id>/execute POST execute_action
/domain GET get_domain
/socket.io POST handle_request
/ GET hello
/model PUT load_model
/model/parse POST parse
/conversations/<conversation_id>/predict POST predict
/conversations/<conversation_id>/tracker/events PUT replace_events
/conversations/<conversation_id>/story GET retrieve_story
/conversations/<conversation_id>/tracker GET retrieve_tracker
/webhooks/socketio GET socketio_webhook.health
/status GET status
/model/predict POST tracker_predict
/model/train POST train
/conversations/<conversation_id>/trigger_intent POST trigger_intent
/model DELETE unload_model
/version GET version
2020-04-21 07:45:03,804 [INFO ] Starting Rasa server on http://localhost:5005
2020-04-21 07:45:03,804 [DEBUG] Using the default number of Sanic workers (1).
2020-04-21 07:45:03,806 [INFO ] Enabling coroutine debugging. Loop id 94055192.
2020-04-21 07:45:04,154 [DEBUG] Extracted model to '/tmp/tmparlo_dm9'.
2020-04-21 07:45:04,486 [DEBUG] Loading the model ...
2020-04-21 07:45:05,578 [DEBUG] Finished loading the model.
2020-04-21 07:45:05,579 [DEBUG] Building tensorflow prediction graph...
2020-04-21 07:45:10,674 [DEBUG] Finished building tensorflow prediction graph.
2020-04-21 07:45:10,680 [DEBUG] Failed to load model for 'ResponseSelector'. Maybe you did not provide enough training data and no model was trained or the path '/tmp/tmparlo_dm9/nlu' doesn't exist?
2020-04-21 07:45:10,691 [DEBUG] Connected to InMemoryTrackerStore.
2020-04-21 07:45:10,691 [DEBUG] Connected to lock store 'InMemoryLockStore'.
2020-04-21 07:45:11,070 [DEBUG] Extracted model to '/tmp/tmp8_f862mu'.
2020-04-21 07:45:11,074 [DEBUG] Using yaml library: /usr/local/lib/python3.6/dist-packages/ruamel/yaml/__init__.py
2020-04-21 07:45:11,098 [DEBUG] Loading the model ...
2020-04-21 07:45:11,364 [DEBUG] Finished loading the model.
2020-04-21 07:45:11,367 [DEBUG] Building tensorflow prediction graph...
2020-04-21 07:45:12,278 [DEBUG] Finished building tensorflow prediction graph.
2020-04-21 07:45:12,283 [DEBUG] Instantiated NLG to 'TemplatedNaturalLanguageGenerator'.