Rasa error disappears after deleting database conversations Why? - rasa-nlu

I have a strange problem in Rasa. When I deleted some action name or changed some action name. At the same time, I also deleted some slots. Rasa will get the following error:
2019-10-12 13:57:58 ERROR rasa.core.trackers - Tried to set non existent slot 'mathcalc'. Make sure you added all your slots to your domain file.
2019-10-12 13:57:58 ERROR rasa.core.trackers - Tried to set non existent slot 'mathcalc'. Make sure you added all your slots to your domain file.
2019-10-12 13:57:58 ERROR rasa.core.trackers - Tried to set non existent slot 'mathcalc'. Make sure you added all your slots to your domain file.
2019-10-12 13:57:58 ERROR rasa.core.trackers - Tried to set non existent slot 'mathcalc'. Make sure you added all your slots to your domain file.
2019-10-12 13:57:58 ERROR rasa.core.trackers - Tried to set non existent slot 'mathcalc'. Make sure you added all your slots to your domain file.
2019-10-12 13:57:58 WARNING rasa.core.featurizers - Feature 'intent_calc_plus' could not be found in feature map.
2019-10-12 13:57:58 WARNING rasa.core.featurizers - Feature 'active_form_meeting_form' could not be found in feature map.
2019-10-12 13:57:58 WARNING rasa.core.featurizers - Feature 'entity_mathcalc' could not be found in feature map.
2019-10-12 13:57:58 WARNING rasa.core.featurizers - Feature 'intent_calc_plus' could not be found in feature map.
2019-10-12 13:57:58 WARNING rasa.core.featurizers - Feature 'active_form_meeting_form' could not be found in feature map.
2019-10-12 13:57:58 WARNING rasa.core.featurizers - Feature 'entity_mathcalc' could not be found in feature map.
2019-10-12 13:57:58 WARNING rasa.core.featurizers - Feature 'intent_calc_plus' could not be found in feature map.
2019-10-13 09:08:21 WARNING rasa.core.domain - Failed to use action 'action_calc_plus' in history. Please make sure all actions are listed in the domains action list. If you recently removed an action, don't worry about this warning. It should stop appearing after a while.
2019-10-13 09:08:21 WARNING rasa.core.domain - Failed to use action 'action_calc_minus' in history. Please make sure all actions are listed in the domains action list. If you recently removed an action, don't worry about this warning. It should stop appearing after a while.
2019-10-13 09:08:21 WARNING rasa.core.domain - Failed to use action 'action_calc_divide' in history. Please make sure all actions are listed in the domains action list. If you recently removed an action, don't worry about this warning. It should stop appearing after a while.
2019-10-13 09:08:21 WARNING rasa.core.domain - Failed to use action 'action_calc_multiply' in history. Please make sure all actions are listed in the domains action list. If you recently removed an action, don't worry about this warning. It should stop appearing after a while.
2019-10-13 09:08:21 WARNING rasa.core.domain - Failed to use action 'action_calc_multiply' in history. Please make sure all actions are listed in the domains action list. If you recently removed an action, don't worry about this warning. It should stop appearing after a while.
The slot 'mathcalc' and Feature 'intent_calc_plus' and action 'action_calc_plus' that I have deleted.
The above error disappears only when I delete the conversations table. What is the reason?
I can't delete the conversations table in the production environment.
Thanks a lot!

Looks like I'm a bit late at answering but here it is.
You are getting that error because when the user re-initiates the conversation with the bot, rasa first retrieves the conversation it had before with the user (called a tracker). This means that in the previous conversation, it had the slot mathcalc. But this time when it tries to populate the slot mathcalc, it can't because it doesn't exist this time. This is the reason that you're getting the error.
Here is what a tracker looks like:
{
"sender_id": "2408196042602645",
"slots": {
"mathcalc": null,
...
},
"latest_message": {...},
"events": [...]
}
To solve the problem, you should re-train your bot with the new data and it should be fine.
The reason that this problem disappears when you delete the conversations table is because the previous conversation doesn't exist and so it doesn't try to re-create the last conversation.
Hope that helps.

Related

Dialogflow CX - Form Parameter FINAL and UPDATED Event Not working

The DialogFlow CX with Page Form Parameters was working well in the test simulator and recently, my chat bots are not responding after user inputs.
When i tried to troubleshoot the issue i realized that the $page.params.status='FINAL' or the $page.params.someparameter.status='UPDATED' events are not firing properly. Due to this the routes are getting failed.
Initially i thought the issue is in my Project. Then i tried the prebuilt agent (travel-baggage claim) in another project. Even that behaves the same without any response after the user inputs are collected.
I also reported this issue to the Developer Advocates in Twitter and updating here as well to get some response from the community.
I’ve tried to replicate your use case but I was able to successfully trigger the condition routes $page.params.status = "FINAL" and $page.params.parameter-name.status = "UPDATED" as expected and transition to the defined page from my end.
See the following for reference:
$page.params.status = "FINAL"
$page.params.parameter-name.status = "UPDATED"
To troubleshoot the issue, you may check if all the conditions defined in your condition route are fulfilled. Also, you may need to check if the condition rules applied is OR or AND. If AND, make sure that all conditions are fulfilled in order to transition to the defined page or flow.
As for the prebuilt agent Travel: baggage claim, I was also able to replicate the same behavior. However, I noticed that this is a different issue since the issue is caused by the webhook being unable to provide a response and not caused by triggering the condition route. I was able to verify this by adding a static response on the condition route and by checking the logs from the simulator. See screenshots below for more information.
Static response
Log snippet from simulator
I tried creating a new flow and migrated all the pages and it works well. I suspect the flow got corrupted when i programmatically tried to update via API.

System Extension Validation Fails

im trying to submit a Network Extension activation request but i get this error.
The operation couldn’t be completed. (OSSystemExtensionErrorDomain error 9.)
I looked it up and apparently it means "validationFailed". Is there a way to debug this because i cant seem to figure out why the validation is failing. Thanks.
Finally found the solution. You need to set the NEMachServiceName property on the Info.plist of the extension to match the app group name.

Spring Boot Cloud + RabbitMQ

I'm new to Spring, Spring Boot and RabbitMQ. However I'm an okay programmer/problem solver.
Recently I started go through this book Learning Spring Boot 2.0 - Second Edition, code for this can be found here https://github.com/learning-spring-boot/learning-spring-boot-2nd-edition-code.
I think that things were moving along okay for me...a few hiccups here and there, but whatever as long as I was able to progress and learn I figured I was doing fine.
I'm just over half way through this book and then the author starts talking about Spring Cloud Streams and RabbitMQ.
I'm now working with the code found here https://github.com/learning-spring-boot/learning-spring-boot-2nd-edition-code/tree/master/6/part3.
My code doesn't produce an error, but it doesn't produce the expected results either. When I run the application the image comments do not seem to be saved or shown. I get a log message like this
o.s.a.r.c.CachingConnectionFactory : Created new connection:
SpringAMQP#2920fd66:0/SimpleConnection#59ae4da8
[delegate=amqp://guest#127.0.0.1:5672/, localPort= 57615]
when I submit the data.
However the book states that I should get getting messages like the following:
o.s.integration.channel.DirectChannel : preSent on channel input,
message: GenericMessage [payload=Comment(id=null,imageId=book.jpg)]
Looking at what I should be getting it appears as thought the problem may be related to channels. Maybe the issue is not channels, maybe it's something else.
How should I go about finding the exact cause of this of the problem?
I am not familiar with Greg's book but I just loaded that app into eclipse and it seems to work fine for me. After uploading an image it shows up in the list.
I then comment on the image and I see...
2018-01-19 00:04:41.046 DEBUG 16612 --- [ctor-http-nio-3] o.s.integration.channel.DirectChannel : preSend on channel 'output', message: GenericMessage [payload=com.greglturnquist.learningspringboot.comments.Comment#4d4f990f, headers={id=0651f9b9-e38b-1557-67dd-bc19acae8af1, timestamp=1516338281044}]
...
2018-01-19 00:07:35.570 DEBUG 17539 --- [g-spring-boot-1] o.s.integration.channel.DirectChannel : preSend on channel 'input', message: GenericMessage [payload=byte[64], headers={amqp_receivedDeliveryMode=PERSISTENT, amqp_receivedRoutingKey=learning-spring-boot-comments, amqp_receivedExchange=learning-spring-boot-comments, amqp_deliveryTag=1, amqp_consumerQueue=learning-spring-boot-comments.learning-spring-boot, amqp_redelivered=false, id=43094b53-d9b5-115e-db44-dcef73a0a9c4, amqp_consumerTag=amq.ctag-MGTGY6EuPidJy3BH7_j2zw, contentType=application/json, timestamp=1516338455570}]
But I do see
2018-01-19 00:07:54.291 ERROR 17539 --- [ctor-http-nio-3] o.s.w.s.h.ResponseStatusExceptionHandler : Response status 405 with reason "Request method 'GET' not supported"
I am not sure if you are supposed to see comments that were posted.
How should I go about finding the exact cause of this of the problem?
First step would be to enable DEBUG logging for the root logger (or all of org.springframework).
You might want to reach out to Greg on his web site.

freeswitch not accepting calls, tls invalid

i just made a fresh install for freeswitch on windows
wireshark show invites coming in on sip, but freeswitch show no activity
on startup freeswitch shows several errors, specifically
2016-08-16 16:37:08.919250 [DEBUG] sofia.c:4237 tls-verify-date [true]
2016-08-16 16:37:08.919250 [DEBUG] sofia.c:4237 tls-verify-policy
[none] 2016-08-16 16:37:08.919250 [ERR] sofia_glue.c:329 Invalid
tls-verify-policy value: none 2016-08-16 16:37:08.919250 [DEBUG]
sofia.c:4237 tls-verify-depth [2]
and
2016-08-16 16:37:15.260094 [NOTICE] switch_loadable_module.c:390
Adding File Format 'G723' 2016-08-16 16:37:15.260094 [CRIT]
switch_loadable_module.c:1520 Error Loading module C:/Program
Files/FreeSWITCH/mod/mod_png.dll
**dll open error [126l]
** 2016-08-16 16:37:15.261095 [CONSOLE] switch_loadable_module.c:1538 Successfully Loaded [mod_local_stream] 2016-08-16 16:37:15.261095
[NOTICE] switch_loadable_module.c:338 Adding API Function
'local_stream'
(actually the dll load err happens for several dlls/mods)
i did some research and found that "none" may not be a valid option, so i commented out all settings to none (since its default anyway) , except for the ipv6 file
but the error still happens, should the ipv6 file be loaded? i believe it causes issues, moreover the documentation states that ipv6 is not supported on windows. so how can i disable it?
i appreciate any help, thanks!
I've just found a workaround about the first issue:
The "tls-verify-policy" is a parameter defined in some sip profile. In my case it was a external profile in the route:
/usr/local/freeswitch/conf/sip_profiles/external.xml
And the parameter was defined as:
<param name="tls-verify-policy" value="none"/>
The problem is that the value "none" is not valid and according the documentation the possible values are in/out/all/subjects_in/subjects_out/subjects_all.
But if you don't want to verify any connection, you can leave it blank, as in:
<param name="tls-verify-policy" value=""/>
For the second issue see #NicholasBlasgen answer.
If you have no need for MOD_PNG, then best to just disable it. That should be the quickest solution. Look at autoload_configs/modules.conf.xml to be able to disable it.
The longer solution would be to identify why the linked library won't load. I suspect there was a bug in it and it's likely been fixed by now. You can also use Freeswitch's bug tracker to report a problem.

I can't enable or disable modules. I'm getting "An error occurred while saving this configuration"

I have a new fairly fresh install of Magento 1.7.0.2. I installed two custom payment method modules that are working fine. But now whenever I try to disable or enable any Payment Method or Shipping Method it gives me an error like this:
An error occurred while saving this configuration: Warning:
preg_match() expects parameter 2 to be string, array given in
/app/code/core/Mage/Adminhtml/Model/System/Config/Backend/Encrypted.php
on line 57
Do you have any idea why?
EDIT 2:
I found the problem. It was caused by a patch I applied to fix to a previous BUG. :P
EDIT:
So I disabled all custom modules and I was still getting the same error. I logged what was happening and it seems that some core modules are sending an array to be saved.
Like PayPal Express Checkout for instance. One of the config options is the Time of Day in the SFTP credentials for Settlement Report Settings. It's composed by three different drop-downs and it's sending an array to be saved!
In the Shipping Methods, UPS has a config options called Allowed Methods, and it's also sending an array!
I still don't know why it wasn't happening before and now it is. They shouldn't be encrypting everything. Any clues?
I had the same issue, its a known bug, try this around line no. 135 on code/core/mage/Adminhtml/model/config/data.php
// add this code
$backendClass=false;
if (isset($fieldConfig->backend_model)) {
$backendClass = $fieldConfig->backend_model;
} // before this
if (!$backendClass) {
$backendClass = 'core/config_data';
}
This actually is a reported bug in Magento that seems to affect encrypted config settings (passwords, API keys, etc.).
http://www.magentocommerce.com/bug-tracking/issue?issue=14217
Try disabling the first of the modules, log out of the admin panel, clear /var/cache and see if the problem still exists. If it does, do the same with the second module.
The problem should be gone. Now that you know which module causes the problems you can either decide to try another module instead or debug to see whats going wrong.
Put a breakpoint on line 57 in /app/code/core/Mage/Adminhtml/Model/System/Config/Backend/Encrypted.php to see what goes wrong. It seems that the second parameter is not a string (might be null for example due to wrong configuration or something) when its given to the preg_match function. Maybe this helps you to identify the problem.
Regards
Disabling modules through the back office isn't a good idea as it only disables the block output, if you would like to disable a module you should go to app/etc/modulesand there you should find Module_Name.xmlfile - in this file just put false in active node.

Resources