I notice the following behaviour yesterday on my english instance of Wit.ai.
For the same input text
"hi, i'd like to book a flight from brussels to nice, leaving next monday in the morning and back in the evening"
I am getting sometimes the intent "intro" with confidence 1, and other times i get the intent "request_price", with conficende also 1
How is that even possible?
Regards
Example of both responses
Would you be able to show what your submitted examples are for Request_Price? I would suggest submitting some really small examples in the Understanding tab for both intents, deleting confusing examples from the undesired intent, or perhaps completely deleting the undesired intent and re-adding it with very clear examples (the final suggestion requiring the most work with highest probability of success).
Related
I have a big fat sort of table with lots of elements that make it really buggy.
For context 1 row per user, each user has X projects and each project has 3 month of day display (sort of gantt)
So I built something cool it's working great, but if I scale to more users it begin to be real buggy.
I'm implementing filter to display less users but at some point it need to handle the limit that I have now without being buggy.
What I found is when I'm updating a day it re-render the whole Gantt which is really stupid.
here is a minimal reproduction link: https://stackblitz.com/edit/vitejs-vite-g6azah?file=src%2FApp.vue&terminal=dev
As you can see when I'm updating an input with the v-model:
value.value
I have the attribute with the function
:test="testRerender()"
That trigger for the whole Gantt and I believe it is the issue here.
I saw v-memo that look like what I need but I can't figure out with the doc how can I use it to match my needs (and there is almost no good article on it)
Thanks for your help you would save my life!
I tried to filled proper :key attribute, code optimization, v-memo etc...
I have been working on a simple game in Unreal Engine 4. I am trying to make it so when a player is hit by a cube they take damage. However, I am stuck on creating a condition. I have previously used:
to set up a condition where a player only takes damage if they are touched by the cube (In my cube pawn blueprints).
This doesn't work however - when trying to set-up my health bar:
This shows that I am now using entirely new variables to attempt to get a successful updating bar.
Without setting it up so when a cube hits the player, the player takes damage, the player will take damage from simply jumping at walking into other surfaces.
I have created a function that successfully updates my current health and max health so I don't need to show, or need help with the maths or updating the widget. Is there a way for me to use the branch to create an if statement that checks the contact is form a cube?
I am quite new to blueprints and have mostly developed through the use of tutorials. If you need clarity on my question or you don't understand what I am asking please leave a comment and I will try to update. I have looked long and hard for an answer, but I have found that Unreal Engine 4 hasn't got many questions that I can tailor the answer to my situation. If the answer is already in another post on this website, comment saying so and I will remove this post.
Thanks for any help you can give me :)
(This also has a itch.io page for me to quickly share to my friends so I will also credit the person who helped me there)
If I've understood your question correctly, I believe you are just asking how to check if the cube has hit the player character and not some other actor.
Instead of using an if statement as you suggested, you can just cast the Other Actor property to your first person character. If the cast is successful then it hit the character, if the cast fails, it hit something else. You can then call the damage function which you said you've already created. Below is a basic example you could use in your cube blueprint. You will also need to make sure you have a collision box surrounding your cube mesh (and your character, but I can already see that in your screenshot).
In Gherkin, you can have free-form text that describes a scenario, a feature, etc. These descriptions are not used by, say, a test runner, but are for you to describe important additional information to another human.
The documentation for Gherkin says that these cannot start a line with one of the other keywords, such as Given, When or Then. Yet, sometimes the best description I could give would be to start with one of these keywords.
I'm sort of making this up as I go here, but here is an example of what I wish I could do:
Scenario: Many notifications at the same time get combined
When we have a lot of notifications being posted at once, it causes problems
for humans. They can't make sense of that much new information all at once.
So if we are ever in a situation where we are posting lots of
notifications in a short time period, we will take the one with the highest
severity and show it with the other notifications as "child" notifications,
accessible via a link that says, "And N other issues."
Given a notification posted today at 11:03:25
And a notification posted today at 11:03:26
And a notification posted today at 11:03:26
And a notification posted today at 11:03:27
When a notification is posted at 11:03.28
Then the notification list will contain 1 notification
And that notification should contain 4 child notifications
The problem I have is that because my description starts with a When, it the tools assume that I've started my specific steps, and blows up on the next line, which doesn't start with a keyword.
I've considered:
Commenting out the first line or the entire description (that seems more consistent to me) but to me, there is a semantic difference between a comment with # and a description.
Rewording the thing to not start with a "When". For example, if it started with, "In times where we have a lot of notifications..." but that's less readable, which is the point with Gherkin-style specifications.
If it wasn't the first word in the whole description, I might be able to get away with simply wrapping my lines differently so that the "When" starts in the middle of a line instead of the beginning, but in this case, I don't have that option.
Those options just seem like workarounds that feel sub-optimal.
Is there a way to "escape" these keywords to tell the system that some usage of "When" is really still just part of the description and not a keyword? If not, is there some sort of accepted best practice or guideline for how people should handle situations like this?
You could use # in the beginning of the line (it's used for writing comments).
Ex:
# When a notification is posted...
You are misinterpretting the the language spec. You can describe a feature and use keywords at the beginning of a line. The example you posted gets interpreted as steps in a scenario since the description appears after the scenario keyword.
Just as Mr Cas said in his answer, you need comments.
Feature: Given a feature title
When I use keywords up here
Then it is allowed
Scenario: When I use keywords after the title to describe a scenario
# Then I need to use comments
I've been trying to plot tweets with certain hashtags on a picture in processing but haven't been able to find any good resources or assistance.
It's really hard to answer general "how do I do this" type questions. Stack Overflow is designed more for specific "I tried X, expected Y, but got Z instead" type questions. That being said, I'll try to answer in a general sense:
You need to break your problem down into smaller steps.
Step 1: You can draw your image using the image() function. Get a sketch working that does just this without worrying about the tweets just yet.
Step 2: You can use a library like Twitter4J to retrieve tweets with certain hashtags. Put together a completely separate sketch that just retrieves tweets and prints them out to the console. Don't worry about displaying them.
Step 3: You can use the text() function to display text on your screen. Again, get a simple application working first before you try combining it with the other steps.
When you have each of those steps working by themselves, then you can start thinking about combining them into one sketch. Focus on one small step at a time, and post an MCVE if you get stuck. Good luck!
I'm trying to create a survey containing several slider-type questions, asking participants to select a point on a line to indicate their experience of different emotions. For example:
Not at all anxious -----------------------|---------------------Extremely anxious
These questions are meant to serve as visual analogue scales (VAS), often used in my field but typically administered via paper/pen. A digital VAS would be very useful for my research, but I have 2 problems with the Qualtrics format.
First, I need to hide the vertical slider until participants selects a point on the line (in effect removing any visual "starting point" that can influence responses). Second, I can't force participants to respond to every question, so I need a way to distinguish those who skip a question/don't answer (including code to check if the slider was displayed or not displayed when the participant chooses to continue to the next question might work, so that Qualtrics stores no response for that question if the slider is still hidden).
Thanks for the help! I'm not a coder, so I would really value any and all advice!