We are using moder activities suppose in modern activities if we can put strict,fuzzy,image selector if i select these three option which one take at run time?or how can it will run?
There is no hierarchy as such…which ever method is able to identify faster is used
Hope this helps
cheers
I’m surprised this is what the documentation says. The way it was explained to me on a deep dive conference call with a member of the UiPath product team, it actually tries to use strict first and if it doesn’t find it then it bounces down to fuzzy, if that doesn’t find it then it bounces down to image, then computer vision.
There are supposed to be some changes coming (I think in 2023.10) to make it more efficient as far as managing the timeout etc.
@loginerror I wonder if we could get some official clarification from UiPath on this topic.
Let me forward your message to the UIAutomation team, I am curious myself now
My current assumption is that strict would be in most cases the fastest anyway, with fuzzy being slightly slower and so on, which would mean that both explanation are kind of correct.
Hi,
When you are using Modern activities and you select the options for strict, fuzzy, and image selector, the behavior at runtime will depend on how you have configured these options.
-
Strict Selector: When you select the “Strict” option, UiPath will try to match the selector exactly as defined. If the selector does not match precisely as specified in the selector string, the activity will fail. It’s useful when you have a very specific selector and you want to ensure that the element you’re interacting with matches that selector precisely.
-
Fuzzy Selector: Selecting the “Fuzzy” option makes UiPath more tolerant in matching the selector. It will try to find an element that closely matches the selector but allows for some flexibility. This can be helpful when elements have dynamic attributes or when you expect some minor changes in the selector over time. Fuzzy matching makes the selector more resilient.
-
Image Selector: If you select the “Image” option, UiPath will use an image-based approach to identify the element instead of using a textual selector. It captures an image of the element and tries to match it with the actual screen content during runtime. This can be useful when you cannot reliably use a textual selector, but it’s generally less efficient and may be slower compared to using traditional selectors.
At runtime, UiPath will follow the selected option’s behavior for that specific activity. If you select multiple options (e.g., both Strict and Fuzzy), the activity will typically follow the behavior of the option that you selected last in the properties panel.
It’s important to choose the appropriate selector options based on the reliability of the selectors and the specific requirements of your automation process. Generally, it’s recommended to use Strict selectors when the element’s attributes are stable and reliable, and to use Fuzzy or Image selectors when dealing with more dynamic or hard-to-capture elements.
HI
At runtime, the selection of a strategy depends on the selector you define in your automation workflow. When you configure a selector, you can choose to use “strict,” “fuzzy,” or “image” selector strategy based on the needs of your automation task.
I think the order the user selects the properties should not affect the behavior.
Hi everyone,
@postwick is the winner here in the guessing game. Sorry for all this confusion, I’ll work directly with the documentation team to update the erroneous docs page. Paul is 90% right about the 23.4 behavior: we first try to find the target using the strict selector, then we also enable the fuzzy selector (so we now have both strict & fuzzy searching for a target); if they still both fail, we enable both CV & Image at the same time.
The 23.4 approach has a number of issues:
- it can result in an Image target even if Image sometimes can hallucinate
- CV is a powerful fallback, but it hides issues with selectors - there’s no feedback something’s wrong with them, which is not ideal; ideally, we recommend Unified Target to be fully loaded
This is why we’ve implemented a number of significant changes in our 23.10 GA package:
- there is a new clear hierarchical targeting system, based on each targeting method’s relevancy
- there is a new setting called “Wait for primary targeting method until timeout” (enabled by default for new projects in Studio >=23.10)
- selectors (both strict and fuzzy) are the primary targeting methods
- secondary is CV and Image is tertiary (so it’s de-prioritized vs. CV)
- the output of an inferior targeting method in this hierarchy will only be used in case the upper targeting methods are failing
- Image is disabled by default in Project Settings > UI Automation Modern, since it’s the weakest targeting method available (uses direct pixel values) + we have CV enabled by default, which is a far superior targeting method in terms of accuracy
- we now output warnings in case CV or Image are able to target an element, but the selector are not + we output some recommendations for the selectors possible new values (see attached pdf: Unified Target.pdf [191.7 KB])
Please watch this demo video - it should be more clear vs. a wall of text
And let us know what you think. How do you find these changes?