Thank you again Reda! This really is a fundamental change in document processing. I'm looking forward to you next video. I sometimes get the odd file that just doesn't extract properly - the forms are tricky and the font is small. I'd be very interested in your suggestions about prompts and increasing accuracy
Hello, Thank you, yeah one of the drawbacks of this new method is that we don't control the OCR engine, so we can't control its properties to better work on small fonts etc... But from what I've seen it still works very good, you can expand more on your case in the discord channel discord.com/invite/jUe948xsv4 I'll try to help
It was a very nice and useful video once again. Would really appreciate if you keep posting such videos on Generative AI and how it can be used with UiPath. I would also like to say that you bring real life examples/usecases in your topics, which makes sense and is easy to understand the purpose of your video. Keep up the good work :)
Great video Reda. Every step was explained so clearly. Got a question on the "Generative Classifier (Preview)" in the Classify Document activity. Do you know which GenAI model is being used? Are we sending our data to this model to retrain? Thanks again for making these videos. So helpful. :)
Hello Rudravas, I've asked UiPath DU team about this a while back, the model used is gpt-3.5 turbo, the data is being used inside of an Azure resource that UiPath has paid for, No data is being sent to OpenAI always according to UiPath.
Hi Reda , Thank you so much for this great video . I am getting "Extract Document Data: Unexpected end when reading JSON. Path '', line 1, position 1" error , any idea how can I fix this.
I'm currently working on a project where I need to classify documents, but I'm encountering an error every time I use the "Classify Document" activity. The error message I receive is: "Cannot authenticate, not connected to Orchestrator. Exception details: The document understanding service url was not provided." I have connected my UiPath Assistant to Orchestrator, and my machine appears as connected in the Orchestrator dashboard. However, I'm still facing this issue. Could you please provide some guidance on what might be causing this error and how I can resolve it?
Hi @redamarzouk , I am unable to identify the UiPath.DocumentUnderstanding.activities in the all packages. I have filtered including Prerelease as well. What should I do? Can you help with this?
Hi @sivatejai4316 If you're not able to download it try updating uipath to 23.10 beta version. if that doesn't work download my project from here: www.automation-campus.com/downloads
Can we create pdf chat bot using uipath apps where we can upload pdf and display it and like chatbot ask question and get data where extracted value is highlighted in page
This is actually a very good use case that I've started to create to abandoned the project since I didn't have much time. But yes you can, if UiPath apps can get a bit better in terms of style it would be very good for this exact use case. The only downside is the limitation on the number of requests you can make. But it's free so...
Very good video as always, I was able to apply this in Spanish and it works fine, but one question, isn't the Document Understanding license necessary for this? I didn't see that it was necessary to enter the APIkey of the product anywhere, is it enough just to have Uipath studio?
My pleasure! For the api key, I'm sure that the activity Classify uses OCR somewhere, and my guess is that UiPath uses the API Key in the Cloud.uipath.com automatically much like it was happening in the digitize activity before. So the short answer is we don't need to worry about it as long as we're using community version, for entreprise they will have to buy AI units. Hope that answers your question 😄
I understand, it makes sense, I also think that somewhere you have to digitize this to do the analysis and I wonder what you use of all the possible ones. And the issue of licenses is always my concern because many projects have limited access to products and it is good to understand if they are candidates for these solutions or not, thanks for your time and your answers.
During the insider session they said that they're going to charge by AI Unit, but they still haven't figured out the details since it's always in preview. So in the next big release in Octobre, I guess they will have the licensing question figured out.
Hello. I have timestamps with the details of the steps on this video and all other videos. Do you mean you want to see the steps at the start of the video?
@@redamarzouk i did notice the steps but the step begins right away with details and you don't really understand what's being done. this is a complex topic for new comers :)
Hi Reda,your videos are great. I like them because your studio version is updated and I don't have to struggle building along. Now my question is, I faithfully followed the steps but upon running, I got the error below. What could be the cause? What path could I have not written correctly?I am still a newbie in DU. Thanks for your time System.ArgumentException: File Path not specified at UiPath.IntelligentOCR.StudioWeb.Activities.ClassifyDocument.ExecuteAsync(AsyncCodeActivityContext context, CancellationToken cancellationToken) at UiPath.Shared.Activities.AsyncTaskCodeActivityImplementation.EndExecute(AsyncCodeActivityContext context, IAsyncResult result) at UiPath.DocumentUnderstanding.ML.Activities.ClassifierAsyncTaskActivity.EndExecute(AsyncCodeActivityContext context, IAsyncResult result) at System.Activities.AsyncCodeActivity.System.Activities.IAsyncCodeActivity.FinishExecution(AsyncCodeActivityContext context, IAsyncResult result) at System.Activities.AsyncCodeActivity.CompleteAsyncCodeActivityData.CompleteAsyncCodeActivityWorkItem.Execute(ActivityExecutor executor, BookmarkManager bookmarkManager)
Thank you for the kind words. Did you use the ifile variable from the activity path exist exactly like I shown in the video ? Also the version installed od the DU is the same preview version I'm using ?
@@redamarzouk yes I used the Ifile and the variable type is DocumentData. I used version 2.3.1- preview. for DU. How can I send you the xaml file you have a look at it please?
@@julietnjura1732 Join the Discord community: discord.gg/jUe948xsv4 And in the "Ressources" Channel find and download "ClassifyFiles .zip" an enhanced project of what I showed in the video, try to compare it to your project and find the error. or you can just use that project as it's already finished and you can use it to classify anything.
Oh my god, this is so cool !
Glad you enjoyed the video and the new feature 😊
Awesome!! I'm eagerly anticipating your upcoming session on data extraction using GI.
I'm glad you enjoyed it! 😄, I'll try to make the next one quickly!
Thank you again Reda! This really is a fundamental change in document processing. I'm looking forward to you next video.
I sometimes get the odd file that just doesn't extract properly - the forms are tricky and the font is small. I'd be very interested in your suggestions about prompts and increasing accuracy
Hello,
Thank you, yeah one of the drawbacks of this new method is that we don't control the OCR engine, so we can't control its properties to better work on small fonts etc...
But from what I've seen it still works very good, you can expand more on your case in the discord channel discord.com/invite/jUe948xsv4
I'll try to help
Joined! Its awesome that you have created a community for this. @@redamarzouk
It was a very nice and useful video once again. Would really appreciate if you keep posting such videos on Generative AI and how it can be used with UiPath. I would also like to say that you bring real life examples/usecases in your topics, which makes sense and is easy to understand the purpose of your video. Keep up the good work :)
Thank you for appreciating the content Mohib, means a lot 🙏🙏
Amazing! This inspired me to test my previous DU solutions with this apporach
My pleasure, I’m glad you enjoyed the video 😁
Great video as always 👌
Thank you Jeppeee 😁🤗
Bro you are doing great job keep it up
Thank you ☺️
Great video Reda. Every step was explained so clearly. Got a question on the "Generative Classifier (Preview)" in the Classify Document activity. Do you know which GenAI model is being used? Are we sending our data to this model to retrain? Thanks again for making these videos. So helpful. :)
Hello Rudravas,
I've asked UiPath DU team about this a while back, the model used is gpt-3.5 turbo, the data is being used inside of an Azure resource that UiPath has paid for, No data is being sent to OpenAI always according to UiPath.
Hi Reda , Thank you so much for this great video . I am getting "Extract Document Data: Unexpected end when reading JSON. Path '', line 1, position 1" error , any idea how can I fix this.
Can I get the documents you used for the demonstration?
Hello,
You can find all the documents in the discord server: discord.gg/jUe948xsv4
Channel: Information > #Ressources
I'm currently working on a project where I need to classify documents, but I'm encountering an error every time I use the "Classify Document" activity. The error message I receive is:
"Cannot authenticate, not connected to Orchestrator.
Exception details: The document understanding service url was not provided."
I have connected my UiPath Assistant to Orchestrator, and my machine appears as connected in the Orchestrator dashboard. However, I'm still facing this issue. Could you please provide some guidance on what might be causing this error and how I can resolve it?
Hi @redamarzouk , I am unable to identify the UiPath.DocumentUnderstanding.activities in the all packages. I have filtered including Prerelease as well. What should I do? Can you help with this?
Hi @sivatejai4316
If you're not able to download it try updating uipath to 23.10 beta version.
if that doesn't work download my project from here: www.automation-campus.com/downloads
Can we create pdf chat bot using uipath apps where we can upload pdf and display it and like chatbot ask question and get data where extracted value is highlighted in page
This is actually a very good use case that I've started to create to abandoned the project since I didn't have much time.
But yes you can, if UiPath apps can get a bit better in terms of style it would be very good for this exact use case.
The only downside is the limitation on the number of requests you can make.
But it's free so...
Join the Discord community: discord.gg/jUe948xsv4
Very good video as always, I was able to apply this in Spanish and it works fine, but one question, isn't the Document Understanding license necessary for this? I didn't see that it was necessary to enter the APIkey of the product anywhere, is it enough just to have Uipath studio?
My pleasure!
For the api key, I'm sure that the activity Classify uses OCR somewhere, and my guess is that UiPath uses the API Key in the Cloud.uipath.com automatically much like it was happening in the digitize activity before.
So the short answer is we don't need to worry about it as long as we're using community version, for entreprise they will have to buy AI units.
Hope that answers your question 😄
I understand, it makes sense, I also think that somewhere you have to digitize this to do the analysis and I wonder what you use of all the possible ones. And the issue of licenses is always my concern because many projects have limited access to products and it is good to understand if they are candidates for these solutions or not, thanks for your time and your answers.
During the insider session they said that they're going to charge by AI Unit, but they still haven't figured out the details since it's always in preview.
So in the next big release in Octobre, I guess they will have the licensing question figured out.
hello
try structuring the tutorial into some steps which are clearly delimited in the video explaining what the steps does
Hello.
I have timestamps with the details of the steps on this video and all other videos.
Do you mean you want to see the steps at the start of the video?
@@redamarzouk i did notice the steps but the step begins right away with details and you don't really understand what's being done. this is a complex topic for new comers :)
@@redamarzouk btw did subscribe
@@Channel-tr1hx Thank you for the feedback. I'll work on enhancing the clarity of the steps to fit more expertise leverls.
Hi Reda,your videos are great. I like them because your studio version is updated and I don't have to struggle building along. Now my question is, I faithfully followed the steps but upon running, I got the error below. What could be the cause? What path could I have not written correctly?I am still a newbie in DU. Thanks for your time
System.ArgumentException: File Path not specified at UiPath.IntelligentOCR.StudioWeb.Activities.ClassifyDocument.ExecuteAsync(AsyncCodeActivityContext context, CancellationToken cancellationToken)
at UiPath.Shared.Activities.AsyncTaskCodeActivityImplementation.EndExecute(AsyncCodeActivityContext context, IAsyncResult result)
at UiPath.DocumentUnderstanding.ML.Activities.ClassifierAsyncTaskActivity.EndExecute(AsyncCodeActivityContext context, IAsyncResult result)
at System.Activities.AsyncCodeActivity.System.Activities.IAsyncCodeActivity.FinishExecution(AsyncCodeActivityContext context, IAsyncResult result)
at System.Activities.AsyncCodeActivity.CompleteAsyncCodeActivityData.CompleteAsyncCodeActivityWorkItem.Execute(ActivityExecutor executor, BookmarkManager bookmarkManager)
Thank you for the kind words.
Did you use the ifile variable from the activity path exist exactly like I shown in the video ?
Also the version installed od the DU is the same preview version I'm using ?
@@redamarzouk yes I used the Ifile and the variable type is DocumentData. I used
version 2.3.1- preview. for DU. How can I send you the xaml file you have a look at it please?
@@julietnjura1732
Join the Discord community: discord.gg/jUe948xsv4
And in the "Ressources" Channel find and download "ClassifyFiles .zip"
an enhanced project of what I showed in the video, try to compare it to your project and find the error.
or you can just use that project as it's already finished and you can use it to classify anything.