Script Reference: if( GetSelectedCount(Title)>=1, endpoints.ScriptAggrStr('{"RequestType":"endpoint", "endpoint":{"connectionname":"OpenAI1", "column":"choices.text"}}', 'What are people saying: ' & aggr(concat(Title,'. '), model) ), 'please select ... ') This video covers some important tips and tricks and need to know advice when using the OpenAI Connector with a chart expression when embedding generative insights into your Qlik Cloud Analytics apps. Video 2: ua-cam.com/video/XCaaRenozb8/v-deo.html Playlist: ua-cam.com/play/PLW1uf5CQ_gSqF5bcmbBrk1q7Q4-h899V1.html OpenAI API Reference: platform.openai.com/docs/api-reference/authentication #OpenAI Model Reference: platform.openai.com/docs/models/gpt-4 #chatgpt
9:00 here you can use this script and change your connection name and columns for prompt: if( GetSelectedCount(Title)>=1, endpoints.ScriptAggrStr('{"RequestType":"endpoint", "endpoint":{"connectionname":"OpenAI1", "column":"choices.text"}}', 'What are people saying: ' & aggr(concat(Title,'. '), model) ), 'please select ... ')
Thank you - I attempted to put this in the description and it would not allow me - let me see if I can add this to the main comment as well! Appreciate you - Regards, Mike
I wonder how I would use it. Since often with these AI you pay for X amount of tokens... based on your example I would rather loop through all my console products to get the answers from the AI and save this to a QVD. Like that, I would not spend multiple times tokens for the exact same result.
Yep - I believe I mentioned that towards the end of Part 2 - ua-cam.com/video/XCaaRenozb8/v-deo.html - using the data load approach and adding the data to the model is the most cost effective approach.
Perfect example of the value of Qlik's ability to store information for reuse in QVD's. Assuming you also have an INCREMENTAL approach to your data you could load the bulk initially, then only call OpenAI for response on the updated/inserted rows to your data set rather than reloading them all in bulk as well.
Hi Christian it is an Extension - github.com/murraygm/qse-mgoimagegrid - I am getting images from public S3 folder- and the object is configured in that manner - please note that the Content Security Policy found in the Management console - (cloud tenants) - is also set to allow a number of directives from the source image folder url - Thanks for your question - Mike T
Thank you - it is released, it just may not be turned on for you. The Qlik Management Console has a feature settings section where this can be switched on - and we will soon have permissions that control who can access Application Automation. You may need to check with your BI admin.
We couldn't find it first, either. It actually needs to be switched on here: Qlik Sense SaaS Management Console / Settings / Feature Control / Machine learning endpoints: set it to enabled
I am getting an error 'Operation was rejected because the system is not in a state reuired for the operation's execution' error details: 'grpc::StatusCode::FAILED_PRECONDITON:'Error returned by the end-point: TooManyRequests'. Any tips on how to resolve this?
Hi there this is because you most likely are using the free account which is limited - you need to have a paid relationship with OpenAI and purchase an appropriate amount of tokens / credits.
Script Reference:
if( GetSelectedCount(Title)>=1,
endpoints.ScriptAggrStr('{"RequestType":"endpoint", "endpoint":{"connectionname":"OpenAI1", "column":"choices.text"}}', 'What are people saying: ' & aggr(concat(Title,'. '), model) ),
'please select ... ')
This video covers some important tips and tricks and need to know advice when using the OpenAI Connector with a chart expression when embedding generative insights into your Qlik Cloud Analytics apps.
Video 2:
ua-cam.com/video/XCaaRenozb8/v-deo.html
Playlist:
ua-cam.com/play/PLW1uf5CQ_gSqF5bcmbBrk1q7Q4-h899V1.html
OpenAI API Reference:
platform.openai.com/docs/api-reference/authentication
#OpenAI Model Reference:
platform.openai.com/docs/models/gpt-4
#chatgpt
REALLY appreciate the fact you emphasized the column choices.text as without specifying what column you want from the return you only get the ID.
Dude as always you did a great job. Love the irony of RETRO gaming with LEADING EDGE technology.
9:00 here you can use this script and change your connection name and columns for prompt: if( GetSelectedCount(Title)>=1,
endpoints.ScriptAggrStr('{"RequestType":"endpoint", "endpoint":{"connectionname":"OpenAI1", "column":"choices.text"}}', 'What are people saying: ' & aggr(concat(Title,'. '), model) ),
'please select ... ')
Thank you - I attempted to put this in the description and it would not allow me - let me see if I can add this to the main comment as well! Appreciate you - Regards, Mike
I wonder how I would use it. Since often with these AI you pay for X amount of tokens... based on your example I would rather loop through all my console products to get the answers from the AI and save this to a QVD. Like that, I would not spend multiple times tokens for the exact same result.
Yep - I believe I mentioned that towards the end of Part 2 - ua-cam.com/video/XCaaRenozb8/v-deo.html - using the data load approach and adding the data to the model is the most cost effective approach.
Perfect example of the value of Qlik's ability to store information for reuse in QVD's. Assuming you also have an INCREMENTAL approach to your data you could load the bulk initially, then only call OpenAI for response on the updated/inserted rows to your data set rather than reloading them all in bulk as well.
What's the name of the visual that shows the photo and the price of the consoles?
Hi Christian it is an Extension - github.com/murraygm/qse-mgoimagegrid - I am getting images from public S3 folder- and the object is configured in that manner - please note that the Content Security Policy found in the Management console - (cloud tenants) - is also set to allow a number of directives from the source image folder url - Thanks for your question - Mike T
This is good feature but it is not visible in Qlik SaaS. When is it planning to release for users
Thank you - it is released, it just may not be turned on for you. The Qlik Management Console has a feature settings section where this can be switched on - and we will soon have permissions that control who can access Application Automation. You may need to check with your BI admin.
We couldn't find it first, either. It actually needs to be switched on here: Qlik Sense SaaS Management Console / Settings / Feature Control / Machine learning endpoints: set it to enabled
I am getting an error 'Operation was rejected because the system is not in a state reuired for the operation's execution' error details: 'grpc::StatusCode::FAILED_PRECONDITON:'Error returned by the end-point: TooManyRequests'. Any tips on how to resolve this?
Hi there this is because you most likely are using the free account which is limited - you need to have a paid relationship with OpenAI and purchase an appropriate amount of tokens / credits.