0.1.9

[tomatyss]

Release Notes: Prompt Mixer v0.1.9

New Features

1. Cell Value Checking: Users can now click on individual cells within a dataset to view and verify the content of data points. A new tab will open displaying the selected cell's value.

2. Delete All Prompt Chains and Collections: Users can now delete all prompt chains and collections with a single action. A confirmation dialog will be displayed, warning users that this action is irreversible.

3. Model Search in Model Selector: Users can now search for models in the model selector, allowing them to quickly find the model they need to work with. The model list will automatically update to show only matching models as the user types in the search query.

4. Clear All Outputs: Users can now clear all outputs from the Prompt Mixer interface with a single click, helping them maintain a clean working environment.

5. Search Connectors: Users can now search for specific connectors within the Prompt Mixer app. The connector list will be updated to show only the matching connectors, with the results paginated and loaded through infinite scrolling.

Enhancements

• Improved performance and stability of the Prompt Mixer application.
• Enhanced user interface for a more intuitive and streamlined experience.

---

This discussion was created from the release 0.1.9.

[Spring3]

@tomatyss could you please explain the intended use of the datasets? When I saw this feature, I thought that it was the data that would be sent as the prompt context, but it looks like the PromptMixer runs the prompt for each dataset row

So if I have let's say dataset containing 1000 rows, clicking "Run" on the prompt like the following

[[Test Dataset]]

With the given csv return all popular items

it would make N requests where N is the size of the dataset.

Am I understanding the use of the datasets incorrectly? Could you perhaps give a correct example?

Thanks in advance!

[tomatyss]

Thank you for your question! Here's the answer addressing the key points:

1. Understanding the intended use of datasets: You are correct in your understanding. The PromptMixer runs the prompt for each row of the dataset, which means that if you have a dataset with 1000 rows, clicking "Run" on the prompt will make 1000 requests, one for each row of the dataset.

2. Correct example of using datasets: Suppose you have a dataset with three columns: "Name", "Topic", and "Context". Your prompt could look like this:

   Write a response for [[Dataset Name.Name]] about [[Dataset Name.Topic]]. Here is the context: [[Dataset Name.Context]]
   

   In this case, the PromptMixer will run the prompt for each row of the dataset, substituting the values from the respective columns. This allows you to test your prompt with different names, topics, and contexts in a single click.

3. Using a CSV file as a single context: If you want to use a CSV file as a single context for your prompt, you can consider using a tool like Llamaindex Connector, as you mentioned. This would allow you to pass the CSV file as a link, and the prompt would use the entire file as the context, rather than iterating through each row. https://github.com/PromptMixerDev/prompt-mixer-llamaindex-example-connector

5. Specific case example: If you have a specific use case, you can provide more details (email us to [email protected]), and we can help you with a custom connector or approach to handle your specific requirements.

In summary, the datasets feature in PromptMixer is intended to allow you to test your prompts with multiple variations of input data in a single run. If you need to use a CSV file as a single context, consider using a tool like Llamaindex Connector. Please feel free to provide more details about your specific use case, and we'll be happy to assist you further.