Document your #powerbi model with #chatGPT and a #tabulareditor script and browse it in your model

Inspired by Darrens Gosbell’s excellent blog post – Automatically generating measure descriptions for Power BI and Analysis Services with ChatGPT and Tabular Editor – Random Procrastination (gosbell.com) – and my session at SQLBits 2023 where I showed how to call the chatGPT from within Power Query and how you can use the New chat (openai.com) – to describe your M code.

I thought it would be nice if I could add this as a description to my tables in Power BI model.

How hard can that be 🙂

Well, it turns out that we are not allowed to modify the table object via the Tabular #ditor and this includes adding a description to the table

So, if I created a script that added a description – it worked fine but returning to the desktop I was asked to refresh the queries and the descriptions where removed by Power BI desktop –

Well, what is supported as write operations to a model?

Calculation groups is supported – hmmm… could I use this and then actually create a calculation group with a calculation item for each of the tables in my model – that would also mean that I could create a table in my report and expose the descriptions in a report page instead of the description should be hidden away on the tooltip of a table in the field list – This could actually be quite awesome!

C# Script here we go

I know very little about how to write C# code so Darrens example and the documentation of Tabular editor was a big help and please bear in mind that I am a C# novice when reading my code

The steps needed to this is as follows (I have used Tabular Editor 3 but it should be possible to us the free version of Tabular Editor as well

Step 1 – Open your Power BI desktop file and find the external tool ribbon and click the Tabular Editor

Step 2 – Add a Calculation Group called “Documentation”

Step 3 – Choose to create a New C# Script

Step 3 – Use the following script and run it

#r "System.Net.Http"
using System.Net.Http;
using System.Text;
using Newtonsoft.Json.Linq;

// You need to signin to https://platform.openai.com/ and create an API key for your profile then paste that key 
// into the apiKey constant below
const string apiKey = "<YOUR_API_KEY>";
const string uri = "https://api.openai.com/v1/completions";
const string question = "Please describe this power query code for me:\n\n";

using (var client = new HttpClient()) {
    client.DefaultRequestHeaders.Clear();
    client.DefaultRequestHeaders.Add("Authorization", "Bearer " + apiKey);

    foreach (var t in Model.Tables)
    {
        
        foreach ( var p in t.Partitions)
        {
            // Only uncomment the following when running from the command line or the script will 
            // show a popup after each measure
            //
            //var body = new requestBody() { prompt = question + m.Expression   };
                string _type = Convert.ToString(p.SourceType);
                string _exp = Convert.ToString(p.Expression);
        if ( _type == "M" )
            {var body = 
                "{ \"prompt\": " + JsonConvert.SerializeObject( question + p.Expression ) + 
                ",\"model\": \"text-davinci-003\" " +
                ",\"temperature\": 1 " +
                ",\"max_tokens\": 256 " +
                ",\"stop\": \".\" }";

            var res = client.PostAsync(uri, new StringContent(body, Encoding.UTF8,"application/json"));
            //res.Result.EnsureSuccessStatusCode();
            var result = res.Result.Content.ReadAsStringAsync().Result;
            var obj = JObject.Parse(result);
            var desc = obj["choices"][0]["text"].ToString().Trim(); 
          
            //Reference to your calculation group that should hold the calculation Items
            var x =(Model.Tables["Documentation"] as CalculationGroupTable).CalculationItems[t.Name];
            
            //deletes the old version
            x.Delete();

            var calculationItem1 = (Model.Tables["Documentation"] as CalculationGroupTable).AddCalculationItem();
            
            //removes any quotes in the chatGPT description
            var s = desc.Replace("\"", "");

            calculationItem1.Expression = "\"" +  s + "\"";
            calculationItem1.Name = t.Name;
            
            //Info("Processing " + t.Name) ;
            
            }
        }

    }
}

This will create a calculation item for each of your Power Query table and add a Expression that contains the chatGPT description of your M code

Example of one of the tables

Step 4 – Save the changes back to your model and you will be prompted to refresh your calculation group

Step 5 – Add a table where you take the Calculation group name and a measure that I call Query description – the value of this measure will in the table be changed to the expression of the calculation item.

This So GIFfrom This GIFs

https://tenor.com/embed.js

We could use the same method to document our DAX measures and put that into a calculation group as well – thereby documenting our full model and exposing the information directly in the report and not just in the tooltips of the fields.

I will try to find time to do a blog post on this as well.

Let me know in the comments if you find this useful – would very much like to hear from you.