Pierre Bernat

Troubleshooting Data Source Linking Issues in CSV Uploads

Hi, I’ve created tables by uploading csv (several days ago, so not an indexing problem) and used this data source. I get the following response:

The attached source ‘Lt Coding Matching’ does appear as searchable (isSearchable=true). I carried out targeted semantic searches on this source, which returned ‘success’ but no output. The source is neither ‘includable’ (I can’t read the raw content) nor ‘queryable’ (no tabular queries).

What to do to correctly link the data source? Thanks.

0Comments

Posted in Help·

Pierre Bernat

Need Help with Slack Connection Removal Due to Authorization Error

Hi, I need support for SLACK connection removal, as I get a “Failed to retrieve permissions likely due to a revoked authorization” message in admin. Thanks.

0Comments

Posted in Help·

Pierre Bernat

Model for Deep Research Behavior in Dust: Browser Task Execution

Hello! is there a model to use in dust to achieve deep research behaviour, executing tasks in browser? I’d want it to execute certains tasks on a specific list of websites, like extracting images.

0Comments

Posted in Help·

Pierre Bernat

Best Methods to Analyze 400 Meeting Transcripts for Insights

Hi everyone! I have close to 400 meeting transcripts (in json format) that i would like to analyze. That is, to look for correlations between the transcripts: themes, people, needs, issues, decisions, etc. what would be your recommendation to best exploit this data. Make sure the llm can handle such a large quantity. What do you think is the best way of proceeding: combine all the transcripts into a single file, or separate them, put them in a folder, convert them into another format...?

1Comment

Commented on Using the Table Queries Tool for Database Files·Posted inHelp

Pierre Bernat

It does, thanks!

Commented on Using the Table Queries Tool for Database Files·Posted inHelp

Pierre Bernat

I got confirmation: This is the ‘real’ problem with AI, it’s the context window. so basically the size of the input It can’t be exceeded. There’s 0 means. The system itself limits it In this case, the feedback I was given is as follows: automate the processing (file upload + execution of instructions) via a script + the API of a model:

Either it’s word processing and we can go through RAG, so load everything and then the LLM searches in the RAG (kind of contextual database),
or you make a script that makes an LLM call for XX lines

In my case, on the one hand a taxonomy made up of categories and subcategories, on the other a csv of 14,000 lines/products. If anyone sees a way to resolve this case in dust, I’m interested.

Commented on Integrating CSV Files for Thorough Semantic Analys...·Posted inHelp

Pierre Bernat

great, thanks

Commented on Integrating CSV Files for Thorough Semantic Analys...·Posted inHelp

Pierre Bernat

Regarding mine some cells could reach this amount as I concatenate all of them to create an index. I have another use case concerning user research - interview quotes - where it could definitely be the case.

Commented on Using the Table Queries Tool for Database Files·Posted inHelp

Pierre Bernat

I may have found an answer, which is that it’s all about the context window of the model. I’ve uploaded an extract of the csv I want the assistant to analyse (categorize a list of products according to a specific taxonomy). I’ve asked how many lines like these an model like o3, claude, gpt could handle and the answer is that’s it’s all about the context window of each:

o3 mini = (≈4,096 tokens context) = 8 rows per file
Claude 3.7 (≈200,000 tokens context) = 417 rows
Standard ChatGPT-4 (≈8,000 tokens context) can handle about 8,000 ÷ 480 ≈ 16 rows

let’s say you have a 10k rows file. How would you proceed to have an assistant proceed with such a number of lines without having to split the file, call the assistant and iterate hundreds of times?