Hello !
I’m currently trying to create an output based on files that I’ve uploaded in OpenAI modules.
Then, I just create an assistant modules where I map the file ID into the related field.
But, I have seen different topic about that but none of them really answer to the question.
By doing this, giving a file to the assistant module. Does he take in count the file that we give to him.
Because after several try, it looks absolutely useless. The output is out context of the document we give.
If that the case, I’m wondering why there’s a file ID field in the assistant.
If anyone has any tips to make it more relevant and get the context of the document in the input.
It would be great !
Thanks for your help !
1 Like
Ok, Maybe i’ve got my answer.
Is it because the files does not have the time to be processed before launching the next module ?
If yes, should I use a sleep module to be sure the file has been processed and then, do my prompt ?
Hello 
,
Using the “Sleep” module to create a short delay between the ChatGPT’s actions definitely might help.
Your issue, however, might be caused by the file content itself or by the selected tool.
I would recommend trying to use only one of the tools (Knowledge Retrieval or Code Interpreter).
That might produce better results.
Cheerio 
3 Likes
Hello !
Thanks for your answer.
Indeed, it works, i want also to add that, at this time, i wanted to send a csv file which was not readable for the API assistant.
Now I’ve use a JSON file which is way more efficient to analyse for the assistant.
If there is someone in my case, create a json by parsing your csv and you’ll have better output.
Have a great day all !
3 Likes
