Hi everyone,
I want to join the latest complains about the laziness and depowered behaviour of GPT API.
In addition to that, I’ve built a tool where people can upload local files and scraped data from social like text, video and images.
We always used a function to get the document from MongoDB (JSON structure) and upload it to OpenAI https://api.openai.com/v1/files
// create file locally from scraped data
for (const data of entityScrapedDataList) {
await fsPromises.writeFile(`${data.id}`, JSON.stringify(data));
console.log(`Created file for data ${data.id}`);
// create openAi file from createdFile
const file = await openAIFilesCreate({
file: createReadStream(data.id),
});
console.log("openAI file ----->", file);
file_ids.push(file.id);
// delete temp files
await fsPromises.unlink(data.id);
}
return file_ids;
In the last days, this function is not working anymore. All the agents are replying:“The myfiles_browser tool is designed for browsing and searching within text-based files, including plain text (.txt), PDFs, and Word documents (.docx). However, it’s not intended for use with JSON files, which are typically data files formatted as JavaScript Object Notation.”
While, in the docs https://platform.openai.com/docs/assistants/tools/knowledge-retrieval you can find:" The maximum file size is 512 MB and no more than 2,000,000 tokens (computed automatically when you attach a file). Retrieval supports a variety of file formats including .pdf
, .md
, .docx
and many more. More details on the file extensions (and their corresponding MIME-types) supported can be found in the Supported files section below."
and if you go to Supported Files https://platform.openai.com/docs/assistants/tools/supported-files
you can see .json
application/json
for code interpreter and retrieval.
So, the question is why for the God’s sake is not working anymore?
Thank you