PSM.AUTO_ONLY still includes OCR results

[olivierpascal]

Describe the bug
The result of recognize() with PSM_AUTO_ONLY still include text.
I guess this is not normal.

To Reproduce
worker.setParameters({ tessedit_pageseg_mode: PSM.AUTO_ONLY });
worker.recognize(image);
Return: { ..., psm: "AUTO_ONLY", oem: "DEFAULT", text: "bla bla" }

Expected behavior
The doc state: "Automatic page segmentation, but no OSD, or OCR."
So I expect { ..., text: null } but this is not the case.

Desktop (please complete the following information):

• OS *
• Browser *

[Balearica]

This mode is marked as "not implemented" in Tesseract, and this behavior is consistent with the desktop version. Therefore, outside of clarifying the documentation (or removing this option from the API to avoid confusion) adding this feature is probably outside of the scope of this project.