Quote:
Originally Posted by PoP
@vdp If it can help, I updated the source, the best I could for my device.
|
Thanks! I made some tweaks and compiled a K3 binary. I could only test on your screenshots, so I am not completely sure it works flawlessly. You can install the files from the attachment as described in OP. If you have usbnetwork installed, you can SSH to device and run something like
Code:
lhack /mnt/us/documents '*.pdf' 0.5
while selecting different titles to see if the result makes sense. I deliberately choose a relatively lower similarity coefficient, because the English model is not a perfect match for some of your files (in French as far as I can tell). For example "Je l'aimais - Anna Gavalda" is recognized as "Ie Paimais - Anna Gavalda". As explained in the OP, higher coefficients lead to a slightly faster search, but sometimes produce false negatives.
Please let me know if it works for you.
Edit: The K3 binary is now attached to the OP.