Latvian language incorrect

[viktors1982]

Hey,

Latvian language is incorrect. Does not recognize long marks (macron).

import easyocr
reader = easyocr.Reader(['lv'])
txt = reader.readtext('avatar.jpg')
print(txt)

Output:

[([[42, 12], [463, 12], [463, 62], [42, 62]], 'Gada ir tikai divas dienas,', 0.057680848985910416), ([[53, 57], [449, 57], [449, 101], [53, 101]], 'kuras mes neko nevaram', 0.41570523381233215), ([[39, 101], [143, 101], [143, 143], [39, 143]], 'iesakt.', 0.9080015420913696), ([[155, 101], [463, 101], [463, 143], [155, 143]], 'Viena ir vakardiena', 0.32598042488098145), ([[185, 143], [469, 143], [469, 187], [185, 187]], 'ritdiena. Tadejadi', 0.2768159508705139), ([[35, 149], [153, 149], [153, 185], [35, 185]], 'un otra', 0.5262686610221863), ([[66, 184], [434, 184], [434, 232], [66, 232]], 'šodien ir ista diena, lai', 0.17492368817329407), ([[63, 231], [442, 231], [442, 272], [63, 272]], 'miletu, ticetu, daritu un', 0.17923244833946228), ([[68, 266], [267, 266], [267, 320], [68, 320]], 'galvenokart', 0.8032127618789673), ([[295, 275], [431, 275], [431, 313], [295, 313]], 'dzivotu!', 0.7159291505813599)]

[rkcosmos]

That's because 1. my list of Latvian characters doesn't have characters with long mark. (see this file ) You can help me by send a PR with updated characters list. 2. Even with updated character list, Latin model currently doesn't have those characters. So you will have to wait for next version of model to be able to recognize that.

[viktors1982]

Thanks for your work. In the attachment I added correct Latvian characters

lv_char.txt

[rkcosmos]

Is there anything between P and R (also p and r)? They are blank space in my text editor.

[viktors1982]

there is nothing between them.

[visutida]

Updated by #112