mirror of
				https://github.com/explosion/spaCy.git
				synced 2025-11-04 09:57:26 +03:00 
			
		
		
		
	
		
			
				
	
	
		
			27 lines
		
	
	
		
			971 B
		
	
	
	
		
			Python
		
	
	
	
	
	
			
		
		
	
	
			27 lines
		
	
	
		
			971 B
		
	
	
	
		
			Python
		
	
	
	
	
	
from ..tokenizer_exceptions import BASE_EXCEPTIONS
 | 
						|
from ...symbols import ORTH, NORM
 | 
						|
from ...util import update_exc
 | 
						|
 | 
						|
 | 
						|
_exc = {}
 | 
						|
 | 
						|
 | 
						|
for exc_data in [
 | 
						|
    {ORTH: "ডঃ", NORM: "ডক্টর"},
 | 
						|
    {ORTH: "ডাঃ", NORM: "ডাক্তার"},
 | 
						|
    {ORTH: "ড.", NORM: "ডক্টর"},
 | 
						|
    {ORTH: "ডা.", NORM: "ডাক্তার"},
 | 
						|
    {ORTH: "মোঃ", NORM: "মোহাম্মদ"},
 | 
						|
    {ORTH: "মো.", NORM: "মোহাম্মদ"},
 | 
						|
    {ORTH: "সে.", NORM: "সেলসিয়াস"},
 | 
						|
    {ORTH: "কি.মি.", NORM: "কিলোমিটার"},
 | 
						|
    {ORTH: "কি.মি", NORM: "কিলোমিটার"},
 | 
						|
    {ORTH: "সে.মি.", NORM: "সেন্টিমিটার"},
 | 
						|
    {ORTH: "সে.মি", NORM: "সেন্টিমিটার"},
 | 
						|
    {ORTH: "মি.লি.", NORM: "মিলিলিটার"},
 | 
						|
]:
 | 
						|
    _exc[exc_data[ORTH]] = [exc_data]
 | 
						|
 | 
						|
 | 
						|
TOKENIZER_EXCEPTIONS = update_exc(BASE_EXCEPTIONS, _exc)
 |