in src/indexing.py [0:0]
def IndexTextQuery(txt_mmap, ref_mmap, idx):
p = int(ref_mmap[idx]) # get starting byte position
i = 0
dim = 10000 # max sentence length in bytes
b = bytearray(dim)
# find EOL
while txt_mmap[p+i] != 10 and i < dim:
b[i] = txt_mmap[p+i]
i += 1
return b[0:i].decode('utf-8')