# Last edited on 2025-11-03 23:30:48 by stolfi
# Note 085 Generating two texts with same language diff spelling

  The goal of this note is to generate texts that could
  serve as controls for investigations into the cause of
  the "language A / language B" split.
  
  The idea is to use Machado de Assis's Dom Casmurro in Portuguese with
  three different spellings, and its Spanish translation.
  
SETUP

  ln -s ../.. work
  ln -s work/projects/langbank bank
  ln -s work/projects/dom-casmurro casm
  ln -s work/error_funcs.py
  
EXTRACTING THE FULL TEXTS

  Create files "out/{lang}/{buk}/full.txt" where {lang} is "port" or "span"
  and {buk} is "cso" (1899 spelling), "csc" (1999 spelling), "csp" (2099 phonetic spelling),
  "cas" (modern Spanish spelling).
  
    create_full_files.sh  
  
CREATING THE PARALLEL SAMPLES

    create_parallel_samples.sh

>>>===REDO===<<<

EXTRACTING THE DISJOINT SAMPLES

  Files
   
    file    ln spelling chapters    source
    ------- -- -------- ----------- --------------------------------------------------
    DC1.txt PT 1899     3,13,23,..73 ~/projects/dom-casmurro/1899/orig.txt
    DC2.txt PT 1999     4,14,24,..74 ~/projects/dom-casmurro/1999/orig.txt
    DC3.txt PT phonetic 5,15,25,..75 ~/projects/dom-casmurro/2099/orig.txt
    DC4.txt ES modern   6,16,26,..76 https://livros01.livrosgratis.com.br/al000069.pdf
    
  Lines starting with "#" are comments and should be ignored.

  Chapter numbers and titles are lines marked "# \chapt".

  Text in italics (sometimes in a foreign language) are marked "\emph{...}".

  Ellipses are "...", and the period of abbreviation is ".~" without
  a following space. Otherwise "." is always a sentence period.
  
  Lines that start with "---" are utterances in dialogs.