#! /usr/bin/python3
# Last edited on 2026-10-07 09:06:31 by stolfi

# Functions for removing punctuation from SBJ (hanzi) and SPS (EVA) text,

import sys, os, re
from sys import stdout as out, stderr as err, stdin as inp
from error_funcs import arg_error, file_line_error, prog_error
from process_funcs import bash, basic_line_loop
from chinese_funcs import read_chinese_char_set
from compute_stats_funcs import compute_and_print_stats
from math import sqrt, hypot, exp, log, pi, inf, nan, floor, ceil

def remove_punctuation(text, utype, data_error):
  # Removes the  punctuation of a file of the given {utype}.
  # The result may be of a different {utype}.
  #
  # See {remove_bencao_punctuation} for more details when {utype} is "ch"
  # of "wh".
  # 
  # See {remove_starps_punctuation} for more details when the {utype}
  # is "ec", "wp", or "wc".
  # 
  if utype == "ch" or utype == "wh":
    text = remove_bencao_punctuation(text, utype, data_error)
  elif utype == "ec" or utype == "wc" or utype == "wp":
    text = remove_starps_punctuation(text, utype, data_error)
  return text
  # ----------------------------------------------------------------------

def remove_starps_punctuation(text, utype, data_error):
  # Removes punctuation from a {text} from an SPS text (line,
  # parag or page) of the given {utype}, which must be 
  # "ec", "wc", or "wp".
  #
  # If {utype} is "wc" or "wp", deletes all commas, dashes, and periods,
  # and the alignment markers [«=»].  The text then becomes of type "ec".
  #
  # If {utype} is "ec", bombs out if the {text} contains any dashes, commas,
  # or periods; otherwise it just deletes any alignment markers [«=»].

  text = text.strip()
  if utype == "ec" or utype == "wc" or utype == "wp":

    # Remove alignment markers:
    text = re.sub(r"^[«=»]", "", text)
    text = re.sub(r"[«=»]$", "", text)

    # Replace punctuation:
    text = re.sub(r"-", ".", text)
    if utype == "ec":
      if re.search(r"[-,.]", text) is not None:
        data_error("punctuation in 'ec' text")
    elif utype == "wc" or utype == "wp":
      text = re.sub(r"[-,.]", "", text)
    else:
      assert False # Should not happen.
  else:
    arg_error(f"invalid {utype = !r}")

  return text
  # ----------------------------------------------------------------------

def remove_bencao_punctuation(text, utype, data_error):
  # Removes the punctuation from an SBJ (hanzi) text of type {utype}.
  # Returns the cleaned {text}.  
  #
  # If the {utype} is "wh", the function removes all ideographic
  # punctuation, including ideographic spaces '　'. The result is then
  # of type "ch".
  #
  # If the {utype} is "ch", the functon fails if the {text} contains any
  # ideographic punctuation, including spaces; otherwise it is a no-op.
  # 
  punct = r"[］［；：、，。［］《》　]"
  if utype == "wh":
    # Remove all punctuation:
    text = re.sub(punct, "", text)
  elif utype == "ch":
    if re,search(punct, text) is not None:
      data_error("punctuation in 'ch' text")
  else:
    arg_error(f"invalid text unit type {utype = !r}")
    
  return text
  # ----------------------------------------------------------------------

def test_stuff():

  def data_error(msg):
    err.write(f"** {msg}\n"); assert False
    # ....................................................................

  return
  # ----------------------------------------------------------------------

if len(sys.argv) > 1 and sys.argv[1] == "BSRF.TEST":
  test_stuff()