#! /usr/bin/python3
# Last edited on 2026-10-12 10:48:47 by stolfi

# Functions for correcting errors and expanding abbreviations 
# in SBJ (hanz) or 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
import bencao_starps_convert_funcs as convf
from math import sqrt, hypot, exp, log, pi, inf, nan, floor, ceil

def normalize_text(text, utype, del_qo, exp_irm, data_error):
  # Modifies a {text} (typically after {cleanup_text}) from an SBJ or
  # SPS transcription file according to {utype}, by replacing certain
  # substrings that are assumed to be scribal errors, abbreviations, or
  # non-significant variants.
  #
  # See {normalize_bencao_text} for the meaning of {utype} is "ch" or "wh".
  # 
  # See {normalize_starps_text} for the meaning of {utype}s "ec", "wp",
  # and "wc". The parameters {del_qo} and {exp_irm} are used only for
  # these {utype}s.
  # 
  if utype == "ch" or utype == "wh":
    text = normalize_bencao_text(text, utype, data_error)
  elif utype == "ec" or utype == "wc" or utype == "wp":
    text = normalize_starps_text(text, utype, del_qo, exp_irm, data_error)
  return text
  # ----------------------------------------------------------------------

def normalize_starps_text(text, utype, del_qo, exp_irm, data_error):
  # Normalizes the raw {text} from an SPS transcription file according
  # to {utype} for use in Note 077 analysis. Returns the normalized text
  #
  # For all {utype}s, the normalization entails fixing assumed scribal
  # mistakes and assumed non-significant handwriting variation.
  #
  # Specifically, if {del_qo} is true, deletes all "qo".
  # If {exp_irm} is true, expands all "ir", "is", and "m".
  # 
  if utype == "ec" or utype == "wc" or utype == "wp":
    # Reduce presumed handwriting variants:
    text = re.sub(r"g", "m", text)
    text = re.sub(r"u", "n", text)

    # Correct presumed scribal and transcription errors:
    text = re.sub(r"ch[?]dy", "chedy", text)
    text = re.sub(r"sh[?]dy", "shedy", text)
    text = re.sub(r"q[?e]", "qo", text)
    text = re.sub(r"hh", "he", text)
    text = re.sub(r"ih", "ch", text)
    text = re.sub(r"i([ktpf])h", r"c\1h", text)

    if exp_irm:
      # Expand presumed scribal abbreviations
      text = re.sub(r"i([-,.]*)r", r"i\1in", text)
      text = re.sub(r"i([-,.]*)s", r"i\1in", text)
      text = re.sub(r"m",  "iin", text)
    
    if del_qo:
      # Remove the "qo" glyphs.
      text = re.sub(r"qo", "", text)

    # Remove double punct that may have been created by normalization:
    if utype == "wc" or utype == "wp":
      text = re.sub(r"[,.]*[.][,.]*", ".", text)
  else:
    arg_error(f"invalid {utype = !r}")
    
  return text
  # ----------------------------------------------------------------------

def normalize_bencao_text(text, utype, data_error):
  # Normalizes the raw {text} from an SBJ transcription file according to {utype}
  # for use in Note 077 analysis. Returns the normalized {text}.
  #
  # If {utype} is "ch", the input raw {text} must have no punctuation,
  # just the hanzi letters.
  #
  # If the {utype} is "wh", it may have
  # punctuation, including ideographic brackets '' around keywords 味,
  # 主, 久服, etc.
  # 
  # For both {utype}s, normalizing entails replacing some keys that 
  # the are incorrect in the CTP version.
  
  convf.check_bencao_conversion(text, utype, data_error)
  
  if utype == "ch" or utype == "wh":
    # Fix CTP errors:
    text = re.sub(r"主治", "主", text) # Seems that SBJ never used 主治.
  else:
    assert False, f"invalid text 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] == "BSNF.TEST":
  test_stuff()