app.py

import streamlit as st
import epitran
import langcodes
from langcodes import LanguageTagError
from pathlib import Path
from operator import itemgetter
from collections import defaultdict
# TODO: reverse transliterate? 



def get_lang_description_from_mapping_name(string_to_check):  
  if "generic-Latn" == string_to_check: 
    return "Generic Latin Script"
    
  if len(string_to_check)<2:
    return None       

  substrings = string_to_check.split("-")
  substrings = substrings[:2] # first two are ISO 639-3 language, and ISO 15924 script
  string_to_check = "-".join(substrings)
  


  description = None
  lang = langcodes.get(string_to_check)
  if lang: 
    items = []
    for key, value in lang.describe().items():
      if key == "language":
        iso_code = lang.to_alpha3()
        value = f"[{value}](https://iso639-3.sil.org/code/{iso_code})"
      items.append(f"{key}: {value}")
      
  
    description = ", ".join(items)  
    if substrings[-1] == "red":
      description = description + " (reduced)"    
  return description

@st.cache
def get_valid_epitran_mappings_list():
  map_path = Path(epitran.__path__[0]) / "data" / "map"
  map_files = map_path.glob("*.*")
  valid_mappings = [map_file.stem for map_file in map_files]
  valid_mappings.append("cmn-Hans") # special case
  
  problem_mappings = ['generic-Latn',
 'tur-Latn-bab',
 'ood-Latn-sax',
 'vie-Latn-so',
 'vie-Latn-ce',
 'vie-Latn-no',
 'kaz-Cyrl-bab'] # https://github.com/dmort27/epitran/issues/98

  filtered_mappings = [mapping for mapping in valid_mappings if mapping not in problem_mappings]
  
  return filtered_mappings 

@st.cache
def get_epitran(selected_mapping):
  if selected_mapping == "cmn-Hans":
#    st.info("Chinese requires a special dictionary. Downloading now")
    epitran.download.cedict()  
  
  epi = epitran.Epitran(selected_mapping)
  return epi
  

if __name__ == "__main__":

  st.write("# Phonemize your text with [Epitran](https://github.com/dmort27/epitran)!")

  st.write("Epitran is a library and tool for transliterating orthographic text as IPA (International Phonetic Alphabet), by Mortensen, David R.  and Dalmia, Siddharth and Littell, Patrick.") 
  
  valid_epitran_mappings = get_valid_epitran_mappings_list()
  index_of_swa_latn = valid_epitran_mappings.index("swa-Latn")
  st.write(f"It supports converting many writing sytems to IPA symbols, including approximately {len(valid_epitran_mappings)} languages/scripts, listed below:")

  #st.write(valid_epitran_mappings)
  
  selected_mapping = st.selectbox("Select input language/script:", valid_epitran_mappings, index=index_of_swa_latn)

  
  description = get_lang_description_from_mapping_name(selected_mapping)
  st.write(f"Selected input language/script: {description}")
    
  st.info("attempting to instantiate epitran transliterator for your language/script")    
  epi = get_epitran(str(selected_mapping))
      
  examples = defaultdict(lambda: 'Try typing some words in the language you chose, and they will be transliterated.')
  examples['cmn-Hans'] = '太初有道，道与神同在，道就是神。'
  examples['swa-Latn'] = 'Mwanzoni Kabla ulimwengu haujaumbwa alikuwepo Neno Huyo Neno alikuwa pamoja na Mungu, na Neno alikuwa Mungu.'

  input_text = st.text_area(label="Whatever you type here will be transliterated!", value=examples[selected_mapping])
  
#  combined_code = "-".join([iso_lang_code, iso_script_code])
#  st.write(f"Combined code: {combined_code}")


  st.info(f"transliterating `{input_text}`\n\tusing {selected_mapping}...")  
  transliteration = epi.transliterate(input_text)
  
  output = {
      "original": input_text, 
      "transliteration":transliteration,  
  }
  
  st.write(output)