langcode-search / app.py
cdleong's picture
Update app.py
58b95fa
raw
history blame
4.45 kB
import streamlit as st
import langcodes
st.write("# Language code/tag search")
st.write("Fed up with language tag confusion? Here's your one-stop shop!")
st.write("Try typing your language below! Accepts either codes like `eng`/`en`, or full names like `English`, and we will use the [langcodes](https://github.com/rspeer/langcodes) library to figure out the correct modern BCP-47 code according to [official W3 Guidelines](https://www.w3.org/International/questions/qa-choosing-language-tags)")
# https://huggingface.co/blog/streamlit-spaces
# https://github.com/psf/requests-html
# https://docs.streamlit.io/library/api-reference/write-magic/st.write
langtext = st.text_input("Language Code/Tag Lookup using langcodes", "english")
st.write("Checking whether the tag is valid. That is, the language, script, territory, and variants (if present) are all tags that have meanings assigned by IANA.")
if langcodes.tag_is_valid(langtext):
st.write(f"...True! '{langtext}' parses meaningfully as a language tag according to IANA.")
else:
st.write(f"...False! '{langtext}' doesn't parse meaningfully as a language tag according to IANA, some of its subcomponents may be invalid or it might be a natural language description.")
try:
lang = langcodes.Language.get(langtext)
# st.write(f"{lang} is the BCP-47 tag.")
if "unknown" in lang.display_name().lower():
st.write(f"Attempting to lookup the code directly gives us '{lang.display_name()}', attempting to search for it as a natural language string.")
lang = None
except langcodes.LanguageTagError as e:
st.write(f"Could not lookup code directly, attempting to search for it as a natural language string.")
lang = None
if lang is None:
try:
found = langcodes.find(langtext)
lang = found
st.write(f"natural language search found the following BCP-47 tag: {lang}")
except LookupError as e:
st.write(f"Unable to look up language code.")
st.write(f"Try also: https://r12a.github.io/app-subtags/")
st.write(f"Try also: https://glottolog.org/glottolog?search={langtext}")
lang = None
def pull_obsolete_codes(iso_code):
from requests_html import HTMLSession
session = HTMLSession()
r= session.get(f"https://iso639-3.sil.org/code/{iso_code}")
# https://www.w3schools.com/cssref/css_selectors.asp
for found_element in r.html.find(".views-field-nothing", clean=True):
lines = found_element.text.splitlines()
# lines = text.splitlines()
obsolete_codes = {}
for line in lines:
for obsolete_code_name in ["639-1","639-2/B", "639-2/T", "639-3"]:
if obsolete_code_name in line:
code = line.split()[-1]
obsolete_codes[obsolete_code_name] = code
return obsolete_codes
#st.write(f"langcodes found the following tag: {type(found)}") # a Language object
if lang is not None:
display = lang.display_name()
st.write(f"Best-match BCP-47 tag for '{langtext}', according to the langcodes library: {lang}")
st.write(f"Breakdown of tag components: {lang.describe()}")
st.write(f"Display name for {lang}: {lang.display_name()}")
st.write(f"Autonym for {lang}: {lang.autonym()}")
b_variant = lang.to_alpha3(variant='B')
t_variant = lang.to_alpha3(variant='T')
st.write(f"ISO 639-3 'alpha3' code, 'terminology' variant (deprecated): {t_variant}")
st.write(f"ISO 639-3 'alpha3' code, 'bibliographic' variant (deprecated): {b_variant}")
st.write(f"If it exists, the ISO 639 Code Tables entry for the T variant would be at https://iso639-3.sil.org/code/{t_variant}")
st.write(f"If it exists, the ISO 639 Code Tables entry for the B variant would be at https://iso639-3.sil.org/code/{b_variant}")
# ethnologue prefers T for german (deu), and T for French
obsolete_codes = pull_obsolete_codes(t_variant)
if obsolete_codes:
st.write(f"Obsolete codes from previous ISO-639 iterations:")
st.write(obsolete_codes)
broader_tags = lang.broader_tags()
st.write(f"Broader tags for this language, if any:")
st.write(broader_tags)
st.write(f"Try also: https://r12a.github.io/app-subtags/?lookup={lang}")
st.write(f"https://glottolog.org/glottolog?search={t_variant} may be of interest, with links to Ethnologue, etc. If that doesn't work, try https://glottolog.org/glottolog?search={b_variant}")
st.write(f"## Correct, standardized, BCP-47 tag for {langtext}, according to the langcodes library: {langcodes.standardize_tag(lang)}")