Improved Glottolog retrieval

app.py

@@ -3,7 +3,7 @@ import langcodes
 from requests_html import HTMLSession
 import urllib
 import requests
-
+session = HTMLSession() 
 # FEATURE: get wikipedia codes, e.g. from https://en.wikipedia.org/wiki/List_of_Wikipedias or https://meta.wikimedia.org/wiki/List_of_Wikipedias, some of which are nonstandard. Then output f"{code}.wikipedia.org"
 # Big TODO: collate all the results into a big dictionary? Then display that. Reduces if statements?
 # TODO: fix 'knh', it has an empty ISO section. Turns out some languages only have 639-3
@@ -51,17 +51,20 @@ def pull_obsolete_codes(iso_code):
 
 def try_retrieving_glottolog_id(langtext):
   languoid_id = ""
-  session = HTMLSession() 
   langtext_quoted = urllib.parse.quote(langtext)
   query_url=f"https://glottolog.org/glottolog?search={langtext_quoted}"
   glottolog_r= session.get(query_url)
   returned_url = glottolog_r.html.url
-
   
   if "languoid" in returned_url:
     last_section = returned_url.split("/")[-1]
     languoid_id = last_section
   return languoid_id    
+  
+def get_glottolog_json(languoid_id):
+  query_url=f"https://glottolog.org/resource/languoid/id/{languoid_id}.json"
+  glottolog_r = session.get(query_url)
+  return glottolog_r.json()
 
 def try_searching_vachan_engine(langtext):
   results_list = []
@@ -76,7 +79,7 @@ def try_searching_vachan_engine(langtext):
 def main():
   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` (or even some in other languages, try '法语' or 'français'), 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)")
+  st.write("Try typing your language below! Accepts either codes like `eng`/`en`, or full names like `English` (or even some in other languages, try '法语' or 'français'), and we will use the [langcodes](https://github.com/rspeer/langcodes) library to figure out the correct modern BCP-47 tag according to [official W3 Guidelines](https://www.w3.org/International/questions/qa-choosing-language-tags)")
   st.write(f"**Feedback:** Provide feedback at https://twitter.com/cleong110, or via slack: https://masakhane-nlp.slack.com/archives/D01DU3MHP7A")
   
   # https://huggingface.co/blog/streamlit-spaces
@@ -121,15 +124,15 @@ def main():
       st.success(f"* Natural language search found the following BCP-47 tag: {lang}")
     except LookupError as e:
       st.error("## Result: failure!")
-      st.error(f"Unable to look up language code. But all hope is not lost...")
+      st.error(f"Unable to look up BCP-47 tag. But all hope is not lost...")
       st.write(f"* You can also try https://r12a.github.io/app-subtags/")    
-      st.write(f"* Or possibly https://glottolog.org/glottolog?search={urllib.parse.quote(langtext)}")
       lang = None
   
   
 
   
-  t_variant = None 
+  t_variant = None
+  b_variant = None 
  
   
   #st.write(f"langcodes found the following tag: {type(found)}") # a Language object
@@ -184,15 +187,16 @@ def main():
       st.write(b_obsolete_codes)
 
   st.write(f"### Glottolog")
-  languoid_id = try_retrieving_glottolog_id(langtext)         
-  if languoid_id:
-    st.write(f"**Glottolog Languoid ID:** Searching for '{langtext}' on Glottolog returns the following 'languoid ID': [{languoid_id}](https://glottolog.org/resource/languoid/id/{languoid_id})")
-  # FIXME: fix this to display something if there's an ISO code to try
-  if t_variant:
-    st.write(f"https://glottolog.org/glottolog?search={t_variant} may be also of interest, with links to various resources including WALS, Wikidata, Odin, and OLAC. ")
-    if t_variant != b_variant:
-      st.write(f"If that doesn't work, try https://glottolog.org/glottolog?search={b_variant}, or put in a [custom search query](https://glottolog.org/glottolog)") 
-  st.write(f"https://glottolog.org/glottolog?search={urllib.parse.quote(langtext)} may pull up something as well.")
+  search_terms_for_glottolog = [langtext, t_variant, b_variant]
+  languoids = []
+  for search_term in search_terms_for_glottolog :
+    if search_term :
+      languoid_id = try_retrieving_glottolog_id(search_term )  
+      if languoid_id:
+        if languoid_id not in languoids:
+          st.write(f"**Glottolog Languoid ID:** Searching for '{search_term}' on Glottolog returns the following 'languoid ID': [{languoid_id}](https://glottolog.org/resource/languoid/id/{languoid_id})")
+#          get_glottolog_json(languoid_id)
+          languoids.append(languoid_id)
           
   results_from_vachan = try_searching_vachan_engine(langtext)    
   if results_from_vachan: