API Error on Import

[Thomas Lazarus]

Is there a reason the API is pinged when importing the library? It seems a little weird and has caused some issues for me when running tests. Thanks!

[Levi John Wolf]

The library fetches the current copy of the metadata upon import, and uses this to structure the API. it's part of the design: we don't want to mirror or maintain duplicates of the main API metadata endpoints.

I suppose, strictly speaking, it doesn't have to be fetched on import, but rather could be pushed down into the first (if any) call to explorer.available() & cached once fetched.

Happy to merge any contribution to do this!

[Thomas Lazarus]

@ljwolf Could you point me to where that initial API fetch is? I'm having trouble finding it. Thanks 👍🏽

[Ronnie Llamado]

@lazarust See line 13 in cenpy/explorer.py:

cenpy/cenpy/explorer.py

Lines 12 to 35 in f767140

  
 resp = raw_APIs = r.get("https://api.census.gov/data.json") 
 try: 
     resp.raise_for_status() 
     raw_APIs = resp.json()["dataset"] 
     APIs = { 
         entry["identifier"] 
         .split("id")[-1] 
         .lstrip("/"): { 
             key: value for key, value in diter(entry) if key != entry["identifier"] 
         } 
         for entry in raw_APIs 
     } 
 except r.HTTPError: 
     raise r.HTTPError( 
         "The main Census API Endpoint (https://api.census.gov/data.json) is not available." 
         " Try visiting https://api.census.gov/data.json in a web browser to verify connectivity." 
     ) 
 except JSONDecodeError: 
     raise JSONDecodeError( 
         "The main Census API Endpoint (https://api.census.gov/data.json) returned malformed content." 
         " Try visiting https://api.census.gov/data.json in a web browser to verify connectivity." 
     )