Fix: Prez SPARQL endpoint's handling of url encoding/decoding of queries in GET requests

[edmondchuc]

For some reason, the previous implementation won't work with query parameters passed to the httpx.URL object as the result becomes a double URL encoded value. This happens when we send a SPARQL query using Postman but not via the Yasgui client.

Postman:

Testing with https://meyerweb.com/eric/tools/dencoder/, it requires the value to be decoded twice.

b'query=PREFIX%2520rdf:%2520%253Chttp:%2F%2Fwww.w3.org%2F1999%2F02%2F22-rdf-syntax-ns%2523%253E%250APREFIX%2520skos:%2520%253Chttp:%2F%2Fwww.w3.org%2F2004%2F02%2Fskos%2Fcore%2523%253E%250ASELECT%2520*%2520WHERE%2520%7B?s%2520?p%2520?obj%2520.FILTER%2520(?p%2520IN%2520(skos:prefLabel)%2520)FILTER%2520regex(str(?s),%2520%2522id%2FLexicon%2FNamedRockUnit%2522)FILTER%2520(regex(lcase(str(?obj)),%2520%2522slate%2522))%7D'

Using FastAPI/Starlette to decode and then performing percent encoding using urllib.parse.quote_plus:

b'query=PREFIX+rdf%3A+%3Chttp%3A%2F%2Fwww.w3.org%2F1999%2F02%2F22-rdf-syntax-ns%23%3E%0APREFIX+skos%3A+%3Chttp%3A%2F%2Fwww.w3.org%2F2004%2F02%2Fskos%2Fcore%23%3E%0ASELECT+%2A+WHERE+%7B%3Fs+%3Fp+%3Fobj+.FILTER+%28%3Fp+IN+%28skos%3AprefLabel%29+%29FILTER+regex%28str%28%3Fs%29%2C+%22id%2FLexicon%2FNamedRockUnit%22%29FILTER+%28regex%28lcase%28str%28%3Fobj%29%29%2C+%22slate%22%29%29%7D'

This PR fixes this issue by getting FastAPI/Starlette to always decode the incoming query parameter and encoding it ourselves using urllib.parse.quote_plus. By doing this, it doesn't look like httpx.URL will double encode the value.

I've also refactored the code slightly to keep all concerns with incoming HTTP in the route handler.

[recalcitrantsupplant]

@lalewis1 could you please test this locally, you can use postman (I use a similar service called Hoppscotch) Could you do a SPARQL query that includes something like:

{
  VALUES ?EscapedSearchTerm {"\"Illumina HiSeq 4000\"" "Illumina HiSeq 4000"
  }

Could you also add/create a test_endpoints_sparql.py file with queries in different query= formats