Calling search on a Query that returns no results causes an AttributeError exception to be raised

[Owen O'Connor]

Bug Description

When calling the search method on a Query object that would return no results, an error similar to the following is raised

/home/o-linux/anaconda3/envs/proteins/lib/python3.7/site-packages/pypdb/util/http_requests.py:61: UserWarning: Too many failures on requests. Exiting...
  warnings.warn("Too many failures on requests. Exiting...")
Traceback (most recent call last):
  File "pypdb/mwe.py", line 4, in <module>
    found_pdbs = Query('McNoman, Norman', query_type='AdvancedAuthorQuery').search()
  File "/home/o-linux/anaconda3/envs/proteins/lib/python3.7/site-packages/pypdb/pypdb.py", line 253, in search
    if response.status_code == 200:
AttributeError: 'NoneType' object has no attribute 'status_code'

System info

• Operating system: Linux Debian, kernel 5.9.0-5-amd64
• Anaconda version: 4.9.2
• Python version: 3.7.9
• pypdb version: 2.0, from conda-forge

Duplication

This issue can be duplicated by using this python script

from pypdb import *
found_pdbs = Query('McNoman, Norman', query_type='AdvancedAuthorQuery').search()

[William Gilpin]

Hi, thanks for this report. it looks like this is only working for certain names. For example, this works:

found_pdbs = Query('Perutz, M.F.', query_type='AdvancedAuthorQuery').search()
print(found_pdbs)

but this fails:

found_pdbs = Query('Perutz', query_type='AdvancedAuthorQuery').search()
print(found_pdbs)

So that means that the database lookup is super sensitive to the format of the author name, which doesn't seem ideal. For now, I would suggest doing a keyword search (which I agree is not ideal). As far as I can tell, there are no hits for that specific query, but let me know if there's a query that returns results on the website, but not from keyword search.

I should mention that @lacoperon has developed a much more full featured API that will eventually replace the old one. I think that the way to do this search with the new API (using the latest GitHub version) looks like this

from pypdb.clients.search.search_client import perform_search
from pypdb.clients.search.search_client import SearchService, ReturnType
from pypdb.clients.search.operators import text_operators

search_service = SearchService.TEXT
search_operator = text_operators.ContainsPhraseOperator(value="McNoman",
                                            attribute="audit_author.name")
return_type = ReturnType.ASSEMBLY

results = perform_search(search_service, search_operator, return_type)

print(results[:5])

This still fails for the McNoman query, but it does work for Perutz. Eventually we will convert all the old functions to use the new API under the hood.

[Owen O'Connor]

Yeah, it seems that for almost all of the basic Query type searches will cause the issue if there are no hits in the database, for example using Query('KFC'). I was poking around in the code a bit and I think I found the issue, so I can get a pr put together if you want.

[Owen O'Connor]

PR is in