Better error handling NCBI down (#377)

* better error handling NCBI down

* fix test

* return 503 if NCBI fails

* fix test

Author: manulera

src/opencloning/ncbi_requests.py

@@ -9,7 +9,10 @@
 
 async def async_get(url, headers, params=None) -> Response:
     async with get_http_client() as client:
-        return await client.get(url, headers=headers, params=params, timeout=20.0)
+        resp = await client.get(url, headers=headers, params=params, timeout=20.0)
+        if resp.status_code == 500:
+            raise HTTPException(503, 'NCBI is down, try again later')
+        return resp
 
 
 # TODO: this does not return old assembly accessions, see https://github.com/ncbi/datasets/issues/380#issuecomment-2231142816
@@ -120,9 +123,9 @@ async def get_genbank_sequence(sequence_accession, start=None, end=None, strand=
     elif resp.status_code == 400:
         raise HTTPException(404, 'wrong sequence accession')
     elif resp.status_code == 503:
-        raise HTTPException(503, 'NCBI returned an error')
+        raise HTTPException(503, 'NCBI returned an internal server error')
     else:
-        raise HTTPException(500, 'NCBI returned an unexpected error')
+        raise HTTPException(503, 'NCBI returned an unexpected error')
 
 
 def validate_coordinates_pre_request(start, end, strand):

tests/test_endpoints_external_import.py

@@ -936,6 +936,18 @@ def test_exceptions(self):
         response = client.post('/genome_coordinates', json=viral_source.model_dump())
         self.assertStatusCode(response.status_code, 400)
 
+    def test_ncbi_down(self):
+        correct_source = GenomeCoordinatesSource.model_validate(
+            request_examples.genome_region_examples['full']['value']
+        )
+        with respx.mock:
+            respx.get(
+                'https://api.ncbi.nlm.nih.gov/datasets/v2alpha/genome/accession/GCF_000002945.2/annotation_report'
+            ).mock(return_value=httpx.Response(500, text='NCBI is down'))
+            response = client.post('/genome_coordinates', json=correct_source.model_dump())
+            self.assertEqual(response.status_code, 503)
+            self.assertIn('NCBI is down', response.json()['detail'])
+
 
 class SEVASourceTest(unittest.TestCase):
     def test_seva_url(self):

tests/test_ncbi_requests.py

@@ -17,12 +17,18 @@ async def test_get_genbank_sequence_subset(self):
         with pytest.raises(HTTPException) as e:
             await ncbi_requests.get_genbank_sequence('blah', 1, 10, 1)
         assert e.value.status_code == 503
-        assert e.value.detail == 'NCBI returned an error'
+        assert e.value.detail == 'NCBI returned an internal server error'
 
         respx.get('https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi').respond(500, text='')
         with pytest.raises(HTTPException) as e:
             await ncbi_requests.get_genbank_sequence('blah', 1, 10, 1)
-        assert e.value.status_code == 500
+        assert e.value.status_code == 503
+        assert e.value.detail == 'NCBI is down, try again later'
+
+        respx.get('https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi').respond(504, text='')
+        with pytest.raises(HTTPException) as e:
+            await ncbi_requests.get_genbank_sequence('blah', 1, 10, 1)
+        assert e.value.status_code == 503
         assert e.value.detail == 'NCBI returned an unexpected error'
 
     async def test_get_annotations_from_query(self):