Now tracks failed profilings

.devcontainer/devcontainer.json

@@ -12,7 +12,7 @@
 	// "forwardPorts": [],
 
 	// Use 'postCreateCommand' to run commands after the container is created.
-	"postCreateCommand": "pip3 install --user -r requirements.txt",
+	"postCreateCommand": "pip3 install --user -r requirements.txt && pipx install . -e",
 	"customizations": {
 		"vscode": {
 			"extensions": [

src/automlst/cli/exactmatch.py

@@ -41,7 +41,10 @@ async def run(args):
         gen_strings = read_multiple_fastas(args.fastas)
         async with await bigsdb_index.build_profiler_from_seqdefdb(args.seqdefdb, args.schema) as mlst_profiler:
             mlst_profiles = mlst_profiler.profile_multiple_strings(gen_strings)
-            await write_mlst_profiles_as_csv(mlst_profiles, args.out)
+            failed = await write_mlst_profiles_as_csv(mlst_profiles, args.out)
+            if len(failed) > 0:
+                print(f"A total of {len(failed)} IDs failed:\n{"\n".join(failed)}")
+            print(f"Completed fetching MLSTs for {len(args.fastas)} sequences.")
 
 def run_asynchronously(args):
     asyncio.run(run(args))

src/automlst/engine/local/csv.py

@@ -15,11 +15,15 @@ def dict_loci_alleles_variants_from_loci(alleles_map: Mapping[str, Sequence[Alle
     return result_dict
 
 
-async def write_mlst_profiles_as_csv(mlst_profiles_iterable: AsyncIterable[tuple[str, MLSTProfile]], handle: Union[str, bytes, PathLike[str], PathLike[bytes]]):
+async def write_mlst_profiles_as_csv(mlst_profiles_iterable: AsyncIterable[tuple[str, Union[MLSTProfile, None]]], handle: Union[str, bytes, PathLike[str], PathLike[bytes]]) -> Sequence[str]:
+    failed = list()
     with open(handle, "w", newline='') as filehandle:
         header = None
         writer: Union[csv.DictWriter, None] = None
         async for name, mlst_profile in mlst_profiles_iterable:
+            if mlst_profile is None:
+                failed.append(name)
+                continue
             if writer is None:
                 header = ["id", "st", "clonal-complex", *mlst_profile.alleles.keys()]
                 writer = csv.DictWriter(filehandle, fieldnames=header)
@@ -30,4 +34,5 @@ async def write_mlst_profiles_as_csv(mlst_profiles_iterable: AsyncIterable[tuple
                 "id": name,
                 **dict_loci_alleles_variants_from_loci(mlst_profile.alleles)
             }
-            writer.writerow(rowdict=row_dictionary)
+            writer.writerow(rowdict=row_dictionary)
+    return failed

src/automlst/engine/remote/databases/bigsdb.py

@@ -46,17 +46,14 @@ class BIGSdbMLSTProfiler(AbstractAsyncContextManager):
             "designations": allele_request_dict
         }
         async with self._http_client.post(uri_path, json=request_json) as response:
-            response_json = await response.json()
+            response_json: dict = await response.json()
             if "exact_matches" not in response_json:
                 raise NoBIGSdbExactMatchesException(self._database_name, self._schema_id)
             schema_exact_matches: dict = response_json["exact_matches"]
-            if "fields" not in response_json:
-                schema_fields_returned = {
-                    "ST": "Unknown",
-                    "Clonal Complex": "Unknown"
-                }
-            else:
-                schema_fields_returned: Mapping[str, str] = response_json["fields"]
+            response_json.setdefault("fields", dict)
+            schema_fields_returned: dict[str, str] = response_json["fields"]
+            schema_fields_returned.setdefault("ST", "unknown")
+            schema_fields_returned.setdefault("clonal_complex", "unknown")
             allele_map: dict[str, list[Allele]] = defaultdict(list)
             for exact_match_loci, exact_match_alleles in schema_exact_matches.items():
                 for exact_match_allele in exact_match_alleles:
@@ -68,7 +65,7 @@ class BIGSdbMLSTProfiler(AbstractAsyncContextManager):
         return await self.fetch_mlst_st(alleles)
 
 
-    async def profile_multiple_strings(self, namedStrings: AsyncIterable[NamedString], stop_on_fail: bool = False) -> AsyncGenerator[Union[tuple[str, MLSTProfile], tuple[str, None]], Any]:
+    async def profile_multiple_strings(self, namedStrings: AsyncIterable[NamedString], stop_on_fail: bool = False) -> AsyncGenerator[tuple[str, Union[MLSTProfile, None]], Any]:
         async for named_string in namedStrings:
             try:
                 yield (named_string.name, await self.profile_string(named_string.sequence))