Fix/topic modeling

.env.sample

@@ -55,4 +55,5 @@ TEXTEXTRACTION_ECS_ENDPOINT=
 SUMMARIZATION_V3_ECS_ENDPOINT=
 ENTRYEXTRACTION_ECS_ENDPOINT=
 GEOLOCATION_ECS_ENDPOINT=
+TOPICMODEL_ECS_ENDPOINT=
 

.github/workflows/build.yml

@@ -58,6 +58,7 @@ jobs:
       SUMMARIZATION_V3_ECS_ENDPOINT: ''
       ENTRYEXTRACTION_ECS_ENDPOINT: ''
       GEOLOCATION_ECS_ENDPOINT: ''
+      TOPICMODEL_ECS_ENDPOINT: ''
 
       # Celery
       CELERY_BROKER_URL: ''

analysis_module/mockserver.py

@@ -188,7 +188,7 @@ def process_topicmodeling(body) -> Any:
         for x in range(0, len(excerpt_ids), ceil(len(excerpt_ids) / clusters))
     ]
 
-    data = {key: val for key, val in enumerate(data)}
+    data = dict(enumerate(data))
 
     filepath = save_data_local_and_get_url(
         dir_name="topicmodel", client_id=client_id, data=data

analysis_module/tests/test_apis.py

@@ -36,12 +36,10 @@ def test_topicmodel_incomplete_data(self):
             errors = resp.json()["field_errors"]
             assert param in errors
 
-    @patch('analysis_module.views.analysis_module.spin_ecs_container')
-    def test_topicmodel_valid_request(self, spin_ecs_mock):
+    def test_topicmodel_valid_request(self):
         """
         This tests for a topicmodel api with valid data
         """
-        requests_count = NLPRequest.objects.count()
         valid_data = {
             "entries_url": "https://someurl.com/entries",
             "cluster_size": 2,
@@ -52,15 +50,6 @@ def test_topicmodel_valid_request(self, spin_ecs_mock):
             self.set_credentials()
             resp = self.client.post(self.TOPICMODELING_URL, valid_data)
         assert resp.status_code == 202
-        spin_ecs_mock.delay.assert_called_once()
-        new_requests_count = NLPRequest.objects.count()
-        assert \
-            new_requests_count == requests_count + 1, \
-            "One more NLPRequest object should be created"
-        assert NLPRequest.objects.filter(
-            type="topicmodel",
-            created_by=self.user,
-        ).exists()
 
     def test_ngrams_incomplete_data(self):
         """

analysis_module/utils.py

@@ -17,7 +17,8 @@
     SUMMARIZATION_V3_ECS_ENDPOINT,
     TEXT_EXTRACTION_ECS_ENDPOINT,
     ENTRYEXTRACTION_ECS_ENDPOINT,
-    GEOLOCATION_ECS_ENDPOINT
+    GEOLOCATION_ECS_ENDPOINT,
+    TOPICMODEL_ECS_ENDPOINT
 )
 
 logger = logging.getLogger(__name__)
@@ -253,6 +254,8 @@ def get_ecs_id_param_name(request_type: NLPRequest.FeaturesType):
         return "entryextraction_id"  # not needed probably, just to be in line with the rest.
     if request_type == NLPRequest.FeaturesType.GEOLOCATION:
         return "geolocation_id"
+    if request_type == NLPRequest.FeaturesType.TOPICMODEL:
+        return "topicmodel_id"
     return None
 
 
@@ -265,4 +268,6 @@ def get_ecs_url(request_type: NLPRequest.FeaturesType):
         return urljoin(ENTRYEXTRACTION_ECS_ENDPOINT, "/extract_entries")
     elif request_type == NLPRequest.FeaturesType.GEOLOCATION:
         return urljoin(GEOLOCATION_ECS_ENDPOINT, "/get_geolocations")
+    elif request_type == NLPRequest.FeaturesType.TOPICMODEL:
+        return urljoin(TOPICMODEL_ECS_ENDPOINT, "/get_excerpt_clusters")
     return None

analysis_module/views/analysis_module.py

@@ -65,7 +65,32 @@ def process_request(
 @api_view(["POST"])
 @permission_classes([IsAuthenticated])
 def topic_modeling(request: Request):
-    return process_request(TopicModelDeepRequest, request, "topicmodel")
+    serializer = TopicModelDeepRequest(data=request.data)
+    serializer.is_valid(raise_exception=True)
+
+    if serializer.validated_data.get("mock") or IS_MOCKSERVER:
+        return process_mock_request(
+            request=serializer.validated_data,
+            request_type=NLPRequest.FeaturesType.TOPICMODEL
+        )
+
+    nlp_request = NLPRequest.objects.create(
+        client_id=serializer.validated_data["client_id"],
+        type=NLPRequest.FeaturesType.TOPICMODEL,
+        request_params=serializer.validated_data,
+        created_by=request.user
+    )
+    transaction.on_commit(lambda: send_ecs_http_request(nlp_request))
+    resp = {
+        "client_id": serializer.data.get("client_id"),
+        "type": NLPRequest.FeaturesType.TOPICMODEL,
+        "message": "Request has been successfully processed.",
+        "request_id": str(nlp_request.unique_id),
+    }
+    return Response(
+        resp,
+        status=status.HTTP_202_ACCEPTED,
+    )
 
 
 @api_view(["POST"])

core_server/settings.py

@@ -57,6 +57,7 @@
 TEXT_EXTRACTION_ECS_ENDPOINT = env("TEXTEXTRACTION_ECS_ENDPOINT")
 ENTRYEXTRACTION_ECS_ENDPOINT = env("ENTRYEXTRACTION_ECS_ENDPOINT")
 GEOLOCATION_ECS_ENDPOINT = env("GEOLOCATION_ECS_ENDPOINT")
+TOPICMODEL_ECS_ENDPOINT = env("TOPICMODEL_ECS_ENDPOINT")
 
 
 CALLBACK_MAX_RETRIES_LIMIT = env("CALLBACK_MAX_RETRIES_LIMIT")

docker-compose-prod.yml

@@ -47,6 +47,7 @@ x-server: &base-server-config
     TEXTEXTRACTION_ECS_ENDPOINT: ${TEXTEXTRACTION_ECS_ENDPOINT:?Provide text extraction endpoint}
     ENTRYEXTRACTION_ECS_ENDPOINT: ${ENTRYEXTRACTION_ECS_ENDPOINT:?Provide entry extraction endpoint}
     GEOLOCATION_ECS_ENDPOINT: ${GEOLOCATION_ECS_ENDPOINT:?Provide geolocation endpoint}
+    TOPICMODEL_ECS_ENDPOINT: ${TOPICMODEL_ECS_ENDPOINT:?Provide topic model endpoint}
 
     # MODEL_INFO
     CLASSIFICATION_MODEL_ID: ${CLASSIFICATION_MODEL_ID:-classification-model}

docker-compose.yml

@@ -56,6 +56,7 @@ x-server: &base-server-config
     TEXTEXTRACTION_ECS_ENDPOINT: ${TEXTEXTRACTION_ECS_ENDPOINT:?Provide text extraction endpoint}
     ENTRYEXTRACTION_ECS_ENDPOINT: ${ENTRYEXTRACTION_ECS_ENDPOINT:?Provide entry extraction endpoint}
     GEOLOCATION_ECS_ENDPOINT: ${GEOLOCATION_ECS_ENDPOINT:?Provide geolocation endpoint}
+    TOPICMODEL_ECS_ENDPOINT: ${TOPICMODEL_ECS_ENDPOINT:?Provide topic model endpoint}
 
     # SENTRY
     SENTRY_DSN: ${SENTRY_DSN:-}