merge develop into working branch

.travis.yml

@@ -8,26 +8,27 @@ before_install:
 - curl -L https://codeclimate.com/downloads/test-reporter/test-reporter-latest-linux-amd64 > ./cc-test-reporter
 - chmod +x ./cc-test-reporter
 before_script:
+- mkdir -p ~/.docker/cli-plugins/ && curl -SL https://github.com/docker/compose/releases/download/v2.30.1/docker-compose-linux-x86_64 -o ~/.docker/cli-plugins/docker-compose && chmod +x ~/.docker/cli-plugins/docker-compose && docker compose version
 - ./cc-test-reporter before-build
 - echo "$DOCKER_PASSWORD" | docker login -u "$DOCKER_USERNAME" --password-stdin
-- docker-compose build --pull
-- docker-compose -f docker-compose.yml -f docker-test.yml up -d
-- docker-compose logs -t -f &
+- docker compose build
+- docker compose -f docker-compose.yml -f docker-test.yml up -d
+- docker compose logs -t -f &
 - echo "Waiting for Elasticsearch indexes..." && until curl --silent --fail -I "http://localhost:9200/alegre_similarity_test"; do sleep 1; done
 - until curl --silent --fail -I "http://localhost:3100"; do sleep 1; done
 - echo "Waiting for model servers..." && while [[ ! '2' =~ $(redis-cli -n 1 SCARD 'SharedModel') ]]; do sleep 1; done
 #comment until fix timeout curl: (28) Operation timed out
-# - docker-compose exec alegre bash -c "curl --max-time 600.0 -OL https://raw.githubusercontent.com/meedan/check-api/develop/spec/pacts/check_api-alegre.json"
+# - docker compose exec alegre bash -c "curl --max-time 600.0 -OL https://raw.githubusercontent.com/meedan/check-api/develop/spec/pacts/check_api-alegre.json"
 jobs:
   include:
     - stage: tests
       name: unit-tests
-      script: docker-compose exec alegre make test
+      script: docker compose exec alegre make test
     - stage: tests
       name: contract-testing
-      script: docker-compose exec alegre make contract_testing
+      script: docker compose exec alegre make contract_testing
 after_script:
-- docker-compose exec alegre coverage xml
+- docker compose exec alegre coverage xml
 - if [[ "$TRAVIS_PULL_REQUEST" == "false" && "$TRAVIS_JOB_NAME" != "contract-testing" ]]; then ./cc-test-reporter after-build -t coverage.py -r $CC_TEST_REPORTER_ID --exit-code $TRAVIS_TEST_RESULT; fi
 notifications:
   slack:

app/main/lib/elastic_crud.py

@@ -52,7 +52,6 @@ def get_blocked_presto_response(task, model, modality):
     obj, temporary = get_object(task, model)
     doc_id = obj["doc_id"]
     callback_url = Presto.add_item_callback_url(app.config['ALEGRE_HOST'], modality)
-    app.logger.info(f"Object for {task} of model {model} with id of {doc_id} has requires_encoding value of {requires_encoding(obj)}")
     if requires_encoding(obj):
         blocked_results = []
         for model_key in obj.pop("models", []):

app/main/lib/media_crud.py

@@ -144,7 +144,6 @@ def get_blocked_presto_response(task, model, modality):
     callback_url =  Presto.add_item_callback_url(app.config['ALEGRE_HOST'], modality)
     if task.get("doc_id") is None:
         task["doc_id"] = str(uuid.uuid4())
-    app.logger.error(f"Object for {task} of model {model} with id of {obj.id} has requires_encoding value of {obj.requires_encoding}")
     if obj.requires_encoding:
         response = get_presto_request_response(modality, callback_url, task)
         # Warning: this is a blocking hold to wait until we get a response in 

app/main/lib/similarity.py

@@ -131,7 +131,7 @@ def callback_add_item(item, similarity_type):
     elif similarity_type == "text":
         function = callback_add_text
     if function:
-        response = function(item)
+        response = {"item": function(item)}
         app.logger.info(f"[Alegre Similarity] CallbackAddItem: [Item {item}, Similarity type: {similarity_type}] Response looks like {response}")
         return response
     else:

app/main/lib/text_similarity.py

@@ -109,7 +109,7 @@ def get_model_and_threshold(search_params):
   if 'per_model_threshold' in search_params and isinstance(search_params['per_model_threshold'], list) and [e for e in search_params['per_model_threshold'] if e["model"] == model_key]:
       threshold = [e for e in search_params['per_model_threshold'] if e["model"] == model_key][0]["value"]
   if threshold is None:
-      app.logger.error(
+      app.logger.warn(
           f"[Alegre Similarity] get_model_and_threshold - no threshold was specified, backing down to default of 0.9 - search_params is {search_params}")
       threshold = 0.9
   return model_key, threshold
@@ -194,7 +194,7 @@ def return_sources(results):
         and should noisily break since it would have other downstream consequences.
     """
     #TODO: remove underscore version after dependencies updated https://meedan.atlassian.net/browse/CV2-5546
-    return [dict(**r["_source"], **{"_id": r["_id"], "id": r["_id"], "index": r["_index"], "_score": r["_score"],"score": r["_score"]}) for r in results]
+    return [dict(**r["_source"], **{"_id": r["_id"], "id": r["_id"], "index": r["_index"], "_score": r["_score"], "score": r["_score"]}) for r in results]
 
 def strip_vectors(results):
     for result in results: