Flatten OCW topics so all of them get mapped to PWT topics when runni…

…ng the ETL pipeline (#1343)

learning_resources/etl/ocw.py

@@ -302,13 +302,8 @@ def transform_course(course_data: dict) -> dict:
     readable_term = f"+{slugify(term)}" if term else ""
     readable_year = f"_{course_data.get('year')}" if year else ""
     readable_id = f"{course_data[PRIMARY_COURSE_ID]}{readable_term}{readable_year}"
-    # The json returns basically a tuple of topics with subtopics - we only need
-    # to care about the subtopic, unless there's not one.
     topics = transform_topics(
-        [
-            {"name": topic[1] if len(topic) > 1 else topic[0]}
-            for topic in course_data.get("topics")
-        ],
+        [{"name": topic} for topics in course_data.get("topics") for topic in topics],
         OFFERED_BY["code"],
     )
     image_src = course_data.get("image_src")

learning_resources/etl/utils.py

@@ -70,7 +70,10 @@ def load_offeror_topic_map(offeror_code: str):
     mappings = {}
 
     for pmt_mapping in pmt_mappings:
-        mappings[pmt_mapping.topic_name] = pmt_mapping.topic.name
+        if pmt_mapping.topic_name not in mappings:
+            mappings[pmt_mapping.topic_name] = []
+
+        mappings[pmt_mapping.topic_name].append(pmt_mapping.topic.name)
 
     return mappings
 
@@ -93,7 +96,10 @@ def transform_topics(topics: list, offeror_code: str):
 
     for topic in topics:
         if topic["name"] in topic_mappings:
-            transformed_topics.append({"name": topic_mappings.get(topic["name"])})
+            [
+                transformed_topics.append({"name": mapped_topic})
+                for mapped_topic in topic_mappings.get(topic["name"])
+            ]
         else:
             base_topic = LearningResourceTopic.objects.filter(
                 name=topic["name"]