Add checking to Pydantic so that null and empty questions cannot be sent to Harmony

Author: woodthom2

src/harmony/matching/default_matcher.py

@@ -82,6 +82,10 @@ def match_instruments(
         clustering_algorithm: str = "affinity_propagation",
         num_clusters_for_kmeans: int = None
 ) -> MatchResult:
+    for instrument in instruments:
+        for question in instrument.questions:
+            if question.question_text is None or question.question_text == "":
+                raise ValueError("Invalid argument: you cannot send an empty question to Harmony. Please remove all null and empty questions.")
     return match_instruments_with_function(
         instruments=instruments,
         query=query,

src/harmony/matching/matcher.py

@@ -616,6 +616,10 @@ def match_instruments_with_function(
     for instrument in instruments:
         all_questions.extend(instrument.questions)
 
+    for question in all_questions:
+        if question.question_text is None or question.question_text == "":
+            raise ValueError("Invalid argument: you cannot send an empty question to Harmony. Please remove all null and empty questions.")
+            
     text_vectors, new_vectors_dict = create_full_text_vectors(
         all_questions=[q.question_text for q in all_questions],
         query=query,

src/harmony/schemas/requests/text.py

@@ -61,7 +61,7 @@ class RawFile(BaseModel):
 class Question(BaseModel):
     question_no: Optional[str] = Field(None, description="Number of the question")
     question_intro: Optional[str] = Field(None, description="Introductory text applying to the question")
-    question_text: str = Field(description="Text of the question")
+    question_text: str = Field(description="Text of the question", min_length=1)
     options: List[str] = Field([], description="The possible answer options")
     source_page: int = Field(0, description="The page of the PDF on which the question was located, zero-indexed")
     instrument_id: Optional[str] = Field(None, description="Unique identifier for the instrument (UUID-4)")