fix: Handle empty list of phrases when finding keywords

Although we were handling an empty string for the `find_keywords/2`
function, it was still possible to have a non-empty string that resulted
in an empty list of phrases from the token classification and clean up
process. This rearranges this logic to create another function named
`determine_keywords/2` that has a clause to handle an empty list of
phrases and return an empty list of keywords.

lib/medical_transcription_web/components/transcription_text_component.ex

@@ -130,27 +130,38 @@ defmodule MedicalTranscriptionWeb.Components.TranscriptionTextComponent do
 
   defp find_keywords(live_view_pid, text) do
     # First, we use token classification to determine parts of speech and then retrieve the verb and adjective+noun phrases.
-    %{entities: entities} =
-      Nx.Serving.batched_run(MedicalTranscription.TokenClassificationServing, text)
+    phrases = classify_tokens(text)
 
-    phrases = KeywordFinder.cleanup_phrases(entities)
+    # Then, we use one of two processes to determine which to show as keywords
+    full_keywords = determine_keywords(text, phrases)
 
-    # Then, we use one of two processes to determine which to show as keywords:
-    # 1. A slower process that looks to classify the text by the extracted phrases.
-    # serving = KeywordFinder.prepare_zero_shot_classification_serving(phrases)
-    # %{predictions: predictions} = Nx.Serving.run(serving, text)
-
-    # 2. A fast process finding the phrase closest in vector space to the whole text.
-    predictions = KeywordFinder.find_most_similar_label(text, phrases, 2)
+    send(live_view_pid, {:new_keywords, full_keywords})
 
     # For now, retrieve the top three keywords that have a score of more than 0.25
     keywords =
-      predictions
+      full_keywords
       |> Enum.filter(fn keyword -> keyword.score > 0.25 end)
       |> Enum.take(3)
 
-    send(live_view_pid, {:new_keywords, predictions})
-
     {:ok, %{keywords: keywords}}
   end
+
+  defp classify_tokens(text) do
+    %{entities: entities} =
+      Nx.Serving.batched_run(MedicalTranscription.TokenClassificationServing, text)
+
+    KeywordFinder.cleanup_phrases(entities)
+  end
+
+  # This clause handles cases where there is transcribed text, but no phrases were found.
+  defp determine_keywords(_text, []), do: []
+
+  defp determine_keywords(text, phrases) do
+    # 1. A slower process that looks to classify the text by the extracted phrases.
+    # serving = KeywordFinder.prepare_zero_shot_classification_serving(phrases)
+    # %{predictions: predictions} = Nx.Serving.run(serving, text)
+
+    # 2. A fast process finding the phrase closest in vector space to the whole text.
+    KeywordFinder.find_most_similar_label(text, phrases, 2)
+  end
 end