feat: initial sketch of WhisperFE (WFE)

WhisperFE will be slightly different to other FEs in that there is/are specific
target words to be searched for. Not specifying these could be an error (this
commit specifies this as such) but a better approach may be to downgrade that to
a (logging) notice, and simply match nothing / early exit.

pipeline/feature_extractors.py

@@ -400,3 +400,40 @@ class JSONFeatureExtractor(FeatureExtractor):
 
     def teardown(self):
         pass
+
+
+class WordFeatureExtractor(FeatureExtractor):
+    """Feature extractor for specific word detection (uses Whisper)"""
+    def __init__(self, input_files=None, config=None):
+        if not input_files:
+            raise ValueError("No input files provided!")
+        self.input_files = input_files
+        self.config = config
+        self.features = []
+
+    def setup(self, words=[]):
+        """Setup the word feature extractor -- validate input files & config
+
+        Whisper expects a list of words to search for in the audio
+        """
+        logger.debug("WordFeatureExtractor setup")
+
+        # Validate input files
+        if not self.input_files:
+            raise ValueError("No input files provided")
+
+        # Validate words
+        if not words:
+            raise ValueError("No words provided")
+        words = words
+        # TODO: consider stripping punctuation since Whisper produces words+punctuation
+        # and we might want to strip the punctuation there too
+
+    def run(self):
+        """Extract features corresponding to supplied target words (defined in setup) for each input file
+
+        Use Whisper to detect words in the audio, then match these to target words and create features
+        """
+
+    def teardown(self):
+        """Clean up after Whisper"""