fix: remove audio primitives and fix imports (deepset-ai#22)

* remove audio primitives and fix imports

* pylint

* hooks

* pass

* wrong import

* tests

* check only filename

* path to string

* keep all other fields

nodes/text2speech/tests/test_nodes.py

@@ -18,14 +18,14 @@
 
 from transformers import WhisperProcessor, WhisperForConditionalGeneration
 
-from haystack.schema import Span, Answer, SpeechAnswer, Document, SpeechDocument
-from haystack.nodes.audio import AnswerToSpeech, DocumentToSpeech
-from haystack.nodes.audio._text_to_speech import TextToSpeech
-
+from haystack.schema import Span, Answer, Document
+from text2speech import AnswerToSpeech, DocumentToSpeech
+from text2speech.utils import TextToSpeech
 
 
 SAMPLES_PATH = Path(__file__).parent / "samples"
 
+
 class WhisperHelper:
  def __init__(self, model):
  self._processor = WhisperProcessor.from_pretrained(model)
@@ -50,6 +50,7 @@ def transcribe(self, media_file: str):
 def whisper_helper():
  return WhisperHelper("openai/whisper-medium")
 
+
 @pytest.mark.integration
 class TestTextToSpeech:
  def test_text_to_speech_audio_data(self, tmp_path, whisper_helper: WhisperHelper):
@@ -144,20 +145,19 @@ def test_answer_to_speech(self, tmp_path, whisper_helper: WhisperHelper):
  )
  results, _ = answer2speech.run(answers=[text_answer])
 
- audio_answer: SpeechAnswer = results["answers"][0]
- assert isinstance(audio_answer, SpeechAnswer)
- assert audio_answer.type == "generative"
- assert audio_answer.answer_audio.name == expected_audio_answer.name
- assert audio_answer.context_audio.name == expected_audio_context.name
- assert audio_answer.answer == "answer"
- assert audio_answer.context == "the context for this answer is here"
+ audio_answer: Answer = results["answers"][0]
+ assert isinstance(audio_answer, Answer)
+ assert audio_answer.answer.split(os.path.sep)[-1] == str(expected_audio_answer).split(os.path.sep)[-1]
+ assert audio_answer.context.split(os.path.sep)[-1] == str(expected_audio_context).split(os.path.sep)[-1]
  assert audio_answer.offsets_in_document == [Span(31, 37)]
  assert audio_answer.offsets_in_context == [Span(21, 27)]
+ assert audio_answer.meta["answer_text"] == "answer"
+ assert audio_answer.meta["context_text"] == "the context for this answer is here"
  assert audio_answer.meta["some_meta"] == "some_value"
  assert audio_answer.meta["audio_format"] == "wav"
 
  expected_doc = whisper_helper.transcribe(str(expected_audio_answer))
- generated_doc = whisper_helper.transcribe(str(audio_answer.answer_audio))
+ generated_doc = whisper_helper.transcribe(str(audio_answer.answer))
 
  assert expected_doc[0] in generated_doc[0]
 
@@ -178,15 +178,15 @@ def test_document_to_speech(self, tmp_path, whisper_helper: WhisperHelper):
 
  results, _ = doc2speech.run(documents=[text_doc])
 
- audio_doc: SpeechDocument = results["documents"][0]
- assert isinstance(audio_doc, SpeechDocument)
+ audio_doc: Document = results["documents"][0]
+ assert isinstance(audio_doc, Document)
  assert audio_doc.content_type == "audio"
- assert audio_doc.content_audio.name == expected_audio_content.name
- assert audio_doc.content == "this is the content of the document"
+ assert audio_doc.content.split(os.path.sep)[-1] == str(expected_audio_content).split(os.path.sep)[-1]
+ assert audio_doc.meta["content_text"] == "this is the content of the document"
  assert audio_doc.meta["name"] == "test_document.txt"
  assert audio_doc.meta["audio_format"] == "wav"
 
  expected_doc = whisper_helper.transcribe(str(expected_audio_content))
- generated_doc = whisper_helper.transcribe(str(audio_doc.content_audio))
+ generated_doc = whisper_helper.transcribe(str(audio_doc.content))
 
- assert expected_doc[0] in generated_doc[0]
+ assert expected_doc[0] in generated_doc[0]

nodes/text2speech/text2speech/answer_to_speech.py

@@ -6,8 +6,8 @@
 from tqdm.auto import tqdm
 
 from haystack.nodes import BaseComponent
-from haystack.schema import Answer, SpeechAnswer
-from haystack.nodes.audio._text_to_speech import TextToSpeech
+from haystack.schema import Answer
+from text2speech.utils import TextToSpeech
 
 
 class AnswerToSpeech(BaseComponent):
@@ -79,16 +79,17 @@ def run(self, answers: List[Answer]) -> Tuple[Dict[str, List[Answer]], str]: #
  text=answer.context, generated_audio_dir=self.generated_audio_dir, **self.params
  )
 
- audio_answer = SpeechAnswer.from_text_answer(
- answer_object=answer,
- audio_answer=answer_audio,
- audio_context=context_audio,
- additional_meta={
+ audio_answer = Answer.from_dict(answer.to_dict())
+ audio_answer.answer = str(answer_audio)
+ audio_answer.context = str(context_audio)
+ audio_answer.meta.update(
+ {
+ "answer_text": answer.answer,
+ "context_text": answer.context,
  "audio_format": self.params.get("audio_format", answer_audio.suffix.replace(".", "")),
  "sample_rate": self.converter.model.fs,
- },
+ }
  )
- audio_answer.type = "generative"
  audio_answers.append(audio_answer)
 
  return {"answers": audio_answers}, "output_1"

nodes/text2speech/text2speech/document_to_speech.py

@@ -4,8 +4,8 @@
 from tqdm.auto import tqdm
 
 from haystack.nodes import BaseComponent
-from haystack.schema import Document, SpeechDocument
-from haystack.nodes.audio._text_to_speech import TextToSpeech
+from haystack.schema import Document
+from text2speech.utils import TextToSpeech
 
 
 class DocumentToSpeech(BaseComponent):
@@ -60,15 +60,16 @@ def run(self, documents: List[Document]) -> Tuple[Dict[str, List[Document]], str
  content_audio = self.converter.text_to_audio_file(
  text=doc.content, generated_audio_dir=self.generated_audio_dir, **self.params
  )
- audio_document = SpeechDocument.from_text_document(
- document_object=doc,
- audio_content=content_audio,
- additional_meta={
+ audio_document = Document.from_dict(doc.to_dict())
+ audio_document.content = str(content_audio)
+ audio_document.content_type = "audio"
+ audio_document.meta.update(
+ {
+ "content_text": doc.content,
  "audio_format": self.params.get("audio_format", content_audio.suffix.replace(".", "")),
  "sample_rate": self.converter.model.fs,
- },
+ }
  )
- audio_document.type = "generative"
  audio_documents.append(audio_document)
 
  return {"documents": audio_documents}, "output_1"

nodes/text2speech/text2speech/utils/text_to_speech.py

@@ -9,7 +9,7 @@
 import torch
 from pydub import AudioSegment
 
-from haystack.errors import AudioNodeError
+from haystack.errors import NodeError
 from haystack.modeling.utils import initialize_device_settings
 
 
@@ -27,6 +27,12 @@
  )
 
 
+class Text2SpeechError(NodeError):
+ """
+ Error class for text2speech nodes
+ """
+
+
 class TextToSpeech:
  """
  This class converts text into audio using text-to-speech models.
@@ -138,12 +144,12 @@ def text_to_audio_data(self, text: str, _models_output_key: str = "wav") -> np.a
  """
  prediction = self.model(text)
  if not prediction:
- raise AudioNodeError(
+ raise Text2SpeechError(
  "The model returned no predictions. Make sure you selected a valid text-to-speech model."
  )
  output = prediction.get(_models_output_key, None)
  if output is None:
- raise AudioNodeError(
+ raise Text2SpeechError(
  f"The model returned no output under the {_models_output_key} key."
  f"The available output keys are {prediction.keys()}. Make sure you selected the right key."
  )