[+] improve worker.py

deps/test-task-2025-07-17-v2/python/online/fxreader/pr34/test_task_2025_07_17_v2/transform/worker.py

@@ -1,10 +1,19 @@
 import transformers
 import transformers.pipelines
 
-from typing import (Any, cast, Callable, Protocol, Literal,)
+from typing import (
+	Any, cast, Callable, Protocol, Literal, TypedDict,
+	TypeAlias,
+)
 
 class SummarizerPipeline(Protocol):
-	def predict(self, data: str) -> str: ...
+	class predict_t:
+		class output_t(TypedDict):
+			summary_text: str
+
+		res_t : TypeAlias = list[output_t]
+
+	def predict(self, data: str) -> predict_t.res_t: ...
 
 class Pipeline(Protocol):
 	def __call__(
@@ -14,34 +23,40 @@ class Pipeline(Protocol):
 		tokenizer: Any,
 	) -> 'SummarizerPipeline': ...
 
+
 class Summarizer:
 	def __init__(self) -> None:
-		self.model = cast(
-			Callable[[str], Any],
-			getattr(transformers.AutoTokenizer, 'from_pretrained')(
-				'sshleifer/distilbart-cnn-12-6',
-			)
-		)
 		self.tokenizer = cast(
 			Callable[[str], Any],
-			getattr(transformers.AutoModelForSeq2SeqLM, 'from_pretrained')(
+			getattr(transformers.AutoTokenizer, 'from_pretrained')
+		)(
 			'sshleifer/distilbart-cnn-12-6',
 		)
+		self.model = cast(
+			Callable[[str], Any],
+			getattr(transformers.AutoModelForSeq2SeqLM, 'from_pretrained')
+		)(
+			'sshleifer/distilbart-cnn-12-6',
 		)
 
 		self.summarizer = cast(
 			Pipeline,
-			getattr(transformers.pipelines, 'pipeline')
+			# getattr(transformers.pipelines, 'pipeline')
+			getattr(transformers, 'pipeline')
 		)(
 			'summarization',
 			model=self.model,
 			tokenizer=self.tokenizer,
+			# framework='pt',
 		)
 
 	def summarize(
 		self,
 		data: list[str]
 	) -> list[str]:
-		return self.summarizer.predict(
+		res = self.summarizer.predict(
 			' '.join(data)
-		).split()
+		)
+		assert len(res) == 1
+
+		return res[0]['summary_text'].split()