move streaming testing to test_sampling.py

tests/python_tests/test_llm_pipeline.py

@@ -11,7 +11,7 @@
 from pathlib import Path
 import torch
 
-from common import run_llm_pipeline_with_ref, convert_to_hf, StreamerWithResults
+from common import run_llm_pipeline_with_ref, convert_to_hf
 from ov_genai_test_utils import (
     get_models_list,
     read_model,
@@ -194,27 +194,6 @@ def test_callback_kwargs_one_string(callback):
     pipe.generate('table is made of', max_new_tokens=10, streamer=callback)
 
 
-@pytest.mark.parametrize("streamer", [StreamerWithResults()])
-@pytest.mark.parametrize("prompt", [
-    'table is made of', 
-    'The Sun is yellow because', 
-    '你好！ 你好嗎？'
-    'I have an interview about product speccing with the company Weekend Health. Give me an example of a question they might ask with regards about a new feature'
-])
-@pytest.mark.parametrize("generation_config", [dict(max_new_tokens=10), dict(max_new_tokens=300)])
-@pytest.mark.parametrize("model_descr", get_models_list())
-@pytest.mark.parametrize("use_cb", [True, False])
-@pytest.mark.precommit
-@pytest.mark.nightly
-def test_streamer_compare_texts(model_descr, generation_config, prompt, streamer, use_cb):
-    run_llm_pipeline_with_ref(model_id=model_descr[0], 
-                              prompts=[prompt], 
-                              generation_config=generation_config, 
-                              tmp_path=model_descr[1],
-                              use_cb=use_cb,
-                              streamer=streamer)
-
-
 @pytest.mark.parametrize("callback", [print, user_defined_callback, lambda subword: print(subword)])
 @pytest.mark.precommit
 @pytest.mark.nightly

tests/python_tests/test_sampling.py

@@ -9,7 +9,7 @@
 from openvino_genai import GenerationConfig, StopCriteria
 from typing import List, TypedDict
 
-from common import get_hugging_face_models, convert_models, run_llm_pipeline_with_ref, run_llm_pipeline
+from common import get_hugging_face_models, convert_models, run_llm_pipeline_with_ref, run_llm_pipeline, StreamerWithResults
 
 
 @pytest.mark.precommit
@@ -58,13 +58,26 @@ def test_stop_strings(tmp_path, generation_config):
 @pytest.mark.precommit
 @pytest.mark.parametrize("generation_config",
                          [dict(max_new_tokens=30),
-                          dict(max_new_tokens=30, repetition_penalty=2.0),],
-                         ids=["basic",
-                              "repetition_penalty",])
-def test_greedy(tmp_path, generation_config):
-    prompts = [ "What is OpenVINO?" ]
+                          dict(max_new_tokens=30, repetition_penalty=2.0),
+                          dict(max_new_tokens=300)],
+                         ids=["basic", "repetition_penalty", "long_max_new_tokens"])
+@pytest.mark.parametrize("streamer", [StreamerWithResults()])
+@pytest.mark.parametrize("prompt", [
+    'What is OpenVINO?',
+    'table is made of', 
+    'The Sun is yellow because', 
+    '你好！ 你好嗎？'
+    'I have an interview about product speccing with the company Weekend Health. Give me an example of a question they might ask with regards about a new feature'
+])
+@pytest.mark.parametrize("use_cb", [True, False])
+def test_greedy(tmp_path, generation_config, prompt, streamer, use_cb):
     model_id : str = "katuni4ka/tiny-random-phi3"
-    run_llm_pipeline_with_ref(model_id, prompts, generation_config, tmp_path)
+    run_llm_pipeline_with_ref(model_id=model_id, 
+                            prompts=[prompt], 
+                            generation_config=generation_config, 
+                            tmp_path=tmp_path,
+                            use_cb=use_cb,
+                            streamer=streamer)
 
 
 @pytest.mark.precommit