From 002bc56c0e68a79b7e5311b6799c946d6dd633bf Mon Sep 17 00:00:00 2001
From: zhanglei <zhanglei@fuzhi.ai>
Date: Thu, 18 Jan 2024 12:37:28 +0800
Subject: [PATCH 1/2] add: openai speech to text

---
 metagpt/provider/openai_api.py        | 4 ++++
 tests/metagpt/provider/test_openai.py | 9 +++++++++
 2 files changed, 13 insertions(+)

diff --git a/metagpt/provider/openai_api.py b/metagpt/provider/openai_api.py
index 3a9aca870..d6944eae6 100644
--- a/metagpt/provider/openai_api.py
+++ b/metagpt/provider/openai_api.py
@@ -239,3 +239,7 @@ class OpenAILLM(BaseLLM):
     async def atext_to_speech(self, **kwargs):
         """text to speech"""
         return await self.aclient.audio.speech.create(**kwargs)
+
+    async def aspeech_to_text(self, **kwargs):
+        """speech to text"""
+        return await self.aclient.audio.transcriptions.create(**kwargs)
diff --git a/tests/metagpt/provider/test_openai.py b/tests/metagpt/provider/test_openai.py
index 2d52ad10e..7a0dbe5c4 100644
--- a/tests/metagpt/provider/test_openai.py
+++ b/tests/metagpt/provider/test_openai.py
@@ -1,5 +1,6 @@
 import pytest
 
+from metagpt.const import TEST_DATA_PATH
 from metagpt.llm import LLM
 from metagpt.logs import logger
 from metagpt.provider import OpenAILLM
@@ -53,6 +54,14 @@ async def test_text_to_speech():
     assert 200 == resp.response.status_code
 
 
+@pytest.mark.asyncio
+async def test_speech_to_text():
+    llm = LLM()
+    audio_file = open(f"{TEST_DATA_PATH}/audio/hello.mp3", "rb")
+    resp = await llm.aspeech_to_text(file=audio_file, model="whisper-1")
+    assert "你好" == resp.text
+
+
 class TestOpenAI:
     def test_make_client_kwargs_without_proxy(self):
         instance = OpenAILLM(mock_llm_config)

From 99511fd264ec9354cb411c0762bd75f8850d7c74 Mon Sep 17 00:00:00 2001
From: zhanglei <zhanglei@fuzhi.ai>
Date: Thu, 18 Jan 2024 12:41:53 +0800
Subject: [PATCH 2/2] update:OpenAI text to speech unittest

---
 tests/metagpt/provider/test_openai.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tests/metagpt/provider/test_openai.py b/tests/metagpt/provider/test_openai.py
index 7a0dbe5c4..bc7f92f33 100644
--- a/tests/metagpt/provider/test_openai.py
+++ b/tests/metagpt/provider/test_openai.py
@@ -49,7 +49,7 @@ async def test_text_to_speech():
     resp = await llm.atext_to_speech(
         model="tts-1",
         voice="alloy",
-        input="人生说起来长，但知道一个岁月回头看，许多事件仅是仓促的。一段一段拼凑一起，合成了人生。苦难当头时，当下不免觉得是折磨；回头看，也不够是一段短短的人生旅程。",
+        input="人生说起来长，但直到一个岁月回头看，许多事件仅是仓促的。一段一段拼凑一起，合成了人生。苦难当头时，当下不免觉得是折磨；回头看，也不够是一段短短的人生旅程。",
     )
     assert 200 == resp.response.status_code