feat: +OAS framework

metagpt/actions/azure_tts.py

@@ -1,53 +0,0 @@
-#!/usr/bin/env python
-# -*- coding: utf-8 -*-
-"""
-@Time    : 2023/6/9 22:22
-@Author  : Leo Xiao
-@File    : azure_tts.py
-"""
-from azure.cognitiveservices.speech import AudioConfig, SpeechConfig, SpeechSynthesizer
-
-from metagpt.actions.action import Action
-from metagpt.config import Config
-
-
-class AzureTTS(Action):
-    def __init__(self, name, context=None, llm=None):
-        super().__init__(name, context, llm)
-        self.config = Config()
-
-    # 参数参考：https://learn.microsoft.com/zh-cn/azure/cognitive-services/speech-service/language-support?tabs=tts#voice-styles-and-roles
-    def synthesize_speech(self, lang, voice, role, text, output_file):
-        subscription_key = self.config.get('AZURE_TTS_SUBSCRIPTION_KEY')
-        region = self.config.get('AZURE_TTS_REGION')
-        speech_config = SpeechConfig(
-            subscription=subscription_key, region=region)
-
-        speech_config.speech_synthesis_voice_name = voice
-        audio_config = AudioConfig(filename=output_file)
-        synthesizer = SpeechSynthesizer(
-            speech_config=speech_config,
-            audio_config=audio_config)
-
-        # if voice=="zh-CN-YunxiNeural":
-        ssml_string = f"""
-            <speak version='1.0' xmlns='http://www.w3.org/2001/10/synthesis' xml:lang='{lang}' xmlns:mstts='http://www.w3.org/2001/mstts'>
-                <voice name='{voice}'>
-                    <mstts:express-as style='affectionate' role='{role}'>
-                        {text}
-                    </mstts:express-as>
-                </voice>
-            </speak>
-            """
-
-        synthesizer.speak_ssml_async(ssml_string).get()
-
-
-if __name__ == "__main__":
-    azure_tts = AzureTTS("azure_tts")
-    azure_tts.synthesize_speech(
-        "zh-CN",
-        "zh-CN-YunxiNeural",
-        "Boy",
-        "你好，我是卡卡",
-        "output.wav")

metagpt/tools/azure_tts.py

@@ -0,0 +1,114 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+"""
+@Time    : 2023/8/17
+@Author  : mashenquan
+@File    : azure_tts.py
+@Desc    : azure TTS openapi, which provides text-to-speech functionality
+"""
+from pathlib import Path
+from uuid import uuid4
+import base64
+import sys
+
+sys.path.append(str(Path(__file__).resolve().parent.parent.parent))  # fix-bug: No module named 'metagpt'
+from metagpt.utils.common import initalize_enviroment
+from metagpt.logs import logger
+
+from azure.cognitiveservices.speech import AudioConfig, SpeechConfig, SpeechSynthesizer
+import os
+
+
+class AzureTTS:
+    """Azure Text-to-Speech"""
+
+    def __init__(self, subscription_key, region):
+        """
+        :param subscription_key: key is used to access your Azure AI service API, see: `https://portal.azure.com/` > `Resource Management` > `Keys and Endpoint`
+        :param region: This is the location (or region) of your resource. You may need to use this field when making calls to this API.
+        """
+        self.subscription_key = subscription_key if subscription_key else os.environ.get('AZURE_TTS_SUBSCRIPTION_KEY')
+        self.region = region if region else os.environ.get('AZURE_TTS_REGION')
+
+    # 参数参考：https://learn.microsoft.com/zh-cn/azure/cognitive-services/speech-service/language-support?tabs=tts#voice-styles-and-roles
+    def synthesize_speech(self, lang, voice, text, output_file):
+        speech_config = SpeechConfig(
+            subscription=self.subscription_key, region=self.region)
+        speech_config.speech_synthesis_voice_name = voice
+        audio_config = AudioConfig(filename=output_file)
+        synthesizer = SpeechSynthesizer(
+            speech_config=speech_config,
+            audio_config=audio_config)
+
+        # More detail: https://learn.microsoft.com/en-us/azure/ai-services/speech-service/speech-synthesis-markup-voice
+        ssml_string = "<speak version='1.0' xmlns='http://www.w3.org/2001/10/synthesis' " \
+                      f"xml:lang='{lang}' xmlns:mstts='http://www.w3.org/2001/mstts'>" \
+                      f"<voice name='{voice}'>{text}</voice></speak>"
+
+        return synthesizer.speak_ssml_async(ssml_string).get()
+
+    @staticmethod
+    def role_style_text(role, style, text):
+        return f'<mstts:express-as role="{role}" style="{style}">{text}</mstts:express-as>'
+
+    @staticmethod
+    def role_text(role, text):
+        return f'<mstts:express-as role="{role}">{text}</mstts:express-as>'
+
+    @staticmethod
+    def style_text(style, text):
+        return f'<mstts:express-as style="{style}">{text}</mstts:express-as>'
+
+
+# Export
+def openapi_azsure_tts(text, lang="", voice="", style="", role="", subscription_key="", region=""):
+    """openapi/tts/azsure
+    For more details, check out:`https://learn.microsoft.com/en-us/azure/ai-services/speech-service/language-support?tabs=tts`
+
+    :param lang: The value can contain a language code such as en (English), or a locale such as en-US (English - United States). For more details, checkout: `https://learn.microsoft.com/en-us/azure/ai-services/speech-service/language-support?tabs=tts`
+    :param voice: For more details, checkout: `https://learn.microsoft.com/en-us/azure/ai-services/speech-service/language-support?tabs=tts`, `https://speech.microsoft.com/portal/voicegallery`
+    :param style: Speaking style to express different emotions like cheerfulness, empathy, and calm. For more details, checkout: `https://learn.microsoft.com/en-us/azure/ai-services/speech-service/language-support?tabs=tts`
+    :param role: With roles, the same voice can act as a different age and gender. For more details, checkout: `https://learn.microsoft.com/en-us/azure/ai-services/speech-service/language-support?tabs=tts`
+    :param text: Text to convert
+    :param subscription_key: key is used to access your Azure AI service API, see: `https://portal.azure.com/` > `Resource Management` > `Keys and Endpoint`
+    :param region: This is the location (or region) of your resource. You may need to use this field when making calls to this API.
+    :return: Returns the Base64-encoded .wav file data if successful, otherwise an empty string.
+
+    """
+    if not text:
+        return ""
+
+    if not lang:
+        lang = "zh-CN"
+    if not voice:
+        voice = "zh-CN-XiaomoNeural"
+    if not role:
+        role = "Girl"
+    if not style:
+        style = "affectionate"
+    if not subscription_key:
+        subscription_key = os.environ.get("AZURE_TTS_SUBSCRIPTION_KEY")
+    if not region:
+        region = os.environ.get("AZURE_TTS_REGION")
+
+    xml_value = AzureTTS.role_style_text(role=role, style=style, text=text)
+    tts = AzureTTS(subscription_key=subscription_key, region=region)
+    filename = Path(__file__).resolve().parent / (str(uuid4()).replace("-", "") + ".wav")
+    try:
+        tts.synthesize_speech(lang=lang, voice=voice, text=xml_value, output_file=str(filename))
+        with open(str(filename), mode="rb") as reader:
+            data = reader.read()
+            base64_string = base64.b64encode(data).decode('utf-8')
+        filename.unlink()
+    except Exception as e:
+        logger.error(f"text:{text}, error:{e}")
+        return ""
+
+    return base64_string
+
+
+if __name__ == "__main__":
+    initalize_enviroment()
+
+    v = openapi_azsure_tts("测试，test")
+    print(v)

metagpt/tools/hello.py

@@ -0,0 +1,27 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+"""
+@Time    : 2023/5/2 16:03
+@Author  : mashenquan
+@File    : hello.py
+@Desc    : Implement the OpenAPI Specification 3.0 demo and use the following command to test the HTTP service:
+
+        curl -X 'POST' \
+        'http://localhost:8080/openapi/greeting/dave' \
+        -H 'accept: text/plain' \
+        -H 'Content-Type: application/json' \
+        -d '{}'
+"""
+
+import connexion
+
+
+# openapi implement
+def post_greeting(name: str) -> str:
+    return f"Hello {name}\n"
+
+
+if __name__ == "__main__":
+    app = connexion.AioHttpApp(__name__, specification_dir='../../spec/')
+    app.add_api("openapi.yaml", arguments={"title": "Hello World Example"})
+    app.run(port=8080)

metagpt/tools/metagpt_openapi_svc.py

@@ -0,0 +1,20 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+"""
+@Time    : 2023/8/17
+@Author  : mashenquan
+@File    : metagpt_openapi_svc.py
+@Desc    : MetaGPT OpenAPI REST API service
+"""
+from pathlib import Path
+import sys
+import connexion
+sys.path.append(str(Path(__file__).resolve().parent.parent.parent))  # fix-bug: No module named 'metagpt'
+from metagpt.utils.common import initalize_enviroment
+
+if __name__ == "__main__":
+    initalize_enviroment()
+
+    app = connexion.AioHttpApp(__name__, specification_dir='../../spec/')
+    app.add_api("metagpt_openapi.yaml")
+    app.run(port=8080)

metagpt/utils/common.py

@@ -4,14 +4,18 @@
 @Time    : 2023/4/29 16:07
 @Author  : alexanderwu
 @File    : common.py
+@Modified By: mashenquan, 2023-8-17, add `initalize_enviroment()` to load `config/config.yaml` to `os.environ`
 """
 import ast
 import contextlib
 import inspect
 import os
 import re
+from pathlib import Path
 from typing import List, Tuple
 
+import yaml
+
 from metagpt.logs import logger
 
 
@@ -254,3 +258,12 @@ def parse_recipient(text):
     pattern = r"## Send To:\s*([A-Za-z]+)\s*?"  # hard code for now
     recipient = re.search(pattern, text)
     return recipient.group(1) if recipient else ""
+
+
+def initalize_enviroment():
+    """Load `config/config.yaml` to `os.environ`"""
+    yaml_file_path = Path(__file__).resolve().parent.parent.parent / "config/config.yaml"
+    with open(str(yaml_file_path), "r") as yaml_file:
+        data = yaml.safe_load(yaml_file)
+        for k, v in data.items():
+            os.environ[k] = str(v)

requirements.txt

@@ -36,4 +36,6 @@ anthropic==0.3.6
 typing-inspect==0.8.0
 typing_extensions==4.5.0
 libcst==1.0.1
-qdrant-client==1.4.0
+qdrant-client==1.4.0
+connexion[swagger-ui]
+aiohttp_jinja2

spec/metagpt_openapi.yaml

@@ -0,0 +1,64 @@
+openapi: "3.0.0"
+
+info:
+  title: "MetaGPT Export OpenAPIs"
+  version: "1.0"
+servers:
+  - url: "/openapi"
+
+paths:
+  /tts/azsure:
+    post:
+      summary: "Convert Text to Base64-encoded .wav File Stream"
+      description: "For more details, check out: [Azure Text-to_Speech](https://learn.microsoft.com/en-us/azure/ai-services/speech-service/language-support?tabs=tts)"
+      operationId: azure_tts.openapi_azsure_tts
+      requestBody:
+        required: true
+        content:
+          application/json:
+            schema:
+              type: object
+              required:
+                - text
+              properties:
+                text:
+                  type: string
+                  description: Text to convert
+                lang:
+                  type: string
+                  description: The language code or locale, e.g., en-US (English - United States)
+                  default: "zh-CN"
+                voice:
+                  type: string
+                  description: "Voice style, see: [Azure Text-to_Speech](https://learn.microsoft.com/en-us/azure/ai-services/speech-service/language-support?tabs=tts), [Voice Gallery](https://speech.microsoft.com/portal/voicegallery)"
+                  default: "zh-CN-XiaomoNeural"
+                style:
+                  type: string
+                  description: "Speaking style to express different emotions. For more details, checkout: [Azure Text-to_Speech](https://learn.microsoft.com/en-us/azure/ai-services/speech-service/language-support?tabs=tts)"
+                  default: "affectionate"
+                role:
+                  type: string
+                  description: "Role to specify age and gender. For more details, checkout: [Azure Text-to_Speech](https://learn.microsoft.com/en-us/azure/ai-services/speech-service/language-support?tabs=tts)"
+                  default: "Girl"
+                subscription_key:
+                  type: string
+                  description: "Key used to access Azure AI service API, see: [Azure Portal](https://portal.azure.com/) > `Resource Management` > `Keys and Endpoint`"
+                  default: ""
+                region:
+                  type: string
+                  description: "Location (or region) of your resource, see: [Azure Portal](https://portal.azure.com/) > `Resource Management` > `Keys and Endpoint`"
+                  default: ""
+      responses:
+        '200':
+          description: "Base64-encoded .wav file data if successful, otherwise an empty string."
+          content:
+            application/json:
+              schema:
+                type: object
+                properties:
+                  result:
+                    type: string
+        '400':
+          description: Bad Request
+        '500':
+          description: Bad Request

spec/openapi.yaml

@@ -0,0 +1,35 @@
+openapi: "3.0.0"
+
+info:
+  title: Hello World
+  version: "1.0"
+servers:
+  - url: /openapi
+
+paths:
+  /greeting/{name}:
+    post:
+      summary: Generate greeting
+      description: Generates a greeting message.
+      operationId: hello.post_greeting
+      responses:
+        200:
+          description: greeting response
+          content:
+            text/plain:
+              schema:
+                type: string
+                example: "hello dave!"
+      parameters:
+        - name: name
+          in: path
+          description: Name of the person to greet.
+          required: true
+          schema:
+            type: string
+            example: "dave"
+      requestBody:
+        content:
+          application/json:
+            schema:
+              type: object

tests/metagpt/tools/test_azure_tts.py

@@ -4,8 +4,13 @@
 @Time    : 2023/7/1 22:50
 @Author  : alexanderwu
 @File    : test_azure_tts.py
+@Modified By: mashenquan, 2023-8-17, move to `tools` folder.
 """
-from metagpt.actions.azure_tts import AzureTTS
+import sys
+from pathlib import Path
+
+sys.path.append(str(Path(__file__).resolve().parent.parent.parent.parent))  # fix-bug: No module named 'metagpt'
+from metagpt.tools.azure_tts import AzureTTS
 
 
 def test_azure_tts():