Merge pull request #21 from iorisa/feature/skills

Feature/skills

metagpt/learn/skill_loader.py

@@ -7,11 +7,13 @@
 @Desc    : Skill YAML Configuration Loader.
 """
 from pathlib import Path
-from typing import List, Dict, Optional
+from typing import Dict, List, Optional
 
 import yaml
 from pydantic import BaseModel, Field
 
+from metagpt.config import CONFIG
+
 
 class Example(BaseModel):
     ask: str
@@ -52,7 +54,7 @@ class SkillLoader:
     def __init__(self, skill_yaml_file_name: Path = None):
         if not skill_yaml_file_name:
             skill_yaml_file_name = Path(__file__).parent.parent.parent / ".well-known/skills.yaml"
-        with open(str(skill_yaml_file_name), 'r') as file:
+        with open(str(skill_yaml_file_name), "r") as file:
             skills = yaml.safe_load(file)
         self._skills = SkillsDeclaration(**skills)
 
@@ -62,8 +64,18 @@ class SkillLoader:
         if not entity_skills:
             return {}
 
+        agent_skills = CONFIG.agent_skills
+        if not agent_skills:
+            return {}
+
+        class AgentSkill(BaseModel):
+            name: str
+
+        names = [AgentSkill(**i).name for i in agent_skills]
         description_to_name_mappings = {}
         for s in entity_skills.skills:
+            if s.name not in names:
+                continue
             description_to_name_mappings[s.description] = s.name
 
         return description_to_name_mappings

metagpt/learn/text_to_image.py

@@ -33,7 +33,7 @@ async def text_to_image(text, size_type: str = "512x512", openai_api_key="", mod
         raise openai.error.InvalidRequestError("缺少必要的参数")
 
     s3 = S3()
-    url = await s3.cache(base64_data, BASE64_FORMAT)
+    url = await s3.cache(data=base64_data, file_ext=".png", format=BASE64_FORMAT)
     if url:
-        return url
+        return f"[{text}]({url})"
     return image_declaration + base64_data if base64_data else ""

metagpt/learn/text_to_speech.py

@@ -22,7 +22,7 @@ async def text_to_speech(
     role="Girl",
     subscription_key="",
     region="",
-    **kwargs
+    **kwargs,
 ):
     """Text to speech
     For more details, check out:`https://learn.microsoft.com/en-us/azure/ai-services/speech-service/language-support?tabs=tts`
@@ -41,9 +41,9 @@ async def text_to_speech(
     if (CONFIG.AZURE_TTS_SUBSCRIPTION_KEY and CONFIG.AZURE_TTS_REGION) or (subscription_key and region):
         base64_data = await oas3_azsure_tts(text, lang, voice, style, role, subscription_key, region)
         s3 = S3()
-        url = await s3.cache(base64_data, BASE64_FORMAT)
+        url = await s3.cache(data=base64_data, file_ext=".wav", format=BASE64_FORMAT)
         if url:
-            return url
+            return f"[{text}]({url})"
         return audio_declaration + base64_data if base64_data else base64_data
 
     raise openai.error.InvalidRequestError("缺少必要的参数")