Add HuggingFace sentiment analysis skill category

skills/huggingface/README.md

@@ -0,0 +1,19 @@
+# HuggingFace Skill Category
+
+This skill category enables natural language understanding tools via HuggingFace Transformers.
+
+## Included Skills
+
+### 🧠 Sentiment Analysis
+
+**Description:**  
+Analyzes the sentiment of input text (e.g., "I love this!") and returns a label (`POSITIVE`, `NEGATIVE`, etc.) along with a confidence score.
+
+## Configuration Example
+
+```yaml
+skills:
+  huggingface:
+    states:
+      sentiment_analysis: public
+

skills/huggingface/__init__.py

@@ -0,0 +1,36 @@
+from typing import TypedDict
+from abstracts.skill import SkillStoreABC
+from skills.base import SkillConfig, SkillState
+from skills.huggingface.base import HuggingFaceBaseTool
+from skills.huggingface.sentiment_analysis import SentimentAnalysis
+
+_cache: dict[str, HuggingFaceBaseTool] = {}
+
+class SkillStates(TypedDict):
+    sentiment_analysis: SkillState
+
+class Config(SkillConfig):
+    states: SkillStates
+
+async def get_skills(
+    config: "Config",
+    is_private: bool,
+    store: SkillStoreABC,
+    **_,
+) -> list[HuggingFaceBaseTool]:
+    available_skills = []
+    for skill_name, state in config["states"].items():
+        if state == "disabled":
+            continue
+        elif state == "public" or (state == "private" and is_private):
+            available_skills.append(skill_name)
+
+    return [get_huggingface_skill(name, store) for name in available_skills]
+
+def get_huggingface_skill(name: str, store: SkillStoreABC) -> HuggingFaceBaseTool:
+    if name == "sentiment_analysis":
+        if name not in _cache:
+            _cache[name] = SentimentAnalysis(skill_store=store)
+        return _cache[name]
+    raise ValueError(f"Unknown HuggingFace skill: {name}")
+

skills/huggingface/base.py

@@ -0,0 +1,17 @@
+from typing import Type
+from pydantic import BaseModel, Field
+from abstracts.skill import SkillStoreABC
+from skills.base import IntentKitSkill
+
+class HuggingFaceBaseTool(IntentKitSkill):
+    """Base class for HuggingFace NLP tools."""
+
+    name: str = Field(description="Tool name")
+    description: str = Field(description="What the tool does")
+    args_schema: Type[BaseModel]
+    skill_store: SkillStoreABC = Field(description="Skill store for data persistence")
+
+    @property
+    def category(self) -> str:
+        return "huggingface"
+

skills/huggingface/schema.json

@@ -0,0 +1,23 @@
+{
+  "$schema": "http://json-schema.org/draft-07/schema#",
+  "type": "object",
+  "title": "HuggingFace Skills",
+  "description": "Configuration schema for HuggingFace-powered skills",
+  "properties": {
+    "states": {
+      "type": "object",
+      "properties": {
+        "sentiment_analysis": {
+          "type": "string",
+          "title": "Sentiment Analysis",
+          "enum": ["disabled", "public", "private"],
+          "description": "State of the sentiment analysis skill"
+        }
+      },
+      "description": "States for HuggingFace skills"
+    }
+  },
+  "required": ["states"],
+  "additionalProperties": true
+}
+

skills/huggingface/sentiment_analysis.py

@@ -0,0 +1,22 @@
+from typing import Type
+from pydantic import BaseModel, Field
+from skills.huggingface.base import HuggingFaceBaseTool
+from transformers import pipeline
+
+class SentimentAnalysisInput(BaseModel):
+    text: str = Field(description="Text to analyze sentiment for")
+
+class SentimentAnalysis(HuggingFaceBaseTool):
+    name: str = "sentiment_analysis"
+    description: str = "Analyze sentiment of a given text using HuggingFace Transformers"
+    args_schema: Type[BaseModel] = SentimentAnalysisInput
+
+    _analyzer = pipeline("sentiment-analysis")
+
+    async def _arun(self, text: str, **kwargs) -> str:
+        try:
+            result = self._analyzer(text)[0]
+            return f"Sentiment: {result['label']} with score {result['score']:.2f}"
+        except Exception as e:
+            return f"Failed to analyze sentiment: {str(e)}"
+