Inference API (FastAPI) – first push

Dockerfile

@@ -0,0 +1,6 @@
+FROM python:3.11-slim
+WORKDIR /app
+COPY requirements.txt .
+RUN pip install -r requirements.txt
+COPY . .
+CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "9000"]

README.md

@@ -0,0 +1,10 @@
+# Incident ML Inference API
+
+This FastAPI app provides endpoints for:
+- `/health` - service health check
+- `/analyze` - classify incident category & sentiment
+
+## Deploy to Hugging Face Spaces
+- Create new Space (FastAPI SDK)
+- Push contents of this `inference/` folder to the Space repo
+- Public URL will serve the endpoints

app.py

@@ -0,0 +1,31 @@
+# inference/app.py
+from fastapi import FastAPI
+from pydantic import BaseModel
+from pathlib import Path
+from transformers import pipeline
+
+app = FastAPI(title="Incident ML Inference API")
+
+LOCAL_MODEL = Path(__file__).resolve().parents[1] / "models" / "incident_classifier"
+
+# Category classifier (your fine-tuned model if available)
+if LOCAL_MODEL.exists():
+    incident_classifier = pipeline("text-classification", model=str(LOCAL_MODEL))
+else:
+    incident_classifier = pipeline("text-classification", model="cardiffnlp/twitter-xlm-roberta-base")
+
+# Sentiment (keep public model for now)
+sentiment_analyzer = pipeline("sentiment-analysis", model="cardiffnlp/twitter-xlm-roberta-base-sentiment")
+
+class AnalyzeIn(BaseModel):
+    text: str
+
+@app.get("/health")
+def health(): return {"ok": True, "using_local_model": LOCAL_MODEL.exists()}
+
+@app.post("/analyze")
+def analyze(data: AnalyzeIn):
+    return {
+        "category": incident_classifier(data.text),
+        "sentiment": sentiment_analyzer(data.text)
+    }

requirements.txt

@@ -0,0 +1,4 @@
+fastapi
+uvicorn[standard]
+transformers
+torch

runtime.txt

@@ -0,0 +1 @@
+python-3.11