Upload folder using huggingface_hub

Dockerfile

@@ -0,0 +1,15 @@
+FROM python:3.12-slim
+
+WORKDIR /app
+
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt \
+    --extra-index-url https://download.pytorch.org/whl/cpu
+
+COPY app.py .
+
+# Writable cache for the model download (Spaces run as non-root uid 1000)
+ENV HF_HOME=/tmp/hf
+
+EXPOSE 7860
+CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]

README.md

@@ -1,10 +1,26 @@
 ---
-title: Nigerian Transaction Classifier Api
-emoji: 📈
+title: Nigerian Transaction Classifier API
+emoji: 🏦
 colorFrom: green
 colorTo: gray
 sdk: docker
+app_port: 7860
 pinned: false
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# Nigerian Transaction Classifier API
+
+Serves `killykilly/nigerian-transaction-classifier` (DistilBERT) over a plain
+HTTP API for the finance-analyzer backend, since the model is not available
+through HF serverless Inference Providers.
+
+The model is downloaded from the Hub at container startup, so pushing
+retrained weights to the model repo and restarting this Space is all that is
+needed to upgrade.
+
+## API
+
+- `GET /health` → `{"status": "ok", "model": "..."}`
+- `POST /classify` with `{"texts": ["debit | uber trip lagos", ...]}` →
+  `{"predictions": [{"label": "Transport", "score": 0.97}, ...]}`
+  (one prediction per input, same order)

app.py

@@ -0,0 +1,40 @@
+import os
+
+from fastapi import FastAPI, HTTPException
+from pydantic import BaseModel
+from transformers import pipeline
+
+MODEL_ID = os.getenv("MODEL_ID", "killykilly/nigerian-transaction-classifier")
+MAX_BATCH = int(os.getenv("MAX_BATCH", "256"))
+
+classifier = pipeline(
+    "text-classification",
+    model=MODEL_ID,
+    token=os.getenv("HF_TOKEN") or None,
+    truncation=True,
+)
+
+app = FastAPI(title="Nigerian Transaction Classifier API")
+
+
+class ClassifyRequest(BaseModel):
+    texts: list[str]
+
+
+@app.get("/health")
+def health():
+    return {"status": "ok", "model": MODEL_ID}
+
+
+@app.post("/classify")
+def classify(request: ClassifyRequest):
+    if len(request.texts) > MAX_BATCH:
+        raise HTTPException(413, f"batch too large (max {MAX_BATCH})")
+    if not request.texts:
+        return {"predictions": []}
+    results = classifier(request.texts, batch_size=32)
+    return {
+        "predictions": [
+            {"label": r["label"], "score": float(r["score"])} for r in results
+        ]
+    }

requirements.txt

@@ -0,0 +1,4 @@
+fastapi
+uvicorn
+transformers
+torch