d

.gitignore

@@ -0,0 +1,34 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*.so
+
+# Virtual environment
+venv/
+env/
+.venv/
+
+# Build
+build/
+dist/
+
+# Jupyter Notebook checkpoints
+.ipynb_checkpoints/
+
+# Logs and local environment files
+*.log
+*.env
+.env.local
+
+# PyTorch or TensorFlow saved models
+*.pt
+*.pth
+*.h5
+
+# VSCode settings (if using VSCode)
+.vscode/
+
+# Hugging Face cache (optional)
+/content/huggingface/
+# dataset/
+discord/

Dockerfile

@@ -0,0 +1,21 @@
+FROM python:3.10-slim
+
+RUN apt-get update && apt-get install -y git git-lfs wget unzip && rm -rf /var/lib/apt/lists/*
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+WORKDIR /app
+COPY app.py .
+COPY dataset ./dataset
+
+# Hugging Face cache fix
+ENV TRANSFORMERS_CACHE=/app/models/.cache
+
+# Clone models
+RUN git lfs install && \
+    git clone https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2 /app/models/all-MiniLM-L6-v2 && \
+    git clone https://huggingface.co/facebook/opt-1.3b /app/models/facebook-opt-1.3b && \
+    git clone https://huggingface.co/facebook/bart-large-cnn /app/models/bart-large-cnn
+
+EXPOSE 7860
+CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]

app.py

@@ -0,0 +1,109 @@
+import os
+import json
+from fastapi import FastAPI
+from fastapi.middleware.cors import CORSMiddleware
+from pydantic import BaseModel
+from typing import List
+from transformers import pipeline
+from sentence_transformers import SentenceTransformer
+import faiss
+import gradio as gr
+from gradio import mount_gradio_app
+
+# ------------------- Config ------------------- #
+DATA_PATH = "./dataset/pentagon_core.json"
+EMBEDDING_MODEL = "sentence-transformers/all-MiniLM-L6-v2"
+QA_MODEL = "facebook/bart-large-cnn"
+DEVICE = "cuda" if os.environ.get("USE_CUDA") == "1" else "cpu"
+
+# ------------------- Load Models ------------------- #
+embedder = SentenceTransformer(EMBEDDING_MODEL)
+qa_model = pipeline("text2text-generation", model=QA_MODEL, device=0 if DEVICE == "cuda" else -1)
+
+# ------------------- Load Dataset + Index ------------------- #
+if os.path.exists(DATA_PATH):
+    with open(DATA_PATH, "r") as f:
+        knowledge_base = json.load(f)
+else:
+    knowledge_base = []
+
+texts = [item["content"] for item in knowledge_base]
+embeddings = embedder.encode(texts, convert_to_tensor=True)
+index = faiss.IndexFlatL2(embeddings.shape[1])
+index.add(embeddings.cpu().detach().numpy())
+
+# ------------------- FastAPI App ------------------- #
+app = FastAPI()
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],  # For development
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+# --------- Upload Endpoint --------- #
+class UploadData(BaseModel):
+    content: str
+
+@app.post("/upload/")
+def upload_knowledge(data: UploadData):
+    global knowledge_base, index
+
+    knowledge_base.append({"content": data.content})
+    with open(DATA_PATH, "w") as f:
+        json.dump(knowledge_base, f, indent=2)
+
+    new_embedding = embedder.encode([data.content], convert_to_numpy=True)
+    index.add(new_embedding)
+
+    return {"message": "Data uploaded and indexed."}
+
+# --------- Ask Endpoint --------- #
+@app.get("/ask/")
+def ask(question: str, top_k: int = 3):
+    question_embedding = embedder.encode([question], convert_to_numpy=True)
+    distances, indices = index.search(question_embedding, top_k)
+    context = " ".join([knowledge_base[i]["content"] for i in indices[0]])
+
+    prompt = (
+        f"Context: {context}\n\n"
+        f"Answer the following question based only on the above context:\n"
+        f"{question}\n\nAnswer:"
+    )
+    output = qa_model(prompt, max_length=256, do_sample=False)[0]["generated_text"]
+
+    return {
+        "question": question,
+        "context_used": context,
+        "answer": output.strip()
+    }
+
+# --------- Gradio UI --------- #
+def gradio_upload(file):
+    if file is None:
+        return "No file selected."
+
+    try:
+        content = file.read().decode("utf-8")
+        import requests
+
+        base_url = os.getenv("HF_SPACE_URL", "http://localhost:7860")
+        response = requests.post(f"{base_url}/upload/", json={"content": content})
+
+        if response.status_code == 200:
+            return "✅ Data successfully uploaded and indexed!"
+        else:
+            return f"❌ Failed: {response.text}"
+    except Exception as e:
+        return f"❌ Error: {str(e)}"
+
+gr_app = gr.Interface(
+    fn=gradio_upload,
+    inputs=gr.File(label="Upload .txt or .json file"),
+    outputs="text",
+    title="Upload Knowledge",
+)
+
+# Mount Gradio at /ui
+app = mount_gradio_app(app, gr_app, path="/ui")

dataset/pentagon_core.json

@@ -0,0 +1,8 @@
+[
+    {
+        "content": "This is the first knowledge piece."
+    },
+    {
+        "content": "This is the second knowledge piece."
+    }
+]

requirements.txt

@@ -0,0 +1,8 @@
+fastapi
+uvicorn
+gradio
+transformers
+sentence-transformers
+faiss-cpu
+torch
+python-pptx

space.yaml

@@ -0,0 +1,3 @@
+title: Test-api Space
+sdk: docker
+app_port: 7860