dockerfile updates

Dockerfile

@@ -2,41 +2,36 @@ FROM python:3.11-slim
 
 WORKDIR /app
 
-# Set Hugging Face cache directory to a build-in writeable directory
 ENV HF_HOME=/app/.cache/huggingface
-# Tell NLTK where to find downloaded data
 ENV NLTK_DATA=/app/nltk_data
 
-# Install system deps for numpy/umap
+# Install system deps for numpy/umap/llama-cpp
 RUN apt-get update && apt-get install -y --no-install-recommends \
-    build-essential gcc && \
+    build-essential gcc g++ && \
     rm -rf /var/lib/apt/lists/*
 
-# Copy and install Python deps
 COPY pyproject.toml .
 RUN pip install --no-cache-dir .
 
-# Pre-download NLTK data at build time to the app-owned directory
+# Pre-download NLTK data
 RUN mkdir -p /app/nltk_data && \
     python -c "import nltk; nltk.download('punkt', download_dir='/app/nltk_data'); nltk.download('punkt_tab', download_dir='/app/nltk_data'); nltk.download('stopwords', download_dir='/app/nltk_data')"
 
-# Pre-download public HF embedding models at build time (Nomic and Qwen)
-# Note: we exclude the gated google/gemma-3n-embedding-exp to prevent build failures.
+# Pre-download public HF embedding models
 RUN python -c "from huggingface_hub import snapshot_download; \
     snapshot_download(repo_id='nomic-ai/nomic-embed-text-v1.5'); \
     snapshot_download(repo_id='Qwen/Qwen3-Embedding-0.6B')"
 
-# Pre-download public HF Qwen LLM model at build time
-RUN python -c "from huggingface_hub import snapshot_download; \
-    snapshot_download(repo_id='Qwen/Qwen2.5-1.5B-Instruct')"
+# Download lightweight GGUF model instead of raw safetensors
+RUN mkdir -p /app/models && \
+    python -c "from huggingface_hub import hf_hub_download; \
+    hf_hub_download(repo_id='Qwen/Qwen2.5-1.5B-Instruct-GGUF', filename='qwen2.5-1.5b-instruct-q4_k_m.gguf', local_dir='/app/models')"
 
-# Copy app code
 COPY . .
 
-# Grant full read/write/execute permissions on /app for non-root containers in Hugging Face
 RUN chmod -R 777 /app
 
-# HF Spaces expects port 7860
 EXPOSE 7860
 
-CMD ["uvicorn", "backend.main:app", "--host", "0.0.0.0", "--port", "7860"]
+# Force unbuffered Python output so you can actually see error logs in HF Spaces
+CMD ["python", "-u", "-m", "uvicorn", "backend.main:app", "--host", "0.0.0.0", "--port", "7860"]

README.md

@@ -6,11 +6,11 @@ colorTo: red
 sdk: docker
 app_port: 7860
 pinned: false
+startup_duration_timeout: 1h
 ---
 
 # 🔬 RAG Visualizer
 
-
 **An X-Ray machine for Retrieval-Augmented Generation pipelines.**
 
 RAG Visualizer is an interactive, local-first tool that lets you **see** what happens inside a RAG pipeline — from how your text gets chunked, to how those chunks land in vector space, to which chunks get retrieved for a given query. No cloud APIs, no black boxes. Everything runs on your machine with local Ollama models.
@@ -51,7 +51,7 @@ Visualize and compare **5 chunking strategies** side-by-side:
 - **Sonar Query Simulator** — Type a natural language query and watch the sonar ping animate across the canvas in real time.
 - **Sonar Probe** — Click anywhere on the 2D canvas to retrieve the nearest chunks in that region.
 - **Document X-Ray Highlighting** — Retrieved chunks glow dynamically in the document viewer with rank-based styling (gold for Rank 1, dashed for Rank 2, dotted for Rank 3).
-- **Metadata Level Filtering** — Filter your context pool on the fly (retrieve *Only Parents*, *Only Children*, or *All Levels*).
+- **Metadata Level Filtering** — Filter your context pool on the fly (retrieve _Only Parents_, _Only Children_, or _All Levels_).
 - **Cross-Encoder Reranking** — Run a local FlashRank (`ms-marco-MiniLM-L-12-v2`) engine to rerank search results.
 - **Rank Shift Badges** — Visual indicators showing exactly how much chunks moved after reranking (`▲ +3`, `▼ -1`, or `• Unchanged`).
 - **Normalized Match Strength** — Converts raw vector distances into intuitive similarity percentages (e.g. `Match: 87.7%`).