refactor: update branding, titles, and architecture to VECTOR Multilingual RAG Engine

README.md

@@ -1,16 +1,16 @@
 ---
-title: Hacker House Goa 2026 - Voice Indic RAG
-emoji: 🌴
+title: VECTOR - Voice Indic RAG
+emoji: ⚡
 colorFrom: green
-colorTo: yellow
+colorTo: blue
 sdk: docker
 app_port: 7860
 pinned: false
 license: mit
-short_description: Voice Indic RAG with Sub-10ms FAISS Retrieval
+short_description: VECTOR Voice Indic RAG with Sub-10ms FAISS Retrieval
 ---
 
-# 🌴 Hacker House Goa 2026: Voice-Enabled Multilingual Indic RAG
+# ⚡ VECTOR: Voice-Enabled Multilingual Indic RAG Engine
 
 <div align="center">
 
@@ -31,7 +31,7 @@ short_description: Voice Indic RAG with Sub-10ms FAISS Retrieval
 
 ## 📌 Executive Summary
 
-Active runtime deployment is optimized for **3 core languages** (**English [`en`]**, **Hindi [`hi`]**, and **Marathi [`mr`]**) loading **148,854 in-memory vectors** (148,545 native passage vectors + 309 semantic longdoc vectors), achieving **~7.04 ms retrieval latency** (p95: 7.97 ms vs 50.0 ms budget). 
+**VECTOR** is optimized for low-latency, voice-enabled Retrieval-Augmented Generation across Indian languages. Active runtime deployment is optimized for **3 core languages** (**English [`en`]**, **Hindi [`hi`]**, and **Marathi [`mr`]**) loading **148,854 in-memory vectors** (148,545 native passage vectors + 309 semantic longdoc vectors), achieving **~7.04 ms retrieval latency** (p95: 7.97 ms vs 50.0 ms budget). 
 
 The system provides zero-code extensibility across **all 14 Indic languages** (*Assamese, Bengali, Gujarati, Hindi, Kannada, Malayalam, Marathi, Nepali, Odia, Punjabi, Sanskrit, Tamil, Telugu, Urdu*) plus **English** (15 languages total, **~743,000 deduplicated passages**) via a single configuration entry (`config.LANGUAGES`).
 
@@ -40,7 +40,7 @@ The system provides zero-code extensibility across **all 14 Indic languages** (*
 - 🛡️ **Cascaded 4-Tier Guardrails**: Stem regex, Meta Prompt-Guard 86M neural DPI/IPI shield, 6-class intent filter, and own-language centroid weighting.
 - 🔀 **Script-Aware BM25 + Dense Fusion**: Automatic cross-script detection bypassing BM25 penalties for cross-lingual queries.
 - 🧮 **Deterministic Synthesis**: TextRank graph centrality + SVD singular energy context ranking (zero LLM API latency or cost).
-- 🌴 **Command Center UI**: Retro-tropical Web Audio waveform interface with stage-by-stage telemetry waterfall.
+- 🌴 **VECTOR Command Center UI**: Retro-tropical Web Audio waveform interface with stage-by-stage telemetry waterfall.
 
 ---
 
@@ -48,31 +48,31 @@ The system provides zero-code extensibility across **all 14 Indic languages** (*
 
 ```mermaid
 graph TD
-    A[Spoken Voice Audio / Text Bypass] --> B[Sarvam Saaras v3 STT + ffmpeg 16kHz Normalizer]
-    B --> C[Language Resolution: config.LANGUAGES Router]
-    C --> D[Guardrail 1: Tier-1 Fast Regex + Safety Patterns]
-    D -- Safe --> PG[Guardrail 2: Meta Prompt-Guard 86M Neural DPI Shield]
+    A[Spoken Voice Audio / Text Bypass] --> B[Sarvam Saaras STT Engine + ffmpeg 16kHz Normalizer]
+    B --> C[VECTOR Language Resolution Router: config.LANGUAGES]
+    C --> D[VECTOR Guardrail 1: Tier-1 Fast Stem Regex + Safety Patterns]
+    D -- Safe --> PG[VECTOR Guardrail 2: Meta Prompt-Guard 86M Neural DPI Shield]
     D -- Blocked --> X[Declined Response: Safety Violation]
-    PG -- Safe --> IF[Guardrail 3: Pre-Retrieval Query Intent Filter]
+    PG -- Safe --> IF[VECTOR Guardrail 3: Pre-Retrieval Query Intent Filter]
     PG -- Injected --> X
     IF -- Factual --> E[Query Embedding: 'query: ' Prefix multilingual-e5-small INT8]
     IF -- Non-Factual Intent --> X
-    E --> F[Guardrail 4: Centroid Distance Off-Topic Filter]
+    E --> F[VECTOR Guardrail 4: Centroid Distance Off-Topic Filter]
     F -- Off-Topic --> X
     F -- On-Topic --> CACHE{Dynamic Vector & Gold QA Cache}
     CACHE -- Cache Hit <0.5ms --> N[Grounded Response + Zero-Latency Fast Path]
-    CACHE -- Cache Miss --> G[Parallel Multi-Strategy FAISS HNSW Retrieval]
+    CACHE -- Cache Miss --> G[VECTOR Parallel Multi-Strategy FAISS HNSW Engine]
     G --> H1[Passage Native Index: 148,545 Vectors]
     G --> H2[Semantic LongDoc Index: 309 Vectors]
     H1 --> I[Candidate Merge & Reciprocal Rank Fusion RRF k=60]
     H2 --> I
-    I --> J[Adaptive Script-Aware BM25 Score Fusion]
+    I --> J[VECTOR Adaptive Script-Aware BM25 Score Fusion]
     J --> K[Relevance & Disqualification Gate: Dense / CE Threshold]
     K -- Score < Threshold --> Y[Declined Response: No Relevant Info in Corpus]
     K -- High Relevance --> CS[Context Chunk Safety: Batched Prompt-Guard 86M IPI Scan]
     CS -- Poisoned Chunks --> X
-    CS -- Clean Chunks --> L[Deterministic Non-LLM Context Synthesis: TextRank + SVD Energy]
-    L --> M[Post-Generation Grounding & Hallucination Guardrail]
+    CS -- Clean Chunks --> L[VECTOR Deterministic Context Synthesis: TextRank + SVD Energy]
+    L --> M[VECTOR Post-Generation Grounding & Hallucination Guardrail]
     M -- Grounded --> N[Grounded JSON Response + Full 9-Stage Telemetry]
     M -- Insufficient Info --> Y
 ```
@@ -244,19 +244,19 @@ pytest tests/ -v
 
 ```
 VECTOR/
-├── api/                  # FastAPI web server (/query, /health, /languages)
+├── api/                  # VECTOR FastAPI web server (/query, /health, /languages)
 ├── app/                  # Fast ONNX retriever & 50ms benchmark runner
 ├── benchmark/            # Latency, cold-start, & throughput evaluation scripts
 ├── chunking/             # Native, sentence-window, semantic, & RRF splitters
 ├── data/                 # FAISS HNSW indexes, centroids, JSONL corpora scripts
-├── demo/                 # Hacker House Goa 2026 Web Audio UI & assets
+├── demo/                 # VECTOR Web Audio UI & visual assets
 ├── generation/           # TextRank + SVD non-LLM synthesis & LLM fallback
 ├── guardrails/           # 4-tier cascaded pre-retrieval & post-gen grounding
 ├── pipeline/             # Async 9-stage pipeline state machine & schemas
 ├── retrieval/            # multilingual-e5-small INT8 ONNX & FAISS engine
-├── stt/                  # Sarvam Saaras v3 STT & ffmpeg 16kHz audio pipeline
+├── stt/                  # Sarvam Saaras STT & ffmpeg 16kHz audio pipeline
 ├── tests/                # 50/50 unit & integration test suite
-├── app.py                # Space entrypoint application
+├── app.py                # VECTOR Space entrypoint application
 ├── config.py             # Single source of truth configuration
 ├── Dockerfile            # Container definition for Hugging Face Spaces
 └── requirements.txt      # Python dependencies
@@ -264,15 +264,5 @@ VECTOR/
 
 ---
 
-## 🚀 Hugging Face Space Deployment
-
-Deployed via **Docker SDK** on Hugging Face Spaces:
-- **Live Space URL**: [https://ansh123456789-ragingoa.hf.space](https://ansh123456789-ragingoa.hf.space)
-
-> [!TIP]
-> Free `cpu-basic` Spaces sleep after 48h inactivity. Initial container spin-up takes **30–90s**. Warm runtime operates at **~7–16 ms**.
-
----
-
 ## 📜 License
-MIT License. Created for **Hacker House Goa 2026**.
+MIT License. VECTOR Multilingual RAG Engine.

api/main.py

@@ -62,7 +62,7 @@ async def lifespan(app: FastAPI):
 
 
 app = FastAPI(
-    title="Voice-Enabled Indic RAG API",
+    title="⚡ VECTOR Voice Indic RAG API",
     description="Instrumented low-latency Voice RAG pipeline for Indic languages (English, Hindi, Marathi)",
     version="1.0.0",
     lifespan=lifespan,

app.py

@@ -1,5 +1,5 @@
 """
-Hugging Face Space Application for Hacker House Goa 2026: Voice-Enabled Indic RAG.
+Hugging Face Space Application for VECTOR: Voice-Enabled Indic RAG.
 Renders the full retro-tropical Command Center UI and exposes FastAPI endpoints.
 ZeroGPU compatible.
 """
@@ -74,7 +74,7 @@ def _dummy_zerogpu():
 
 
 # Create core FastAPI application
-app = FastAPI(title="🌴 Hacker House Goa 2026 — Voice-Enabled Indic RAG")
+app = FastAPI(title="⚡ VECTOR — Voice-Enabled Indic RAG")
 
 app.add_middleware(
     CORSMiddleware,
@@ -104,7 +104,7 @@ async def serve_index() -> HTMLResponse:
     if demo_file.exists():
         with open(demo_file, "r", encoding="utf-8") as f:
             return HTMLResponse(content=f.read())
-    return HTMLResponse(content="<h1>Hacker House Goa 2026 Command Center</h1>")
+    return HTMLResponse(content="<h1>VECTOR 2026 Command Center</h1>")
 
 
 @app.get("/health", response_class=JSONResponse)
@@ -204,7 +204,7 @@ async def query_pipeline(
 
 
 # Create Gradio interface block and mount on FastAPI app
-with gr.Blocks(title="🌴 Hacker House Goa 2026 — Voice Indic RAG") as demo:
+with gr.Blocks(title="⚡ VECTOR — Voice Indic RAG") as demo:
     gr.HTML(get_custom_html())
     dummy_btn = gr.Button("zero_gpu_anchor", visible=False)
     dummy_btn.click(fn=_dummy_zerogpu)
@@ -215,5 +215,5 @@ app = gr.mount_gradio_app(app, demo, path="/gradio")
 if __name__ == "__main__":
     port = int(os.getenv("PORT", "7860"))
     host = os.getenv("HOST", "0.0.0.0")
-    print(f"[Space Startup] Starting Hacker House Goa Command Center UI on http://{host}:{port}")
+    print(f"[Space Startup] Starting VECTOR Command Center UI on http://{host}:{port}")
     uvicorn.run(app, host=host, port=port)

app/__init__.py

@@ -1,5 +1,5 @@
 """
-Hacker House Goa 2026 Indic Voice RAG Application Package.
+VECTOR Indic Voice RAG Application Package.
 """
 
 from pathlib import Path

demo/index.html

@@ -3,7 +3,7 @@
 <head>
   <meta charset="UTF-8" />
   <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-  <title>Hacker House Goa 2026 — Voice-Enabled Multilingual RAG</title>
+  <title>VECTOR 2026 — Voice-Enabled Multilingual RAG Engine</title>
   
   <!-- Fonts (Creative & Stylish Beach Pop Art & Logo Typography) -->
   <link rel="preconnect" href="https://fonts.googleapis.com">
@@ -488,7 +488,7 @@
         
         <div>
           <div class="flex items-center gap-2">
-            <span class="action-burst text-xs">HACKER HOUSE GOA 2026</span>
+            <span class="action-burst text-xs">VECTOR 2026</span>
             <span class="px-2 py-0.5 bg-black text-beachcyan font-comic text-xs border-2 border-black rounded shadow-brutal-sm">POP ART EDITION</span>
           </div>
           <h1 class="font-pop-art text-3xl md:text-4xl tracking-wide mt-0.5">VOICE INDIC RAG 🌊 SURF THE VECTORS! 🏄‍♂️</h1>

rag_eval_report.md

@@ -1,7 +1,7 @@
 # RAG Evaluation Report
 
 **Author:** Manus AI  
-**Target:** [Hacker House Goa 2026 — Voice-Enabled Multilingual RAG](https://ansh123456789-ragingoa.hf.space/) [1]  
+**Target:** VECTOR — Voice-Enabled Multilingual Indic RAG [1]  
 **Evaluation date:** 16 August 2026  
 **Evaluation type:** Black-box VIGOURLS-style RAG evaluation battery