Spaces: docker SDK + clean README (UTF-8 no BOM, LF)

.gitattributes

@@ -1,2 +1,2 @@
-* text=auto
-*.md text
+* text=auto eol=lf
+README.md text eol=lf

README.md

@@ -1,32 +1,14 @@
 ---
 title: FastAPI RAG API
-emoji: 🧠
-colorFrom: indigo
-colorTo: blue
-sdk: fastapi
-app_file: app/api.py
+sdk: docker
 pinned: false
 ---
 
 # FastAPI RAG API
 
 Minimal backend for RAG (FastAPI + FAISS + Sentence-Transformers).
-Open `/docs` for the interactive API.
-
-## Live
-- API (HF Space): https://huggingface.co/spaces/HamidOmarov/FastAPI-RAG-API
-- Dashboard (HF Space): https://huggingface.co/spaces/HamidOmarov/RAG-Dashboard
-
-## What it does
-Ask questions about your PDFs using vector search (FAISS) + sentence embeddings.
-Robust to numeric/table-heavy docs, with optional AZ→EN translation and fallbacks.
+Open /docs for the interactive API.
 
 ## Quick test
 curl -F "[email protected]" https://<API>/upload_pdf
-curl -H "Content-Type: application/json" -d '{"question":"What is the document about?","top_k":5}' https://<API>/ask_question
-
-## Ops
-- GET /health • GET /stats • GET /get_history • POST /reset_index
-
-## Stack
-FastAPI · sentence-transformers · FAISS · pypdf · Hugging Face Spaces
+curl -H "Content-Type: application/json" -d "{\"question\":\"What is the document about?\",\"top_k\":5}" https://<API>/ask_question

requirements.txt

@@ -8,3 +8,5 @@ sentencepiece
 sacremoses
 python-multipart
 pdfminer.six
+numpy
+requests