[Feature Request] Chat: show the scope of each question, and optionally show the model's reasoning

> AI disclosure: I wrote this request, and the linked implementation, with the help of an AI agent (Claude). Description Two changes to the chat overlay. 1. Make the scope explicit. The chat answers about one document or about all documents. Today the scope follows the current page. Only the input placeholder tells you which one applies. Proposal: a toggle at the top of the chat, "This document" and "All documents". On a document page, both options are available, so you can ask the whole archive while a document is open. The API already allows that. Each question in the history carries a label with the scope it was sent with. 2. Show the model's reasoning, optional. Reasoning models (ollama gpt-oss, deepseek-r1, and OpenAI-compatible backends that return reasoning_content) produce reasoning before the answer. Paperless discards it. With a local model, that reasoning is where most of the wait is. Proposal: a per-user switch, off by default, that shows th…