GeneralMills/pytrends — reverse-engineered prompt

Build me a small Python library that makes Google Trends easy to query without using an official API. I want to be able to connect once, set things like language, timezone, timeout, retries, and optional HTTPS proxies, then ask for trend data with simple method calls.

It should let me look up interest over time for one or more keywords, including hourly historical data when possible, compare multiple date ranges, see interest by region, get related topics and related queries, pull trending and realtime searches, top charts, and keyword suggestions. I also want to filter by country or region, category, date range, and Google property like web, news, images, or YouTube.

Please make the results easy to work with in pandas dataframes, add a few clear examples, and include basic tests so it is not too fragile. If Google rate limits or changes behavior, handle that as gracefully as you can with retries and decent errors. Look up current Google Trends behavior online if you need to.