Fix instructions, upgrade OpenAI API to v 1 or higher and add support to open source model backend

• Some small missing elements in the README have been added.
• The Redmine configuraton example has been corrected (previously it was a Zulip configuration instead).
• The OpenAI configuration example for an open source model has been added.
• The code in redmine_summarize.py has been adapted to work with the OpenAI API version > 1.0.0, which is needed to set the base URL of the Open AI server.
• A new option has been added in the OpenAI configuration (url) to set the base URL of OpenAI.

Note: This seems to work with Vicuna 13B 1.5 (setup with --load-8bit to reduce memory usage) for most tickets, but some tickets are larger and an OOM error is issued in a P100 GPU machine. For this reason, using Vicuna 7B 1.5 needs to be used there. One could also get an A100 machine, in which probably we could use Vicuna 13B 1.5 or even its 16K variant.