CogStack/MedCAT ? reverse-engineered prompt

Build me a Python library for reading clinical notes and pulling out medical concepts from the text. I want it to find things like diagnoses, symptoms, and medications, then link them to SNOMED CT or UMLS terms, and show the matched phrases back in the original note so I can review them easily.

It should feel like a practical tool for EHR text, with a simple way to load a model, run annotation on a batch of notes, and save the results. If possible, include basic support for negation and a few example notebooks or demo scripts so I can see it working right away. Keep the code clean, with tests and docs, and make it easy to install and use from a normal Python project. Look up the current docs online if you need to, and match the project style closely.