Fetch paginated records and filter them by rules and a historical average

An HTTP endpoint serves patients' medical records, ten per page, with no page count given up front. Each record has many fields, most of them irrelevant to you.

1. Fetch every page until the endpoint runs out.
2. Return the records that pass a list of conditions the interviewer gives you. Most are simple comparisons on a field; one compares a record's weight with that patient's average weight over their history, which you compute from the data.

The logic is easy and the specification is long: you have 45 minutes and are expected to finish working code. Say what you do with missing or badly typed values, and ask which records count as a patient's history.