Fix test_vasopressor_units crashing on Series.contains

[Chessing234]

Bug

`test_vasopressor_units` in `mimic-iv/tests/test_medication.py` crashes with `AttributeError` whenever the preceding `gbq.read_gbq` query returns any rows, before ever reaching its real assertion:

```
AttributeError: 'Series' object has no attribute 'contains'
```

Root cause

The test pulls rows with a non-standard `rateuom` for each vasopressor and then checks whether any row's `hadm_id` falls outside a whitelist of known-inspected `hadm_id`s:

```python
df = gbq.read_gbq(query, project_id=project_id, dialect="standard")

if we find new uninspected rows, raise a warning. this will only happen when mimic-iv is updated.

if (~df['hadm_id'].contains(hadm_id_list)).any():
_LOGGER.warn(...)
```

`pandas.Series` has no `.contains()` method — `contains` only lives on the `.str` accessor and does substring-matching against a single pattern, not membership-testing against a list. The correct pandas idiom for "row's value is in this list" is `Series.isin(...)`, which is clearly what the surrounding code intends (hadm_id_list is a plain list of integers, not a regex).

As written, the moment the query returns any row the test hits the `.contains(hadm_id_list)` call, raises `AttributeError`, and fails. The hit path is what the test is supposed to log a warning for, so the bug masks the whole reason this check exists.

Fix

Replace `.contains(hadm_id_list)` with `.isin(hadm_id_list)`:

```diff

•    if (~df['hadm_id'].contains(hadm_id_list)).any():
  

•    if (~df['hadm_id'].isin(hadm_id_list)).any():
  

```

The outer `~...any()` still expresses "any row falls outside the whitelist", matching the comment and the downstream warning log.