Readings That Can't Be Right

A garden weather station saves its readings to a CSV file. Sometimes a sensor glitches and writes an impossible number, like a humidity of 140%. Some cells are also empty, where a reading was missed.

1hour,temp_c,humidity,wind_kmh
26,8.5,91,4
37,9.1,,6
48,10.4,140,5

Before cleaning anything, you want a quick look at which columns contain nonsense.

The starter code gives you write_text(path, text), which saves text into a file called path.

Task: write out_of_range_columns(path, csv_text, limits).

• Save csv_text to path with write_text, then load it with pandas.
• limits is a dictionary like {"humidity": [0, 100]}. Each key is a column name, and its value is [lowest allowed, highest allowed].
• A column is out of range if its smallest reading is below its lowest allowed value, or its largest reading is above its highest allowed value. A reading exactly on a limit is fine.
• Empty cells are missing readings. Ignore them.
• Return a list of the out-of-range column names, in the same order as the keys of limits. If every column is fine, return [].

Every column named in limits holds numbers and has at least one reading.