DataFrame: Get the mean of two rows to a set it to a row

I have a df like,

             ds             value
====================================
1 2023-07-01 17:00:00        100
2 2023-07-01 18:00:00         0
3 2023-07-01 19:00:00        300

I want to set the mean of 1st and 3rd rows value into the 2nd row like this

             ds             value
====================================
1 2023-07-01 17:00:00       100
2 2023-07-01 18:00:00       150
3 2023-07-01 19:00:00       200

And some times it may be more that one rows which will be having 0, is there a way to handle that dynamically.

Any help is much appreciated

Thanks,

>Solution :

IIUC use Series.replace with Series.interpolate:

df['value'] = df['value'].replace(0, np.nan).interpolate()