Hey Gil:
The QC team may give you a better answer but I'll give it a shot.
1. the function get_price_dataframe is a method for normalizing the historical data passed through the function. To do this, it uses log price as a way of creating smoother prices (ie; comparing the relationship of securities that have roughly $5 per share and the other with roughly $2500 a share is made easier when using log prices). After logging the price values, it converts all historical data to the same time zone and then drops NaN values.
2. I believe this is true for all the framework examples, not just the pearson alpha. They are only examples of what can be done, and aren't meant to be used exactly as is (although functionally you could). For example, you could test the securities for cointegration, or anything else for that matter. The pearson test is just an example of one of many tests that can be done to check for pairs