Wrong exception catch when extracted tables don't exist about cosmos

[ryansun117]

I did some debugging following the changes, it seems both TypeError and AttributeError are possible. At line 276, there could be two cases for error since df is a list of dataframes:

1. If df == [None, None], meaning either Camelot or Pdfplumber successfully extracted a table, but one or both of the tables was None. Then, attempting to access None.to_csv() would throw AttibuteError: 'NoneType' object has no attribute 'to_csv'
2. If df == None, meaning both Camelot and Pdfplumber extraction failed. Then, attempting to access None[0] would throw TypeError: 'NoneType' object is not subscriptable

A solution would be to catch both TypeError and AttributeError. This ensures the process continues to run. Should we add more detailed logging statements for each type of exception?