improve speed/efficiency of dataframe_from_inp function

[aerispaha]

Currently, the dataframe_from_inp (and similar functions) scan the text file twice to extract the content. First, by running the get_inp_sections_details function, then by scanning to the section of interest.

It might be possible to avoid the first scan through the of text file by using a pre-populated list of end_strings in the extract_section_of_file function call, thereby possibly removing the need to run get_inp_sections_details. This would be useful when dealing with large files and/or over a network.