add functionality for processing based on presence of margin attribute

[tcatapano]

sample query: find all occurrences of "gold" when it occurs in the left-middle margin

[matthewkumar]

in the most recent push the master, the method search_margins() was added to the BnF class. Using this example,

manuscript=BnF()
manuscript.search_margins('tl', 'gold', 'left-middle')

yields the following:

['004v_1', '004v_3', '016r_1', '024v_1', '036v_3', '039r_2', '057v_1', '072r_2', '081v_4', '085v_4', '094r_1', '098r_2', '098r_3', '104r_2', '110v_2', '112v_1', '118v_1', '126v_3', '139r_1', '141r_1', '144r_1', '145v_1', '151v_1', '152r_1', '155v_1', '156v_1', '169r_1']

[matthewkumar]

This function was updated to convert all text to lowercase, notice that there are a couple new entries in the updated output (after removing duplicates):

['', '010r_1', '029v_6', '032v_3', '100v_2', '106r_2', '111v_4', '116v_2', '120r_2', '120v_6', '121v_1', '121v_2', '123r_1', '124v_2', '126v_1', '128r_1', '128v_1', '131v_1', '134v_2', '135r_1', '135v_1', '139r_1', '145v_2', '152v_1', '154r_2', '154r_3', '156v_1']

I noticed that some terms, like 'goldsmith' will register positively since 'gold' is a substring. There are three possible ways to deal with it.

1. do nothing, and this case will be manually ignored by scholars upon close analysis
2. use Regex to specify that a space or punctuation must surround the term in question
3. specify property type in the function call.

I recommend the first, but all are possible and easily implemented.

[gschare]

I can add a new function to entry.py to do this with lxml.etree.