Styjun

I am trying to extract definitions and the text that goes along with that definition.

The definition extraction is relatively easy because they are typically in parentheses and in quotes, so I can run a Regular Expression to find them.

The part I am having trouble with is getting the text that goes along with the definition- it will typically show up before the definition but I don't know how many words will be part of it.

For example:

"(a) The Company has the requisite corporate power and authority to execute and deliver this Agreement, to perform its obligations hereunder and, subject to receipt of the Requisite Company Vote, to consummate the Transactions. The execution, delivery and performance by the Company of this Agreement and the consummation by the Company of the Transactions have been duly authorized by the Company Board and no other corporate action on the part of the Company is necessary to authorize the execution and delivery by the Company of this Agreement, the Plan of Merger and the consummation by it of the Transactions, in each case, subject only to the authorization and approval of this Agreement, the Plan of Merger and the Transactions by way of (i) a shareholders’ special resolution by the affirmative vote of holders of Shares representing at least two-thirds of the voting power of the Shares present and voting in person or by proxy as a single class at the Shareholders’ Meeting, (ii) a shareholders’ resolution by the affirmative vote of holders of Shares representing a majority of the aggregate voting power of the outstanding Shares of the Company and (iii) a shareholders’ resolution by the affirmative vote of holders of a majority of the total outstanding Class A Shares (collectively clauses (i), (ii) and (iii), the “ Requisite Company Vote ”), in each case, in accordance with Section 233(6) of the CICL and the Ninth Amended and Restated Articles of Association of the Company, adopted by special resolution on December 28, 2015 (the “ Company Articles ”). This Agreement has been duly and validly executed and delivered by the Company and, assuming the due authorization, execution and delivery by Parent and Merger Sub, constitutes a legal, valid and binding obligation of the Company, enforceable against the Company in accordance with its terms, subject to bankruptcy, insolvency, fraudulent transfer, reorganization, moratorium and similar Laws of general applicability relating to or affecting creditors’ rights and to general principles of equity (the “ Bankruptcy and Equity Exception ”)."

has three definition we extract- "Requisite Company Vote", "Company Articles" and "Bankruptcy and Equity Exception".

"Requisite Company Vote" looks like the definition should be "collectively clauses (i), (ii) and (iii)" but really should also probably add "of subsection (a)" or I should actually grab the text of clauses (i), (ii), and (iii)

"Company Articles" looks like it should be "Ninth Amended and Restated Articles of Association of the Company, adopted by special resolution on December 28, 2015"

"Bankruptcy and Equity Exception" looks like it should be "bankruptcy, insolvency, fraudulent transfer, reorganization, moratorium and similar Laws of general applicability relating to or affecting creditors’ rights and to general principles of equity"

These are just three examples and I have to build something dynamic enough to be able to handle different types of definitions.

I started by building Noun Phrases using NLTK and a grammar but the NPs don't seem to be capturing everything I need even with testing various different grammars. I thought about using edit distance measurements- potentially by building the string before word by word and testing distance but all things I can think of seem to have their own flaws so I am looking for ideas.