friendly_brief parses brief titles from a CSV file and emits a new CSV file
with some inferences about the brief. A brief title might look like this.
1. Amicus Brief, BRIEF OF L. S. LEE, INC. AMICUS CURIAE ON BEHALF OF PETITIONER, December 6, 2000, 2000 U.S. S. Ct. Briefs LEXIS 836
For such a title, friendly_brief tries to guess the
* Brief number
* Amici curiae
* Posture of the amici
How to use
--------------
Install it from pip. ::
pip3 install friendly_brief
And run it on a CSV file. The file must contain a column with all of the
brief titles that you care about, and the titles must be in a column called
"brief". The CSV file can have anything else you want in it too. ::
friendly-brief briefs.csv
It can also receive a CSV file over STDIN. ::
cat briefs.csv | friendly-brief
The resulting CSV file is written to STDOUT.
How it works
---------------
Let's discuss how each of the inferences is made.
Brief number
~~~~~~~~~~~~~
We take the first unbroken group of digits as the brief number.
For example, the following brief title starts with a "1", then a "9",
and then a ".".
19. Brief, BRIEF AMICUS CURIAE OF SOCIAL SCIENCE AND COMPARATIVE LAW SCHOLARS IN SUPPORT OF NEITHER PARTY, June 1, 2001, 2001 U.S. S. Ct. Briefs LEXIS 718
We stop upon noticing the non-digit "." and use "19" as the brief number.
Posture
~~~~~~~~~~~~
Posture is guessed based on the presence of certain phrases.
There are five types of posture, and here are their corresponding phrases
Posture 0
"Neither party"
Posture 1
"Petitioner", "Appellant", and "Reversal"
Posture 2
"Respondent", "Appellee", "Affirmance"
Posture 3
"Plaintiff"
Posture 4
"Defendant"
The program looks for the presence of all of these phrases.
If the result is unambiguous, the resulting spreadsheet contains
the number corresponding to the posture.
Ambiguity can occur if no posture phrases are present or if
phrases corresponding to different postures are present.
For example, I would consider a brief title containing both
"plaintiff" and "defendant" to be ambiguous. In cases of
ambiguouity, the posture cell is left blank.
Amici
~~~~~~~~~
The messiest part of this whole process is the guessing of the amici.
I don't even know what it's doing, but here are some of the concepts.
Pretty early on, the date and everything after get removed. For example, this
4. Amicus Brief, BRIEF OF SOCIAL AND ORGANIZATIONAL PSYCHOLOGISTS AS AMICI CURIAE SUPPORTING RESPONDENTS, August 13, 2012, 2012 U.S. S. Ct. Briefs LEXIS 3223
becomes this.
4. Amicus Brief, BRIEF OF SOCIAL AND ORGANIZATIONAL PSYCHOLOGISTS AS AMICI CURIAE SUPPORTING RESPONDENTS
The brief title is split into pieces at things like commas, semicolons
and the word "and", so we wind up with something like this.
* 4. Amicus Brief
* BRIEF OF SOCIAL AND ORGANIZATIONAL PSYCHOLOGISTS AS AMICI CURIAE SUPPORTING RESPONDENTS
Things that don't look like the names of amici get removed. This includes
words like "amici", "amicus", "supporting", "as", and "brief", and we wind
up with the following amicus.
SOCIAL AND ORGANIZATIONAL PSYCHOLOGISTS
Lots of weird things are done to deal with suffixes ("INC", "LLC", "JR"),
non-serial commas ("first, second and third"), strange character encodings,
misspellings, and other typos in the brief title.
