Wow Kent, what a great start!
I found this
http://mail.python.org/pipermail/python-list/2006-April/376149.html which lays
out some patterns of legal citations ie.
1. Two names, consisting of one or more words, separated by a "v."
2. One, two, or three citations, each of which has a volume number ("90")
followed by a Reporter name ("U.S." or "S.Ct." or "L.Ed."), which consists of
one or two words always ending with a ".", followed by a page number ("1893")
3. Each citation may contain a comma and a second page number (", 234 ")
4. Optionally, a parenthesized year ("(1970)") or optional information in
parentheses ("(DCMD Ala.1966)")
5. An ending "."
Some things I've noticed include:
* A sequence of DIFFERENT citations are separated by a ';' as in Carter v. Jury
Commission of Greene County, 396 U.S. 320, 90 S.Ct. 518, 24 L.Ed.2d 549 (1970);
Lathe Turner v. Fouche, 396 U.S. 346, 90 S.Ct. 532, 24 L.Ed.2d 567 (1970);
White v. Crook, 251 F.Supp. 401 (DCMD Ala.1966) with the first separation
containing the names (separated by a v.)
* A sequence of SIMILAR citations are separated by a ',' as in John Doggone
Williams v. Florida, 399 U.S. 78, 90 S.Ct. 1893, 234, 26 L.Ed.2d 446 (1970)
with the first separation containing the names (separated by a v.)
I was pondering the same issue about names ie. how do you know that "Page 500"
is not part of "Carter". My thought was to start from the "v.", step backwards
a word at a time, assume that the first name is valid, for all subsequent words
check if the last character of a word contained the digits [0-9] or these
punctuation marks [.,:;], if so, then it was unlikely to be part of the name.
I've changed the sample text to include examples of multiple page references:
text = "Page 500 Carter v. Jury Commission of Greene County, 396 U.S. 320 876,
90 S.Ct. 518, 24 L.Ed.2d 549 (1970); Lathe Turner v. Fouche, 396 U.S. 346, 90
S.Ct. 532, 24 L.Ed.2d 567 (1970); White v. Crook, 251 F.Supp. 401 (DCMD
Ala.1966). Moreover, the Court has also recognized that the exclusion of a
discernible class from jury service injures not only those defendants who
belong to the excluded class, but other defendants as well, in that it destroys
the possibility that the jury will reflect a representative cross section of
the community. In John Doggone Williams v. Florida, 399 U.S. 78, 90 S.Ct. 1893,
234, 26 L.Ed.2d 446 159-60 (1970), we sought to delineate some of the essential
features of the jury that is guaranteed, in certain circumstances, by the Sixth
Amendment. We concluded that it comprehends, inter alia, 'a fair possibility
for obtaining a representative cross-section of the community.' 399 U.S., at
100, 90 S.Ct., at 1906.9 Thus if the Sixth Amendment were applicable here, and
petitioner were challenging a post-Duncan petit jury, he would clearly have
standing to challenge the systematic exclusion of any identifiable group from
jury service."
Okay, I'd better get to grips with pyparsing!
Dinesh
From: Kent Johnson
Sent: Saturday, February 07, 2009 6:21 AM
To: Dinesh B Vadhia
Cc: tutor@python.org
Subject: Re: [Tutor] Picking up citations
On Sat, Feb 7, 2009 at 1:11 AM, Dinesh B Vadhia
<dineshbvad...@hotmail.com> wrote:
> Hi! I want to process text that contains citations, in this case in legal
> documents, and pull-out each individual citation. Here is a sample text:
<snip>
> The results required are:
>
> Carter v. Jury Commission of Greene County, 396 U.S. 320 (1970)
> Carter v. Jury Commission of Greene County, 90 S.Ct. 518 (1970)
> Carter v. Jury Commission of Greene County, 24 L.Ed.2d 549 (1970)
>
> Lathe Turner v. Fouche, 396 U.S. 346 (1970)
> Lathe Turner v. Fouche, 90 S.Ct. 532 (1970)
> Lathe Turner v. Fouche, 24 L.Ed.2d 567 (1970)
>
> White v. Crook, 251 F.Supp. 401 (DCMD Ala.1966)
>
> John Doggone Williams v. Florida, 399 U.S. 78 (1970)
> John Doggone Williams v. Florida, 90 S.Ct. 1893, 234 (1970)
> John Doggone Williams v. Florida, 26 L.Ed.2d 446 (1970)
Here is a close solution using pyparsing. It only gets the last word
of the first name, and it doesn't handle multiple page numbers so it
missing J. D. Williams entirely. The name is hard - how do you know
that "Page 500" is not part of "Carter" and "In" is not part of "John
Doggone Williams"? The page numbers seem possible in theory but I
don't know how to get pyparsing to do it.
from pprint import pprint as pp
from pyparsing import *
text = "" # your text
Name1 = Word(alphas).setResultsName('name1')
Name2 = Combine(OneOrMore(Word(alphas)), joinString=' ',
adjacent=False).setResultsName('name2')
Volume = Word(nums).setResultsName('volume')
Reporter = Word(alphas, alphanums+".").setResultsName('reporter')
Page = Word(nums).setResultsName('page')
VolumeCitation = (Volume + Reporter +
Page).setResultsName('volume_citation', listAllMatches=True)
VolumeCitations = delimitedList(VolumeCitation)
Date = (Suppress('(') +
Combine(CharsNotIn(')')).setResultsName('date') + Suppress(')'))
FullCitation = Name1 + Suppress('v.') + Name2 + Suppress(',') +
VolumeCitations + Date
for item in FullCitation.scanString(text):
fc = item[0]
# Uncomment the following to see the raw parse results
# pp(fc)
# print
# print fc.name1
# print fc.name2
# for vc in fc.volume_citation:
# pp(vc)
for vc in fc.volume_citation:
print '%s v. %s, %s %s %s (%s)' % (fc.name1, fc.name2,
vc.volume, vc.reporter, vc.page, fc.date)
print
The output is:
Carter v. Jury Commission of Greene County, 396 U.S. 320 (1970)
Carter v. Jury Commission of Greene County, 90 S.Ct. 518 (1970)
Carter v. Jury Commission of Greene County, 24 L.Ed.2d 549 (1970)
Turner v. Fouche, 396 U.S. 346 (1970)
Turner v. Fouche, 90 S.Ct. 532 (1970)
Turner v. Fouche, 24 L.Ed.2d 567 (1970)
White v. Crook, 251 F.Supp. 401 (DCMD Ala.1966)
Kent
_______________________________________________
Tutor maillist - Tutor@python.org
http://mail.python.org/mailman/listinfo/tutor
