Name:          findstruct
Version:       0.0.1
Release:       1%{?dist}
Summary:       Java library for structured the clinical documents
# CDDL or GPLv2 with exceptions:
# ./FindStruct/src/findstruct/AbsoluteConstraints.java
# ./FindStruct/src/findstruct/AbsoluteLayout.java
# ASL 2.0 pom file
# MIT project license see FindStruct_doc.html
License:       ASL 2.0 and MIT and (CDDL or GPLv2 with exceptions)
URL:           http://groups.csail.mit.edu/medg/projects/text/findstruct/
Source0:       http://groups.csail.mit.edu/medg/projects/text/findstruct/FindStruct_multiple.zip
Source1:       http://central.maven.org/maven2/edu/mit/findstruct/findstructapi/%{version}/findstructapi-%{version}.pom
Source2:       http://www.apache.org/licenses/LICENSE-2.0.txt

BuildRequires: ant
BuildRequires: javapackages-local
BuildRequires: jdom

BuildArch:     noarch

%description
Many clinical documents are not well structured in the
sense of being based on a formal markup language, but
nevertheless have significant clues within their text
that allow a heuristic computer program to recover
much of their structure and thereby to create a version
with the formal markup included. The FindStruct program
is designed to do just this task. This program, and
others like it, work best when the clinical documents
are created using templates that indicate the document
structure as embedded labels, headings, etc., or
when documents have been generated by computer programs
that were designed to support readability by human readers,
but not by other computer programs. Fortunately, we
have encountered a significant number of such sources of
documents, and the current program is the result of
generalizing methods from a handful of these.

FindStruct takes an XML file that specifies the clues to
be sought in a text document to mark the beginnings (and
occasionally the patterns) of text sections and subsections.
It then processes a set of text documents, searching for the
specified structures in each, and writes an XML file with the
same content as the original document but with explicit markup
having been inserted to indicate the structures that
have been found. Optionally, it also recognizes numbered lists
of items and outputs XML tags to indicate these.

FindStructAPI is used by Apache cTAKES. It
was originally developed out of mit.edu's NLP group.

%package javadoc
Summary:       Javadoc for %{name}

%description javadoc
This package contains javadoc for %{name}.

%prep
%setup -q -c
# Cleanup
find . -name '*.bat' -print -delete
find . -name '*.class' -print -delete
find . -name '*.jar' -print -delete

# fix non ASCII chars
for s in FindStruct/src/findstruct/DParagraph.java;do
  native2ascii -encoding UTF8 ${s} ${s}
done

# Customize pom file
cp -p %SOURCE1 pom.xml
%pom_add_dep org.jdom:jdom:1.1.3

chmod 644 FindStruct/FindStruct_doc.html

cp -p %SOURCE2 LICENSE.txt
sed -i 's/\r//' LICENSE.txt FindStruct/FindStruct_doc.html

%mvn_file edu.mit.findstruct:findstructapi %{name} findstructapi FindStruct

%build

ant -f FindStruct/build.xml \
 -Djavac.source=1.6  \
 -Djavac.target=1.6 \
 -Djavadoc.additionalparam="-Xdoclint:none" \
 -Djavadoc.windowtitle="FindStruct API" \
 -Dfile.reference.jdom.jar=$(build-classpath jdom)

%install
%mvn_artifact pom.xml FindStruct/dist/FindStruct.jar
%mvn_install -J FindStruct/dist/javadoc

%files -f .mfiles
%doc FindStruct/FindStruct_doc.html
%license LICENSE.txt

%files javadoc -f .mfiles-javadoc
%license LICENSE.txt

%changelog
* Sat Sep 12 2015 gil cattaneo <puntogil@libero.it> 0.0.1-1
- initial rpm