INSTALL

================================================
Installation prerequesites for using KbBridge
================================================

-------------------------
Before compiling KbBridge
-------------------------


1) Install protobuf
If you are working on Ubuntu, use:
    sudo apt-get install protobuf-compiler
You may have to fix the path with
    sudo ln -s /usr/bin/protoc /usr/local/bin/protoc
For other Operating Systems, download from here: https://developers.google.com/protocol-buffers/


2) Download and Install CIIR_Shared
Download or clone repository from: https://github.com/daltonj/CIIRShared
From inside the directory call
    mvn install

3) Download and install LingPipe 4.1.0
Download and unpack from: http://alias-i.com/lingpipe/web/download.html
From download location call
    mvn install:install-file -Dfile=lingpipe-4.1.0.jar -DgroupId=com.aliasi -DartifactId=lingpipe -Dversion=4.1.0 -Dpackaging=jar

------------------
Compiling KbBridge
------------------

cd into the kbbridge directory (where the pom.xml is located)
    mvn clean compile

If you experience any errors, please contact us.


--------------------
Configuring KbBridge
--------------------

create a copy of resources/kbbridge-blake.properties and place it in resources/kbbridge.properties
(Alternatively, you can also place it in resources/kbbridge-$username@$computername.properties)

Make sure all paths are available on your system

Check paths in ../config/galago-fullwiki
Either obtain the Wikipedia index from us, or create your own, see below.


-----------------
Running KbBridge
-----------------

KbBridge is meant to be a framework that you adapt to your needs.
For an example, run the class "EntityLinkerMain".


-------------------
Index your own Data
-------------------

If do not want to use our Wikipedia index, you can create your own from an annotated WEX dump
You need to gave Galago3.4 installed and access to the binary "galago"

create a json file "wikiparser.json" with this content:
---------------------------------
{
"filetype" : "wex" ,
"parser" : {
  "externalParsers" : [
  {
    "filetype" : "wex",
    "class" : "org.lemurproject.galago.core.parse.WikiWexParser"
  }
 ]
}
}
----------------------------------

Call
    core/target/appassembler/bin/galago build --tokenizer/fields+title --tokenizer/fields+title-exact --tokenizer/fields+fbname --tokenizer/fields+fbname-exact --tokenizer/fields+fbtype --tokenizer/fields+category --tokenizer/fields+redirect --tokenizer/fields+redirect-exact --tokenizer/fields+kb_class --tokenizer/fields+anchor --tokenizer/fields+anchor-exact --tokenizer/fields+text --tokenizer/fields+stanf_anchor --tokenizer/fields+stanf_anchor-exact --indexPath=$indexdir --inputPath+$inputdir wikiparser.json


And add $indexname to ../config/galago-fullwiki  (full paths required)