Installing isis-oai-provider
Isis-oai-provider is a software application that provides access to ISIS databases through OAI protocol to be harvested.
The application works with PHP 5.3.x.
Contents
Instalation
Method 1 - Download files
Download the last version from Downloads link at the github repository (zip or tar gz file) and extract the file to your server.
- Creating a structure and extracting application files under server /opt/ directory on a Linux server:
$ cd /opt/ $ mkdir bireme $ cd bireme $ wget https://github.com/bireme/isis-oai-provider/tarball/master -O isis-oai-provider.tar.gz $ tar xvfzp isis-oai-provider.tar.gz $ mv bireme-isis-oai-provider-53c0abc/ isis-oai-provider
NOTE: the term 53c0abc is automatically generated during the creation of the tar.gz file, thus being different at every execution.
- Downloading the application files on a Windows server:
- access the application page available at https://github.com/bireme/isis-oai-provider
- click on the download zip button to download the application files
- unpack to the directory it is intended to
Method 2 - Clone the project git repository
The advantage of this method is that you can easily update the application later, but it is necessary that the server has the git-core package installed.
Example: Creating a structure and downloading the application from repository under /opt/ directory in the server:
$ cd /opt/ $ mkdir bireme $ cd bireme $ git clone git://github.com/bireme/isis-oai-provider.git
Deploy the aplication on the Web Server (ex. Apache Web Server)
Option A - Using a specific VirtualHost for the web application
<VirtualHost *:80>
ServerName isis-oai-provider.local
DocumentRoot /opt/bireme/isis-oai-provider/htdocs/
<Directory "/opt/bireme/isis-oai-provider/cgi-bin">
Options +ExecCGI
AllowOverride None
Order deny,allow
</Directory>
ScriptAlias /cgi-bin/ /opt/bireme/isis-oai-provider/cgi-bin/
LogLevel warn
CustomLog /var/log/apache2/isis-oai-provider.log combined
ServerSignature On
</VirtualHost>
Option B - Using a existing VirtualHost
Assuming you have a website already running you will only need to find out the paths to the following components:
- DocumentRoot --- the directory where static files and PHP scripts are located - usually htdocs
- cgi-bin --- the directory with special permissions for executing CGI scripts - usually cgi-bin
- ISIS databases --- the directory on the server where ISIS databases are located - usually bases
Then move/copy or create symbolic links to the isis-oai-provider directories:
Ex. Assuming a website with the following directory structure:
/home/sites/
/site1/
htdocs/
cgi-bin/
bases/
Follow the steps below to create symbolic link to the application directories:
$ cd /home/sites/site1/htdocs $ ln -s /opt/bireme/isis-oai-provider/htdocs/isis-oai-provider/ . $ cd /home/sites/site1/cgi-bin $ ln -s /opt/bireme/isis-oai-provider/cgi-bin/isis-oai-provider/ .
Before starting the configuration
In order that ISIS-OAI-Provider works correctly, it is mandatory to rename some files according to the platform in use.
Binary executable files
Below cgi-bin/isis-oai-provider/ there are 4 (four) executable binary files, being:
wxis.exe-Linux wxis.exe-Windows wxis1660.exe-Linux wxis1660.exe-Windows
- In case it is a Linux operating system, rename the files as described below:
- from: wxis.exe-Linux
- to: wxis.exe
- from: wxis1660.exe-Linux
- to: wxis1660.exe
- In case it is a Windows operating system, rename the files as described below:
- from: wxis.exe-Windows
- to: wxis.exe
- from: wxis1660.exe-Windows
- to: wxis1660.exe
NOTES:
- Be sure that all executable files have execute rights either in command line, or through CGI interface;
- All executable files are 32 bits version;
- There are 64 bits version files compiled only for Linux.
Binary data tables
Below htdocs/isis-oai-provider/gizmo/ there are 3 subdirectories:
- linux
- windows
- src
The subdirectories linux and windows contain binary data files required for the correct operation of the system, therefore:
- In case it is a Linux operating system, copy the files as described below:
- from: htdocs/isis-oai-provider/gizmo/linux/gnoctrl.mst
- to: htdocs/isis-oai-provider/gizmo/gnoctrl.mst
- from: htdocs/isis-oai-provider/gizmo/linux/gnoctrl.xrf
- to: htdocs/isis-oai-provider/gizmo/gnoctrl.xrf
- In case it is a Windows operating system, copy the files as described below:
- from: htdocs/isis-oai-provider/gizmo/windows/gnoctrl.mst
- to: htdocs/isis-oai-provider/gizmo/gnoctrl.mst
- from: htdocs/isis-oai-provider/gizmo/windows/gnoctrl.xrf
- to: htdocs/isis-oai-provider/gizmo/gnoctrl.xrf
Configuration files
Rename the sample configuration files as follows:
$ cd htdocs/isis-oai-provider/ $ mv oai-config-sample.php oai-config.php $ mv oai-databases-sample.php oai-databases.php
Modify the oai-config.php and oai-databases.php files according to the steps below:
oai-config.php
- change the variable DATABASE_PATH to the 'bases' directory of current installation
DATABASE_PATH=/home/www/my-iah/bases
- change the variable NAME to a label that best identify the instance
NAME=MyIAHdb
- change the variable EMAIL to the mail address of the person in charge of data
EMAIL=me@institution.org
- change the variables IDPREFIX and IDDOMAIN according to the domain information of your instance
IDPREFIX=br
IDDOMAIN=bvs
- set the variable EARLIESTDATESTAMP to a date close to the beginning of data recording
EARLIESTDATESTAMP=1998-12-30
oai-databases.php
Each database must be configured having in mind the following structure and mandatory elements:
[set-spec]
name= description= database= mapping= prefix=oai_date_ isis_key_length=<1660|1030> identifier_field=<tag> datestamp_field=<tag>
- [set-spec] starts and describes a section for that database and identifies the logical name of that set. This information is presented in the verb ListSets and it is also part of the unique identifier for each document within the set.
e.g.
[lilacs]
- The name variable is the label to the data set to be displayed in the ListSets verb.
e.g.
name=LILACS
- the description variable is the content of description showed in the ListSets verb.
e.g.
description="LILACS Database"
- the database variable has the path and the database master file name.
e.g.
database=/home/www/my-iah/bases/iah/dblil
- the mapping refers to the filename format used to do the mapping of contents. (Located in map/ directory)
e.g.
mapping=lilacs-dc.pft
- the prefix stands for a prefix for the date field. This should be inserted in the FST file used by the invertion process.
e.g.
prefix=oai_date_
- the isis_key_length specifies the minimum and maximum key length used to invert data. Pay attention to the version according to the database source (e.g.. LILACS is 1660, IAH is 1030)
e.g.
isis_key_length=1660
- the identifier_field informs the tag for the unique identifier of records
e.g.
identifier_field=2
- the datestamp_field informs the tag for record creation date on the record
e.g.
datestamp_field=93
Therefore, when the user needs to add new databases to the configuration file, it is only required to fullfil the fields above accordingly and append them to the file.
NOTE: provided that section name is also the database name, no database duplicates are allowed. A possible - but not recommended - workaround is to generate symbolic links to each file that comprises the database (.mst, .xrf, .ifp, .l01, .l02, .n01, .n02 and .cnt files)
Follows a sample configuration of Events directory database
[direve] name=direve description="Events directory" path=/home/projects/isis-oai-provider/bases/isis-oai-provider mapping=direve.i2x prefix=oai_date_ isis_key_length=1030 identifier_field=10 datestamp_field=507
Testing
Test the application by opening the internet browser and accessing the domain name as: http://nome_do_servidor/isis-oai-provider/
E.g. http://isis-oai-provider.local/isis-oai-provider/
Testing
Test the application calling your browser and informing your server name like: http://your_server_name/isis-oai-provider/
