ianmulvany / connotea-public Goto Github PK

NAME
    Connotea Code

COPYRIGHT AND LICENSE
    (c) Copyright 2005-2007 Nature Publishing Group.

    This program is free software; you can redistribute it and/or modify it
    under the terms of the GNU General Public License as published by the
    Free Software Foundation; either version 2 of the License, or (at your
    option) any later version.

    This program is distributed in the hope that it will be useful, but
    WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General
    Public License for more details.

    You should have received a copy of the GNU General Public License along
    with this program; if not, write to the Free Software Foundation, Inc.,
    59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.

    Some portions regarding RDF are originally from RDF::Core, derived from
    works Copyright (C) 2001 Ginger Alliance Ltd., and carry their own
    copyright and GPL notices.

NAMING
    You will the see the names Connotea, Bibliotech and Connotea Code used.
    To eliminate any confusion, we'll clarify the meaning of those names
    here.

    Connotea is the name of the online reference management service created
    and run by Nature Publishing Group (NPG). Bibliotech was the initial
    project name used at NPG while the service was being developed, and
    hence this name is used for some class and variable names in the code.
    The release of the underlying technology for Connotea is known as
    Connotea Code. The purpose of this page and the SourceForge project is
    to make the code that runs this site publicly available for review and
    re-use.

    Therefore, it makes sense to refer to Connotea the service, or to the
    Connotea Code. However, Connotea is a trademark of Nature Publishing
    Group, so if you use the code to create your own bookmarking service, we
    ask that you don't brand it as Connotea. We also ask that you include
    the following footer on your site:

      This site is powered by
      <a href="http://sf.net/projects/connotea">Connotea Code</a>,
      the open source software behind
      <a href="http://www.connotea.org/">Connotea</a>.

    The Connotea logo, the site guide and related documentation, other image
    files and stylesheets are copyrighted by NPG and are not released under
    the GPL.

ABOUT THE CODE
    Connotea Code runs a social bookmarking web site for users to save and
    share links, which can have citation data automatically retrieved from
    authoritative sources.

    Connotea Code is written in Perl, and uses MySQL as the data store. It
    runs as a mod_perl handler in Apache2, and uses templates for page
    presentation.

DOWNLOAD
    Download the tarball from the connotea SourceForge project area at
    <http://sf.net/projects/connotea>.

    The current stable release is version 1.8.

UPGRADING
  NEW FEATURES FROM 1.7.1 TO 1.8
    * Web API in regular use.
    * Template Toolkit based templates in regular use.
    * More optimized SQL queries for common requests.
    * Greater use of transactions in MySQL.
    * Greater flexbility for citation source modules.
    * New citation source modules, plus improvements to existing modules.
    * Blog component to create news page from external blog.
    * Wiki component to create custom wiki.
    * Admin component with user search.
    * Integration with Bibutils library for BibTeX and MODS output.
    * Antispam system with captcha and quarantine responses.
    * Click tracker for all posts.
    * Alpha-version proxy module system to handle known proxied post URL's.
    * Alpha-version stand-alone citation server capability.
    * Additional tools such as command-line post by API, user recovery, and
    test suite launch.
    * Automated deployment scripts, now supporting Darcs instead of CVS.
    * Updated code to support newer versions of CPAN modules.
    * More test suite scripts.

  NEW FEATURES FROM 1.5.0 TO 1.7.1
    * Many bugs fixed.
    * Alpha-version Web API.
    * Alpha-version Template Toolkit based output framework.
    * Full text searching feature.
    * Better cache control and throttling.
    * Better bookmarklets.
    * Better URI validation.
    * Better XML encoding for fringe cases.
    * Better character set decoding of downloaded documents for citations.
    * Exception email notification.
    * More support for two instances on same server.
    * More support for split web/database servers for one instance.
    * More comprehensive User Agent support for citation modules.
    * Method to switch from one citation module to another.
    * Optimized SQL for counting totals and some other operations.
    * Added methods for profiling code and dumping SQL statements.
    * Loosen some grammar restrictions, e.g. ok to name a tag "tag".
    * Tighten some grammar restrictions, e.g. num & start must be numeric.
    * Better RIS import based on real-world file examples.
    * New citation modules:

        * Blackwell
        * PMC
        * Wiley
        * ePrints

    * Several optional administrator utilities:

        * retro: script to update citation data rectroactively.
        * bibwatch: load monitoring utility.
        * bibpreempt: preempting and testing utility.
        * resendreg: utility to resend registration details.
        * deluser: utility to delete users.
        * memcache_wrapper: init.d script to keep memcached running.

    * Several developer testing utilities:

        * import: test import modules
        * citation_source_test.pl: test citation modules
        * get_test_urls.pl: retrieve URLs from Yahoo for testing citation
        modules
        * htmlise.pl: convert citation module test results to clickable HTML
        for review in a browser

  UPGRADING FROM 1.7.1 TO 1.8
    See sql/schema_alter.sql for commands to patch your database. Other
    elements of the upgrade should be optional; that is, you can turn them
    on later.

  UPGRADING FROM 1.5.0 TO 1.7.1
    The biggest difference between 1.5.0 and 1.7.1 is that 1.7.1 uses two
    databases at once.

    In order to support fulltext matching, a new feature, we use a MyISAM
    database in MySQL with FULLTEXT keys (see
    <http://dev.mysql.com/doc/refman/4.1/en/fulltext-search.html>).

    However, InnoDB is still faster for JOIN's and offers referential
    integrity, so as a compromise we run two databases and keep them
    synchronized with MySQL replication (see
    <http://dev.mysql.com/doc/refman/4.1/en/replication.html>).

    If you are upgrading from Connotea Code 1.5.0, please see the section
    below on database setup for the secondary search database. To upgrade,
    you will need to:

    * Create a mysql dump that does not mention schema, just data, as in:
         $ mysqldump -c -t -u bibliotech -p bibliotech > /tmp/dump

    * Create the MyISAM search database as described below.
    * Setup replication and restart MySQL as described below.
    * Run sql/wipe.sql to remove all data from your database:
         $ echo 'source sql/wipe.sql' | mysql -u bibliotech -p bibliotech

    * Reimport your dump back into your main InnoDB database, from where it
    will flow to the search database because of replication:
         $ echo 'source /tmp/dump' | mysql -u bibliotech -p bibliotech

    Except for the addition of a MyISAM database, there are no intradatabase
    schema changes between 1.5.0 and 1.7.1.

  UPGRADING FROM VERSIONS PRIOR TO 1.5.0 TO 1.7.1
    To upgrade from versions prior to 1.5.0, please edit
    sql/schema_alter.sql to contain only the statements necessary to alter
    the database schema from your version to the current schema. There are
    no schema changes between 1.5.0 and 1.7.1.

     $ $EDITOR schema_alter.sql
     $ echo 'source schema_alter.sql' | mysql -u root -p bibliotech

    Then follow the directions above for upgrading from 1.5.0.

ACQUIRING SOURCE FOR SPECIALIZED PROGRAMMING
  CREATING A CITATION MODULE
    Connotea's ability to import bibliographic information from third-party
    websites is enabled by a series of plug-ins.

    If you downloaded this source code with the intent of creating a
    citation module, see the comments and code in the file
    Bibliotech/CitationSource.pm which will explain the base class from
    which your citation source module should be derived.

    In previous releases testing your citation module required a full
    instance of Connotea Code. In this release, a script named
    test_util/citation_source_test.pl provides a way to test your module's
    return values without an instance. Your module file should be placed in
    the Bibliotech/CitationSource directory to be recognized by this script.

    You may also test by creating a fully installed instance, which gives
    the added benefit of letting you test via a web browser and ensure that
    citation data is saved properly in MySQL.

    If you create a new citation plug-in, please consider releasing it back
    to the Connotea community.

  CREATING AN IMPORT MODULE
    Connotea's ability to import a batch of links or references depends on a
    series of plug-ins.

    If you downloaded this source code with the intent of creating an import
    module, see the comments and code in the file Bibliotech/Import.pm which
    will explain the base class from which your import module should be
    derived.

    In previous releases testing your citation module required a full
    instance of Connotea Code. In this release, a script named
    test_util/import provides a way to test your module's return values
    without an instance. Your module file should be placed in the
    Bibliotech/Import directory to be recognized by this script.

    You may also test by creating a fully installed instance, which gives
    the added benefit of letting you test via a web browser and ensure that
    imported data is saved properly in MySQL.

    If you create a new import plug-in, please consider releasing it back to
    the Connotea community.

  CREATING A PROXY MODULE
    Connotea's ability to provide proxy translation for specific types of
    URI's depends on a series of plug-ins.

    If you downloaded this source code with the intent of creating a proxy
    module, see the comments and code in the file Bibliotech/Proxy.pm which
    will explain the base class from which your import module should be
    derived.

    You may test by creating a fully installed instance.

    If you create a new proxy plug-in, please consider releasing it back to
    the Connotea community.

  ADDING A STATIC WEB PAGE
    Any Connotea Code instance that contains the "Inc" component has the
    ability to deliver static pages through the template system. A URL path
    that is not recognized by "Bibliotech::Parser" will be tested as a
    filename under the document root with an extension of ".inc" appended.
    The contents of this file should be XHTML. If found, the contents will
    be served within inc.tt or default.tt according to the rules of the
    template system.

  ADDING A DYNAMIC WEB PAGE
    To create a new component for your Connotea Code instance that serves
    dynamic web content requires at least the following:

    In Bibliotech/Parser.pm you must find the grammar definition and add a
    subrule to the page production which will designate the URL path that
    will activate your component. Keep in mind that a path name that is a
    shortened version of another path name will always eclipse the longer
    one if it appears first, so you should add it after (e.g. "urilabel"
    must come before "uri" or "uri" would always match for either).

    In Bibliotech/Page/Standard.pm add a package based on "Bibliotech::Page"
    like the others defined in that file. The name should be
    "Bibliotech::Page::x" where "x" is your path name with a single capital
    letter at the beginning even if it is more than one word (e.g.
    "Bibliotech::Page::Reportspam" for a path of "/reportspam"). Include a
    "main_component()" method that returns a string of the last part of the
    class name of the main component, a "Bibliotech::Component"-derived
    class (e.g. 'ReportSpam' for "Bibliotech::Component::ReportSpam").

    In the Bibliotech/Component directory create a module based on
    Bibliotech::Component. Use the others that appear in that diectory as
    examples and refer directly to the source code in
    Bibliotech/Component.pm, particularly the comments, for descriptions of
    expected methods and their expected return values. For an HTML
    compontent be sure to include "last_updated_basis()" and
    "html_content()". In particular, html_content() should return a
    "Bibliotech::Page::HTML_Content" object; that class is defined in
    Bibliotech/Page.pm.

  SPEAKING TO THE WEB API FROM YOUR APPLICATION
    The Connotea Web API allows communication with an instance, either the
    Connotea web site at <http://www.connotea.org/> or your own private
    instance, using a predefined set of commands to access structured data
    and accomplish normal user actions in a programmatic manner.

    Your software may be written in any language you choose - the basic
    requirements are the ability to create and parse XML and communicate
    using the HTTP protocol. The ability to interpret the XML as RDF and use
    object orientation to model the objects serialized as RDF may prove
    helpful. Libraries and sample code are available.

    See <http://www.connotea.org/wiki/WebAPI> for Web API documentation.

MINIMUM REQUIREMENTS
    This code requires, or has been best tested on:

    * Linux/UNIX operating system (tested on Red Hat Enterprise Linux 4 -
    see <http://www.redhat.com/>)
    * Perl 5.8.0 (see <http://www.perl.org/>)
    * Perl CPAN modules as identified on the list below (see
    <http://cpan.perl.org/>)
    * Apache 2.0.40 (see <http://httpd.apache.org/>)
    * MySQL 5.0.17 (see <http://www.mysql.com/>)
    * Memcached 1.1.12 (see <http://www.danga.com/memcached/>)

  CPAN
    You will need to have the following modules installed from CPAN.

    On all Perl systems you can type:

      $ LANG=C cpan

    ...or...

      $ LANG=C perl -MCPAN -e shell

    ...to get a CPAN shell prompt, and then type:

      cpan> install XXX::YYY

    ..or...

      cpan> force install XXX::YYY

    ...to install a module.

    The "LANG=C" portion of the command line above is highly recommended as
    many modern Linux distributions set your default "LANG" to a
    locale-based setting and this often interferes with Perl module
    compilations. When it does, the error messages will be very misleading
    and never mention the "LANG" variable.

    Before you embark on what will probably be a long install-fest, it is
    also recommended that you type:

     cpan> install Bundle::CPAN

    ...inside the CPAN shell and then restart it. This will ensure that you
    are using the latest version of the CPAN code. Some things will go more
    smoothly.

    When asked whether to follow dependencies, answer yes. When asked about
    optional utilities and scripts that can be installed to /bin or
    /usr/bin, answer however you like, as none are necessary for this code.

    You do not necessarily need the latest version of every module, although
    in one or two cases you do. In general, if your Perl is at least 5.8.0,
    just install the version that a non-force install will give you at the
    CPAN prompt. If you are lower than 5.8.0, upgrade your base Perl
    installation first.

    The list:

    On Red Hat and some other distros, the following are provided in vendor
    packages, and you're better off using those.

    * Apache2
    * Apache::Const
    * Apache::File

    ...but install these from CPAN so you get new versions:

    * IO::String (for Bio::Biblio::IO, better to preinstall)
    * XML::Writer (for Bio::Biblio::IO, better to preinstall)
    * XML::Twig (for Bio::Biblio::IO, better to preinstall)
    * SOAP::Lite (for Bio::Biblio::IO, better to preinstall)
    * Pod::Man (for DateTime, better to preinstall)
    * Bio::Biblio (may need to be forced)
    * Cache::Memcached
    * CGI
    * Class::DBI
    * Config::Scoped
    * Data::Dumper (not just for debugging, actually used in production)
    * Date::Parse
    * DateTime (you may need to force installation of DateTime::Set if your
    timezone is not UTC)
    * DateTime::Format::ISO8601
    * DateTime::Format::MySQL
    * DateTime::Incomplete
    * Digest::MD5
    * Encode (you may need to force installation of Encode if some
    non-English tests fail)
    * Fcntl
    * File::Temp
    * File::Touch
    * FindBin
    * HTML::Entities
    * HTML::Sanitizer (you may need to force installation of HTML::Sanitizer
    due to some year-old bugs already filed on CPAN)
    * HTTP::OAI
    * IO::File
    * JSON
    * LWP::UserAgent
    * List::MoreUtils
    * List::Util (you may need to force installation of List::Util unless
    you have a very new version of Perl)
    * Net::Daemon::Log (you may need to force installation of
    Net::Daemon::Log for failing a fork test - not used by us)
    * Netscape::Bookmarks
    * Parse::RecDescent
    * RDF::Core
    * SQL::Abstract
    * Set::Array
    * Storable
    * Template
    * Test::Exception
    * Time::HiRes
    * URI
    * URI::Escape
    * URI::Heuristic
    * URI::OpenURL
    * URI::QueryParam
    * Want
    * Wiki::Toolkit
    * Wiki::Toolkit::Plugin::Diff
    * XML::Element
    * XML::Feed
    * XML::LibXML
    * XML::RSS
    * YAML (you may need to force installation of Test::Simple which is a
    dependency of YAML, for an unknown reason)
    * Apache::Emulator (not required for core web service service)
    * Text::BibTeX (not required for core web service service)

SETUP
  MYSQL
    Two databases for user posts need to be created. See sql/schema.sql for
    the database schema which needs to be created in MySQL. The first
    database will be created using InnoDB tables to enforce foreign keys and
    constraints and for table joining speed. A second database then should
    be created with a _search suffix using MyISAM tables that have FULLTEXT
    indexes which are queried when searching for words. (FULLTEXT indexes
    are not available for InnoDB yet.)

    The second schema is generated from the first by running:

      $ cd sql
      $ perl mkschema_search < schema.sql > schema_search.sql

    MySQL relication can be used to make the MyISAM database a slave of the
    InnoDB database, even on the same machine. This is a suggested
    configuration for /etc/my.cnf that will do just that:

      [mysqld]
      # local replication of bibliotech to bibliotech_search:
      server-id=1
      log-bin=mysql-bin
      binlog-do-db=bibliotech
      replicate-same-server-id=1
      replicate-rewrite-db=bibliotech->bibliotech_search
      replicate-do-db=bibliotech_search
      master-host=localhost
      master-user=search_repl
      master-password=pass
      # change stopwords in support of bibliotech freematch feature:
      #ft_stopword_file=/etc/mysql_stopwords.txt
      ft_min_word_len=2
      ft_max_word_len=255
      # allow packing of queries
      group_concat_max_len=8192

    Change the master-password line! Also change the database names if you
    are not using "bibliotech".

    You will probably find the MySQL stopwords to be too restrictive in
    practice. The list can be viewed at
    <http://dev.mysql.com/doc/mysql/en/fulltext-stopwords.html>. We
    recommend that you pare down this list to a more suitable one, and use
    the ft_stopword_file keyword to tell MySQL to use your list instead.

    In any case, if you want the search feature to behave predictably, you
    must specify an external text file stopword list to MySQL. The search
    handler will query MySQL to find out the stopword list file being used,
    and read it as well, so it can anticipate MySQL reporting no matches for
    words that otherwise should match.

    You'll need to execute a grant statement similar to this one:

     GRANT REPLICATION SLAVE, REPLICATION CLIENT ON *.* TO
     search_repl@'localhost.localdomain' IDENTIFIED BY 'pass';

    Two notes on the replication grant statement:

    * MySQL seems to consider "localhost.localdomain" different from
    "localhost" and while the shorter version normally works, for
    replication it seems that the longer one is needed. If you have
    problems, try both.
    * You must have the updated privilege table structure. If you have had
    MySQL installed since the 3.x series, your mysql.user table lacks the
    privilege fields mentioned above; check your docs about a script called
    'mysql_fix_privilege_tables'. On many systems this will be a shell
    script in /usr/bin that you can run as root with a "--password=xxx"
    parameter (to specify the MySQL root user password, not the Unix root
    user password).

    The MySQL username used by the Perl handler must have access to both
    databases (username and password as in /etc/bibliotech.conf):

      GRANT SELECT, INSERT, UPDATE, DELETE ON bibliotech.* TO
      user@localhost IDENTIFIED BY 'secret';
      GRANT SELECT, INSERT, UPDATE, DELETE ON bibliotech_search.* TO
      user@localhost IDENTIFIED BY 'secret';

  WIKI::TOOLKIT
    You also need to setup Wiki::Toolkit so that a wiki is available. This
    is required. You should create a blank database, grant a user rights to
    it, and run the provided setup script.

      CREATE DATABASE conwiki;
      GRANT ALL ON conwiki.* TO conwiki@localhost IDENTIFIED BY 'secret';

      $ /usr/bin/wiki-toolkit-setupdb --type mysql \
                                      --name conwiki \
                                      --user conwiki \
                                      --pass secret \
                                      --host localhost

    Remember to populate the "COMPONENT WIKI" block of your configuration
    file with the wiki database details.

  APACHE
    Everything under the site/default subdirectory should be placed or
    linked into an Apache-accessible location, and a location handler should
    be added to httpd.conf (or elsewhere in the Apache configuration) such
    as the following one.

    Update the values to match your IP, domain, and file paths:

      <VirtualHost 1.2.3.4:80>
        ServerName www.yourdomain.com
        ServerAlias yourdomain.com
        ServerAdmin [email protected]
        DocumentRoot /var/www/perl/connotea_code/site/default
        PerlOptions +Parent
        PerlSwitches -I/var/www/perl/connotea_code
        PerlModule Bibliotech::Apache
        PerlModule Bibliotech::AuthCookie
        <Location />
          SetHandler perl-script
          PerlHandler Bibliotech::Apache
          PerlAuthenHandler Bibliotech::AuthCookie::authen_handler
          AuthName Bibliotech
          AuthType basic
          require valid-user
          #ErrorDocument 503 /paused.html
          #ErrorDocument 503 /readonly.html
          ErrorDocument 503 /unavailable.html
        </Location>
      </VirtualHost>

    The 503 lines allow a custom page to be displayed when your site is
    under heavy load (unavailable.html) or when you deliberately pause
    service (paused.html) or make it read-only (readonly.html); you must
    edit your Apache configuration and switch which line is commented for
    the latter two modes.

  MEMCACHED
    Memcached is required, and the code is written to assume that a memcache
    is running. Database timestamps, cached HTML, and uploaded files are all
    stored temporarily in this cache.

  CONFIGURATION
    See config for a configuration that should be copied to
    /etc/bibliotech.conf and edited to suit your needs. Particularly, be
    sure to change *_SECRET and *_PASSWORD variables.

    Default configuration:

      (((config)))

  CUSTOMIZATION
    The look and feel of your Connotea Code installation can be modified by
    creating a new stylesheet and new templates. The template system is
    Template Toolkit documented at the web site at
    <http://www.template-toolkit.org/>. We refer to this system as TT for
    short.

   TEMPLATE LOCATION
    Templates are located by default in site/default. This is controlled by
    options in the configuration. It is recommended that templates have a
    .tt extension.

   TEMPLATE SELECTION
    The template used to service a particular request is determined by the
    page requested and the available template filenames.

    Individual templates can be defined for individual pages; for example,
    to override the template for the add form, create a template called
    add.tt.

    For general bookmark listing queries (e.g. "/tag/tagname"), templates
    beginning with recent can be used. recent.tt will be used for queries
    with no user or tag parameters - recent_user.tt, recent_tag.tt and
    recent_user_tag.tt can be created to specify the behaviour is there is a
    user query, a tag query, or both respectively.

    Unless overridden by a specific template, default.tt is used.

   TEMPLATE EXAMPLES
    Templates should not contain the full HTML for the page you want to
    construct, but only that which should appear between the "<body>" and
    "</body>" tags.

    This is an example default.tt:

      [% prepare_component_begin() %]
      [% prepare_component('main',undef,'main,verbose') %]
      [% prepare_component_end() %]
      <html>
      <head>
      <title>[% main_title %]</title>
      [% rss_link %]
      [% component_javascript_block_if_needed %]
      </head>
      <body[% component_javascript_onload_if_needed %]>
      [% component_html('main',undef,'main,verbose') %]
      </body>
      </html>

    The syntax is from Template Toolkit documented at the web site at
    <http://www.template-toolkit.org/>. We refer to TT for short.

    A Connotea web page is a series of components that are combined
    together, contributing HTML which can be organized in separately-placed
    parts calculated at once, or as one block, and also sometimes Javascript
    to be placed in a "<script>" block in the "<head>" or in the "<body>"
    "onload" attribute. The components are controlled by the template
    selected to represent the HTTP query, and each component running can
    access the current command, the posts that result from the SQL engine
    processing the query of the command, and a variety of support services.

    Several functions are provided to TT by the calling instance.
    "prepare_component_begin()" is called before anything happens, and this
    prepares some internal data structures; then "prepare_component()" is
    called with the base name of the component (e.g. "Blah" corresponds to
    "Bibliotech::Component::Blah") or the special word "main" which does a
    lookup for the main component of a page described in
    Bibliotech/Parser.pm and Bibliotech/Page/Standard.pm, which allows some
    templates to be reused. The second argument is a comma-separated list of
    parts, which is a mechanism used by some components such as
    "ListOfTags", and the third argument is a comma-separated list of
    options to the components, of which the universal ones are "main" and
    "verbose" which should be set to true for the main component (the format
    is key=value but ommiting the value is the same as =1 which is true).
    Later in the template, a call to "component_html()" with the same
    arguments inserts the HTML (if you specified multiple part names in the
    prepare call, you should have multiple calls to insert HTML each with
    one part mentioned).

    Individual templates can be defined for individual requests. For
    example, to override the template for the add form, create a template
    called add.tt. For general bookmark listing queries (e.g.
    "/tag/tagname"), templates beginning with recent can be used. recent.tt
    will be used for queries with no user or tag parameters -
    recent_user.tt, recent_tag.tt and recent_user_tag.tt can be created to
    specify the behaviour is there is a user query, a tag query, or both
    respectively.

    A more realistic example for default.tt would make calls to a
    normalprep.tt and a normal.tt wrapper:

      [% prepare_component_begin() %]
      [% INCLUDE normalprep.tt %]
      [% prepare_component('main',undef,'main,verbose') %]
      [% prepare_component_end() %]
      [% WRAPPER normal.tt %]
      [% component_html('main',undef,'main,verbose') %]
      [% END %]

    In this case, normal.tt would contain a basic look for the web site that
    can be used by many other templates.

    A set of working templates is furnished with this distribution.

   TEMPLATES AND COMPONENTS
    A Connotea web page is a series of components that are combined
    together, contributing HTML which can be organized in separately-placed
    parts calculated at once, or as one block, and also sometimes Javascript
    to be placed in a "<script>" block in the "<head>" or in the "<body>"
    "onload" attribute. The components are controlled by the template
    selected to represent the HTTP query, and each component running can
    access the current command, the posts that result from the SQL engine
    processing the query of the command, and a variety of support services.

    Several functions are provided to TT by the calling instance. Some
    functions are general, and are available in all templates, even small
    snippet templates used by components (conventionally named with a comp
    prefix). Some functions are page-level utilities that largely control
    component insertion.

   TEMPLATE PAGE LEVEL FUNCTIONS
    * "prepare_component_begin()"
        Declare the beginning of "prepare_component()" calls.

    * "prepare_component(module, parts, options)"
        Prepare a component.

        * module
            The base name of the desired component (e.g. "Blah" corresponds
            to "Bibliotech::Component::Blah") or the special word "main"
            which does a lookup for the main component of a page described
            in Bibliotech/Parser.pm and Bibliotech/Page/Standard.pm, which
            allows some templates to be reused.

        * parts
            For most components, use "undef" for one block of HTML output.
            For components that can return multiple parts of HTML, this
            option is a comma-separated list of part names to prepare in one
            calculation for efficiency. Components with parts: "ListOfTags",
            "ListOfUsers", and "ListOfGangs".

        * options
            A comma-separated list of options to the component, of which the
            universal ones are "main" and "verbose" which should be set to
            true for the main component (the format is key=value but
            ommiting the value is the same as =1 which is true).

    * "prepare_component_end()"
        Declare the end of "prepare_component()" calls.

    * "component_html(module, part, options)"
        The arguments are the same as "prepare_component()", except that the
        part should be at most one part, not more than one.

    * "component_javascript_onload()"
        Insert the Javascript addressed at the "onload" handler.

    * "component_javascript_onload_if_needed"
        Insert the Javascript addressed at the "onload" handler, but wrap it
        a space followed by the actual attribute itself, as in, "
        onload="blah"", or if there is no Javascript, insert nothing.

    * "component_javascript_block()"
        Insert the Javascript addressed at the "head" of the HTML document.

    * "component_javascript_block_if_needed()"
        Insert the Javascript addressed at the "head" of the HTML document,
        but wrap it in a "<script>" block, or if there is no Javascript,
        insert nothing.

    * "main_title"
        The HTML document title recommended by the main component, or
        failing that, a default constructed from the site name and page
        name.

    * "main_heading"
        The HTML document heading ("H1") recommended by the main component.

    * "main_description"
        The description recommended by the main component; used in RSS, etc.

    * "css_link"
        Insert a "<link>" representing the CSS files dictated by
        "HOME_CSS_FILE" for the home page or "GLOBAL_CSS_FILE" otherwise
        (configuration options).

    * "rss_link"
        Insert a "<link>" representing the RSS format output for the
        currently viewed page.

   TEMPLATE GENERAL FUNCTIONS
    * "location"
        Base URL for the web site, which can be directly prepended to page
        names, as in:

          <a href="[% location %]news">

    * "sitename"
        The name of the web site as defined in the configuration.

    * "siteemail"
        The email address of the administrator of the web site as defined in
        the configuration.

    * "user"
        User object of the current user looking at the web page, e.g.:

          [% IF user %][% user.username %][% ELSE %]Visitor[% END %]

    * "is_browser_safari", "is_browser_firefox", "is_browser_ie",
    "is_browser_other"
        Can be used in an IF test - true if the user's browser is the type
        indicated.

    * "browser_redirect(url)"
        Immediately abort and issue a Location header to a new URL. The URL
        can be relative to the root of the web site in which case "location"
        is prepended.

    * "is_virgin"
        Can be used in an IF test - true if the user is a first-time
        visitor.

    * "canonical_uri"
        Canonical URI for the current page.

    * "canonical_location"
        Canonical URI using "location".

    * "object_location"
        Canonical URI using "location" and setting format to HTML.

    * "no_num(url)"
        Remove the num=x parameter from a URL.

    * "encode_xml_utf8(str)"
        Escape ampersand, less- and greater-than symbols, normalize HTML
        entities to XML entities, and remove unusual control characters.

    * "encode_xhtml_utf8(str)"
        Encode characters as XML entities where needed and remove unusual
        control characters.

    * "now"
        The current date and time, as a "Bibliotech::DateTime" object, e.g.:

          [% now.label %]
          [% now.ymd %]
          [% now.ymdhm %]
          [% now.iso8601 %]
          [% now.iso8601_utc %]

    * "time"
        The current date and time, as a Unix timestamp.

    * "join(joinstr, ...)"
        Perl's join command.

    * "speech_join(jointype, ...)"
        Join several elements as in speech. Argument "jointype" is "and" or
        "or". This function will combine with commas, spaces, and the
        jointype operator (if at least three items), e.g.:

          speech_join('and', 'bob') -> 'bob'
          speech_join('and', 'bob', 'alice') -> 'bob and alice'
          speech_join('and', 'bob', 'alice', 'tom') -> 'bob, alice, and tom'

    * "plural(amount, singular, plural, no_space)"
        Join a number with the appropriate singular or plural noun, e.g.:

          plural(6, 'second', 'seconds') -> '6 seconds'
          plural(1, 'second', 'seconds') -> '1 second'

    * "commas(num)"
        Decorate a number with commas every three digits (thousands) per the
        American style, e.g.:

          commas(5000000) -> '5,000,000'

    * "divide(a, b, places, multiplier)"
        Divide two numbers, but avoid a division by zero error by returning
        zero. Return a number formatted to the number of decimals indicated
        in "places" (default 1 if omitted), and multiplied by "multiplier"
        (default 1 if omitted), e.g.:

          divide(10, 0) -> 0
          divide(10, 2) -> 5
          divide(10, 4, 2) -> 2.50
          divide(10, 4, 2, 100) -> 250

    * "percent(a, b, places)"
        Same as "divide" but multiplier is 100 and a percent sign is
        appended.

          percent(1, 2) -> 50.0%
          percent(4, 100) -> 4.0%

    * "bookmarklets"
        Insert all the bookmarklets.

    * "bookmarklet(page, popup)"
        Insert a bookmarklet. Argument "page" should be "add", "addcomment",
        or "comments". Argument "popup" should be "direct" or "popup".

    * "bookmarklet_js(page, popup)"
        Same as "bookmarklet" but only insert the Javascript.

    * "user_in_own_library"
        Can be used in an IF test - true if the user is looking at a page
        that has a current filter of "/user" with their username.

    * "user_in_another_library"
        Can be used in an IF test - true if the user is looking at a page
        that has a current filter of "/user" with a username other than
        their own.

    * "click_counter_onclick(url, new_window)"
        Insert a hyperlink run through the click counter, to the URL
        provided, optionally in a new window if "new_window" is true.

   TEMPLATE COMPONENT SNIPPET FUNCTIONS
    * "sticky(parameter)"
        Used primarily inside the "value" attributes of HTML "ipnut" tags,
        this function allows the form to remember values between refreshes,
        so if a message must be displayed to the user causing the form to be
        redisplayed, the user's form responses are not lost.

    * "has_validation_error"
        True if the form is redisplayed with an active error.

    * "has_validation_error_for(field)"
        True if the form is redisplayed with an active error concerning the
        field specified. Note that some errors are not tied to a field.

    * "validation_error_field"
        If the form is redisplayed with an active error, the field name that
        gave rise to the error. Note that some errors are not tied to a
        field.

    * "validation_error"
        If the form is redisplayed with an active error, the error message.

  LOGROTATE
    A log file will be created at the place specified in "LOG_FILE" in the
    configuration. This file can grow quickly so you may like to configure
    logrotate to deal with it on a weekly basis. This is the contents of a
    suggested /etc/logrotate.d/bibliotech file:

      /var/log/bibliotech.log {
              create 644 apache root
              notifempty
              weekly
              rotate 5
              compress
          postrotate
              /bin/kill -HUP `cat /var/run/httpd.pid 2>/dev/null` 2> /dev/null || true
          endscript
      } 

  SUPPORT
    To subscribe to the Connotea Code development mailing list, go to
    <https://lists.sourceforge.net/lists/listinfo/connotea-code-devel>.

    This list is for the discussion of the core code and citation and import
    plug-ins. It is intended for use by people who are installing their own
    instance of Connotea Code, or who are reviewing the code to see how
    Connotea handles their data, or who would like to help enable the
    importing of bibliographic information from more sources.

    There is a separate list, connotea-discuss, for discussion of Connotea
    itself - i.e. for discussion of <http://www.connotea.org/>. That list is
    the appropriate place for questions about how to use the site, or
    requests or suggestions for new features.

  WEBCITE
    The Connotea Code has a module called WebCite which can be employed
    separately as a simple web service providing citation information using
    the Connotea Code citation modules.

    A WebCite instance requires the full codebase to be present, as well as
    the CPAN module dependencies, and a compatible version of Apache, just
    as you might setup for Connotea Code proper. However, MySQL, Memcache,
    and Wiki::Toolkit are not required.

    You should create a configuration file as directed for Connotea Code,
    but only the WEBCITE section and sections for your active citation
    modules are required.

    WebCite provides caching not with memcached but via the filesystem, so
    the results survive Apache restarts.

    WebCite can be turned on inside a normal Connotea Code deployment, but
    will perform its duties separately from the main codebase.

    The main page for WebCite simply presents a form with two fields:

    * "uri" - text field for the URL
    * "fmt" - select field for the format:

        * "ris"
        * "mods"
        * "json"

    A submit button is provided for human users but the value is ignored.

    The form can be called by programs by performing an HTTP POST to the
    installed location with "uri" and "fmt" parameters.

    Programs should evaluate the HTTP status code first.

    * 200 - citation data returned
    * 404 - no citation data found
    * 500 - internal error

    In the case of a 404 code, a brief message such as "No citation" may be
    displayed, but programs should not expect this value, and should rely
    exclusively on the HTTP status code to determine this condition.

    In the case of a 200 code, the "Content-Type" header should be
    appropriate for the format requested, and the data payload should
    present the citation data in the format requested.

    A 500 code will occur if an exception is thrown retrieving the citation
    data, and the data payload will be plain text giving the error message.

    All transactions are separate. There is no concept of sessions.

    There are no authentication checks in the WebCite code, although the
    system administrator is free to add restrictions in the Apache
    configuration.

    Subsequent requests within 90 days for the same URL will return data
    cached from the original request. Data is cached in an internal
    structure form so the same cache entry can produce all output formats.

    The Apache configuration block for WebCite is as follows:

      PerlSwitches -I/var/www/perl/...
      <Location /bibliotech>
        SetHandler perl-script
        PerlHandler Bibliotech::WebCite
      </Location>

    The PerlSwitches line should point to the directory that contains the
    Bibliotech directory.

    This may or may not appear in the same Apache configuration as Connotea
    Code proper.

    A suggested configuration for /etc/bibliotech.conf is as follows:

      WEBCITE {
        CACHE_ENABLED = true
        CACHE_PATH = '/var/cache/webcite'
        CACHE_TIMEOUT = 7776000
        LOG_ENABLED = true
        LOG_FILE = '/var/log/webcite.log'
      }

    Again this may be a file exclusively for WebCite or a file with
    intermixed configuration for Connotea Code proper.

    You should create the cache directory and log file and give the Apache
    user write access before starting Apache.

ACKNOWLEDGMENTS
    The look, structure, documentation and source code of
    <http://www.connotea.org/> are the collective work of Martin Flack, Ben
    Lund, Timo Hannay, Joanna Scott, Stefania Bojano, Grant Farrelly, Euan
    Adie, and Ian Mulvany. The vast majority of the programming was done by
    Martin Flack of NeoReality, Inc., <http://www.neoreality.com/>.

    The materials available from <http://sf.net/projects/connotea> are
    released under the GNU General Public License; reuse of all other
    materials requires the express written permission of Nature Publishing
    Group.

MORE INFORMATION
    Please visit this URL for more information:
    <http://www.connotea.org/code>

    If you have questions, email us or try our mailing lists:
    <http://www.connotea.org/contact>