Getting Started with EPrints 3
Creating an Archive
EPrints 3 can run run multiple archives under one install. Multiple archives will require giving additional DNS aliases to the machine running EPrints, EPrints can then create all the parts of the apache configuration file needed to run the virtual hosts.
Creating the Archive
Make sure MySQL is actually running.
Change to your eprints user (probably "eprints").
Change directory to the eprints directory (/opt/eprints3 by default) and run
You will get the following prompts
- Archive ID - the system name for your archive. It's probably a good idea to think of something short and memorable. Once entered, an archive/<archive_id> directory will be created, and the standard configuration files will be copied in.
- Configure vital settings - Hit enter to say 'yes'. This will lead to more prompting about core settings:
- Hostname - What someone will type into a web browser to get to your archive. Make sure that your systems team have a DNS alias pointing to your server for this.
- Webserver Port - Which port to you want to serve the archive on? The default is 80, so unless you can think of a good reason not to, just hit enter to accept the default.
- Alias - You can enter any number of aliases that will take users to this archive. Enter a '#' when you don't want to enter any more. You could have your archive served on eprints.myorganisation.org and eprints.myorg.org. As with the Hostname, your systems team need to be informed about these aliases too.
- Administrator Email - Enter the email address of the repository administrator. This will allow your repository users to send email to the right person.
- Archive Name - The full name of your archive. By default, this will be used on many of the pages, and in the title bar of the browser.
- Write these core settings - If you don't say 'yes', then you entered all that data for nothing.
Running a Live Archive
Creating a crontab
When you create an archive it will start out as a development system while you learn how to set it up (and your manager keeps changing his mind) but at some point (hopefully) you will declare your archive open for business.
At this point you should schedule certain scripts to run periodically. The best way to do this is to use "cron" which is an integral part of most UNIX systems.
To set up cron, run (as the eprints user):
% crontab -e
Exactly what to add to the cron table is described in the following sections - "Browse Views" and "Subscriptions".
There should be one set of crontab entries per archive.
You should also have made sure that the system is being properly backed up. This is gone into in more detail elsewhere in the documentation.
We would also encourage you to configure the OAI support for your archive and register it. It's quite easy - pretty much fill in the blanks in the ArchiveOAIConfig.pm file in the archive configuration directory.
EPrints 2.1 support OAI versions 1 and 2 at URL paths /perl/oai and /perl/oai2.
Once you register your archive (at http://www.openarchives.org) various search systems will be able to collect the metadata (titles, authors, abstract etc.) and allow more people to find records in your archive.
See http://www.openarchives.org/ for more information on the OAI protocol. For more information setting up the OAI interface archive see the section in this documentation about Configuring an Archive.
Once every so often you should run the "generate_views" script on each archive in your system to regenerate the browse views section of the site.
This is a set of static pages. By default one per subject, and one per year (only years with papers in that year not EVERY year ever!). Some users prefer to browse the system than search it. This also gives search engines a way to reach, and index, the abstract pages.
See the ArchiveConfig.pm config notes on how to edit the views it generates.
See the How-To section for some suggestions on how to set up views.
But I don't want this feature...
If you don't want to use this feature: don't, it's your archive. Remove the link from the template and front page. Don't run the generate_views script.
Setting it up
This is best done by using the UNIX "cron" command (as user "eprints"). Cron will email "eprints" on that machine with the output, so best use the --quiet option so it only bothers you with errors.
How often you want to run this depends on the size of your archive, and how fast the contents changes. This feature is roughly order "n". Which means if you double the number of items in your archive then you double the time it takes to run (ish).
Once an hour would seem a good starting point. If your archive gets real big, say more than 10000 records, then maybe once a day is more realistic - the one thing that you don't want to happen is for a new generate_views to start before the old one finishes as they will mess up each others output.
Run generate_views on the command line to find out how long it takes.
and add the line
23 * * * * /opt/eprints2/bin/generate_views I<archiveid>
This runs at 23 minutes past each hour. If you have more than one archive, don't make them all start rebuilding stuff at the same time, stagger it. Otherwise once an hour everything will slow down as it fights to run several intensive scripts at once.
See the crontab man page man 5 crontab for more information on using cron.
Subscriptions provide a way in which users of your system can receive regular updates, via email, when new items are added which match a search they specified.
To automate sending out these subscriptions you must add some entries in the crontab (as for views). You need one set of these per archive.
For example (with dookuprints being the name of the archive):
# 00:15 every morning 15 0 * * * /opt/eprints2/bin/send_subscriptions dookuprints daily # 00:30 every sunday morning 30 0 * * 0 /opt/eprints2/bin/send_subscriptions dookuprints weekly # 00:45 every first of the month 45 0 1 * * /opt/eprints2/bin/send_subscriptions dookuprints monthly
Note the spacing out so that all 3 don't start at once and hammer the database. You may wish to change the times, but we recommend early morning as the best time to send them (midnight-6am).
But I don't want users to be able to do this!
Then remove the "subscription" power from each type of user in the archives ArchiveConfig.pm file.
EPrints configures a new archive with a set of metadata fields aimed at an archive of research papers.
The initial "types" of eprint (book, poster, conference paper) are configured in metadata-types.xml
The initial subjects are a subset of the library of congress subjects. Feel free to totally replace them with your own subjects, but the more standard your subject tree the more useful your metadata will be to other people.
The authors and editors have the "hasid" option set which allows people to optionally use a unique id for a person in addition to their name (names are NOT unique!) - this can be useful for generating "CV" pages (see the views how-to) and possibly for generating statistics. Without it you will never be sure which "John Smith" wrote that paper. If you don't like this feature remove the "hasid" from the authors and editors - this will require you to recreate the tables, erasing the archive, so decide before you start. If you want to be more clear about what information goes in that field, edit the phrases eprint_fieldname_authors_id and eprint_fieldname_editors_id in the archive phrase file(s).
In general: Change it! It's not a recommended system setup, just a good starting point.
If you are setting up more than one archive which are related to each other, a "community", you may wish to establish common subjects and metadata.
Removing and adding types is easy. Removing and adding fields is a bit more work. All "screen" names of values are stored in the archives own "phrase file" which comes with phrases for the default config.
If you create a good default configuration for a different purpose or language(s) (and would like to share it), please contact the eprints admin who may want to put it online as an example or even include it as an alternate default in a later version.