About ===== ETD is an Electronic Theses & Dissertation site built on top of the Fedora Commons repository. Software Dependencies ===================== php5, with the following components/configurations - browscap.ini - tidy - fileinfo - geoip - sqlite - ldap - oci8 (Oracle) - PHP_Shell - php memory limit 256MB or greater pdfinfo & pdftohtml (version from Poppler is recommended) For Ubuntu setup instructions, see https://techknowhow.library.emory.edu/etd-documentation/development-environment - javascript xforms FormFaces implementation (included with source) - javascript wysiwyg editor FCKeditor (included with source) svn externals ============= ETD depends on the following externals, which are included in src/lib: - Zend Framework (local copy, slightly customized - minor bugfixes not yet in released version) - Emory extensions to Zend Framework (services & extensions meant to be used with ZF) - xml-utilities (customizable xml object) - fedora-php (php wrapper for Fedora APIs, foxml & datastream objects based on xml-utilities xml object) - simpletest (for unit testing, external under test) - OFC (for admin reports), under lib/OFC (php) and public/js/ofc (javascript) Services & Databases ==================== ETD makes use of or interacts with the following: - Fedora repository, with Gsearch - Solr (index based on Fedora data, populated by Fedora Generic Search Service) - LDAP authentication - Emory Shared Data (Oracle db) - ETD Util DB (MySQL db) - Persistent Id Server (locally developed - persis/pidman) - sqlite3 database for ETD usage statistics - CSV graduation feed from Registrar (triggers publication) - ProQuest ftp server (electronic submissions) - Google Analytics cron jobs ========= The following scripts (in src/scripts) should be run as cron jobs: - statistics.php : etd-specific view/download usage, based on apache log, stored in sqlite db; should be run daily. - publish.php : publication & submission to ProQuest; depends on Registrar feed; should be run at least weekly. - embargo_expiration.php : checks for expiring embargoes and sends various notices 60 days, 7 days, and 0 days before an embargo expires. Should be run daily. Code organization ================= ETD is a Zend Framework application with roughly MVC structure. The site is bootstrapped & loaded through src/public/index.php, which hands off control & routing to the Zend Front Controller. The bulk of the code (controllers, models, and views) can be found under src/app. For more details, see Zend Controller quick start documentation: http://framework.zend.com/manual/en/zend.controller.html#zend.controller.quickstart.go An Etd Controller base class along with a few custom controller helpers can be found under src/lib/Etd. Command-line scripts are found in src/scripts, and use a common bootstrap.php to load config files and initialize services. Data migration scripts are in src/scripts/migrations. Unit tests are under test/ and organized roughly comparable to src. As much as possible, these tests can be run individually or in groups, and can be run on the command line or via the web. Console: src/scripts contains a console.php that loads etd models & services (including fedora connection) and can be used as an interactive environment (REPL). In tests/suites there is an equivalent console.php that loads the test environment. Gsearch: Currently Gsearch is indexing RELS-EXT and MODS datastreams. Due to most objects having managed datastreams, the etdFoxmlToSolr.xslt must make API-A-LITE calls to fedora to retrieve datastreams. The policies allow this action from localhost only. If in the future GSearch does not have access to localhost the policies would have to be modified to allow access from the host making requests. PDF File Parsing ================ This section describes the order (and some criteria) that pages must be in for the PDF file to be parsed correctly on submission. * Distribution Agreement - first page * Signature Page (This is where we pull the Title and Advisor / Committee info) - Comes after Distribution Agreement * Abstract Cover (No Data read from this page) - Comes after Signature Page. * Abstract Page 1 - Comes after Abstract Cover * Abstract Page 2 (Optional - Abstract Page 2 is assumed if the next page is not the Title Page) - Comes after Abstract page 1 * Title Page - (No data read from this page) - Comes after Abstract section * Table of Contents Page 1 (Looks for text "Table of Contents") - Comes after Title Page * Table of Contents Additional pages (Optional - Assumes page is a TOC if it finds "Chapter" or "...<number>" EXAMPLE: "Chapter 2" "...15") - Comes after Table of Contents Page 1 Calendar Feed ============= Calendar information comes form a google calender feed. The url to use can be obtained by logging into the google account and going to: calendar -> (under My Calendars section click the down arrow) -> Share this calendar -> Calendar Details -> (under Private Address) -> XML you must change the last part of the url from "basic" to "full" Removing Records ================ 1. Look at the RELS-EXT of the main object and make a list of the associated objects (PDF, AuthorInfo, Original, Supplemental). 2. Delete each associated object then delete the main object. 3. In PIDManager mark the pid for each main object and associated object as inactive. Common Errors ============= **Symptom:** Error: bnash3 (role=superuser) is not authorized to view Page **Solution:** Edit RELS-EXT and remove refs to non-existent objects **Symptom:** Icon is not displayed beside Original File and when the file is downloaded it only contains "Error parsing media type 'application/msword application/msword' " **Solution:** Do ONE of the following: 1. re-save a good version of the file using MSWord in the latest docx format and replace the file in ETD. 2. Manually change the mime type of the FILE datastream to "application/msword". **Symptom:** After an admin uploads a file for a student the student can no longer access the ETD or the file. **Solution:** Change the owner of the file object to the netid of the ETD submitter. See also ======== Additional documentation is found in the Documentation directory. Documentation/README-Roles - summary of the roles used in ETD, and how they work Documentation/README-Services - explanation of the Fedora services used Documentation/README-MultiSchoolConfiguration - explanation of the per-school configuration and fields supported Documentation/README-EmailNotifier - more information about the email notifier Documentation/README-Solr - more information about solr