Posts Tagged ‘projects’

OEDUc: EDH and Pelagios NER working group

Monday, June 19th, 2017

Participants:  Orla Murphy, Sarah Middle, Simona Stoyanova, Núria Garcia Casacuberta

Report: https://github.com/EpiDoc/OEDUc/wiki/EDH-and-Pelagios-NER

The EDH and Pelagios NER working group was part of the Open Epigraphic Data Unconference held on 15 May 2017. Our aim was to use Named Entity Recognition (NER) on the text of inscriptions from the Epigraphic Database Heidelberg (EDH) to identify placenames, which could then be linked to their equivalent terms in the Pleiades gazetteer and thereby integrated with Pelagios Commons.

Data about each inscription, along with the inscription text itself, is stored in one XML file per inscription. In order to perform NER, we therefore first had to extract the inscription text from each XML file (contained within <ab></ab> tags), then strip out any markup from the inscription to leave plain text. There are various Python libraries for processing XML, but most of these turned out to be a bit too complex for what we were trying to do, or simply returned the identifier of the <ab> element rather than the text it contained.

Eventually, we found the Python library Beautiful Soup, which converts an XML document to structured text, from which you can identify your desired element, then strip out the markup to convert the contents of this element to plain text. It is a very simple and elegant solution with only eight lines of code to extract and convert the inscription text from one specific file. The next step is to create a script that will automatically iterate through all files in a particular folder, producing a directory of new files that contain only the plain text of the inscriptions.

Once we have a plain text file for each inscription, we can begin the process of named entity extraction. We decided to follow the methods and instructions shown in the two Sunoikisis DC classes on Named Entity Extraction:

https://github.com/SunoikisisDC/SunoikisisDC-2016-2017/wiki/Named-Entity-Extraction-I

https://github.com/SunoikisisDC/SunoikisisDC-2016-2017/wiki/Named-Entity-Extraction-II

Here is a short outline of the steps might involve when this is done in the future.

  1. Extraction
    1. Split text into tokens, make a python list
    2. Create a baseline
      1. cycle through each token of the text
      2. if the token starts with a capital letter it’s a named entity (only one type, i.e. Entity)
    3. Classical Language Toolkit (CLTK)
      1. for each token in a text, the tagger checks whether that token is contained within a predefined list of possible named entities
      2. Compare to baseline
    4. Natural Language Toolkit (NLTK)
      1. Stanford NER Tagger for Italian works well with Latin
      2. Differentiates between different kinds of entities: place, person, organization or none of the above, more granular than CLTK
      3. Compare to both baseline and CLTK lists
  2. Classification
    1. Part-Of-Speech (POS) tagging – precondition before you can perform any other advanced operation on a text, information on the word class (noun, verb etc.); TreeTagger
    2. Chunking – sub-dividing a section of text into phrases and/or meaningful constituents (which may include 1 or more text tokens); export to IOB notation
    3. Computing entity frequency
  3. Disambiguation

Although we didn’t make as much progress as we would have liked, we have achieved our aim of creating a script to prepare individual files for NER processing, and have therefore laid the groundwork for future developments in this area. We hope to build on this work to successfully apply NER to the inscription texts in the EDH in order to make them more widely accessible to researchers and to facilitate their connection to other, similar resources, like Pelagios.

OEDUc: Images and Image metadata working group

Tuesday, June 13th, 2017

Participants: Sarah Middle, Angie Lumezeanu, Simona Stoyanova
Report: https://github.com/EpiDoc/OEDUc/wiki/Images-and-image-metadata

 

The Images and Image Metadata working group met at the London meeting of the Open Epigraphic Data Unconference on Friday, May 15, 2017, and discussed the issues of copyright, metadata formats, image extraction and licence transparency in the Epigraphik Fotothek Heidelberg, the database which contains images and metadata relating to nearly forty thousand Roman inscriptions from collections around the world. Were the EDH to lose its funding and the website its support, one of the biggest and most useful digital epigraphy projects will start disintegrating. While its data is available for download, its usability will be greatly compromised. Thus, this working group focused on issues pertaining to the EDH image collection. The materials we worked with are the JPG images as seen on the website, and the images metadata files which are available as XML and JSON data dumps on the EDH data download page.

The EDH Photographic Database index page states: “The digital image material of the Photographic Database is with a few exceptions directly accessible. Hitherto it had been the policy that pictures with unclear utilization rights were presented only as thumbnail images. In 2012 as a result of ever increasing requests from the scientific community and with the support of the Heidelberg Academy of the Sciences this policy has been changed. The approval of the institutions which house the monuments and their inscriptions is assumed for the non commercial use for research purposes (otherwise permission should be sought). Rights beyond those just mentioned may not be assumed and require special permission of the photographer and the museum.”

During a discussion with Frank Grieshaber we found out that the information in this paragraph is only available on this webpage, with no individual licence details in the metadata records of the images, either in the XML or the JSON data dumps. It would be useful to be included in the records, though it is not clear how to accomplish this efficiently for each photograph, since all photographers need to be contacted first. Currently, the rights information in the XML records says “Rights Reserved – Free Access on Epigraphischen Fotothek Heidelberg”, which presumably points to the “research purposes” part of the statement on the EDH website.

All other components of EDH – inscriptions, bibliography, geography and people RDF – have been released under Creative Commons Attribution-ShareAlike 3.0 Unported license, which allows for their reuse and repurposing, thus ensuring their sustainability. The images, however, will be the first thing to disappear once the project ends. With unclear licensing and the impossibility of contacting every single photographer, some of whom are not alive anymore and others who might not wish to waive their rights, data reuse becomes particularly problematic.

One possible way of figuring out the copyright of individual images is to check the reciprocal links to the photographic archive of the partner institutions who provided the images, and then read through their own licence information. However, these links are only visible from the HTML and not present in the XML records.

Given that the image metadata in the XML files is relatively detailed and already in place, we decided to focus on the task of image extraction for research purposes, which is covered by the general licensing of the EDH image databank. We prepared a Python script for batch download of the entire image databank, available on the OEDUc GitHub repo. Each image has a unique identifier which is the same as its filename and the final string of its URL. This means that when an inscription has more than one photograph, each one has its individual record and URI, which allows for complete coverage and efficient harvesting. The images are numbered sequentially, and in the case of a missing image, the process skips that entry and continues on to the next one. Since the databank includes some 37,530 plus images, the script pauses for 30 seconds after every 200 files to avoid a timeout. We don’t have access to the high resolution TIFF images, so this script downloads the JPGs from the HTML records.

The EDH images included in the EAGLE MediaWiki are all under an open licence and link back to the EDH databank. A task for the future will be to compare the two lists to get a sense of the EAGLE coverage of EDH images and feed back their licensing information to the EDH image records. One issue is the lack of file-naming conventions in EAGLE, where some photographs carry a publication citation (CIL_III_14216,_8.JPG, AE_1957,_266_1.JPG), a random name (DR_11.jpg) and even a descriptive filename which may contain an EDH reference (Roman_Inscription_in_Aleppo,_Museum,_Syria_(EDH_-_F009848).jpeg). Matching these to the EDH databank will have to be done by cross-referencing the publication citations either in the filename or in the image record.

A further future task could be to embed the image metadata into the image itself. The EAGLE MediaWiki images already have the Exif data (added automatically by the camera) but it might be useful to add descriptive and copyright information internally following the IPTC data set standard (e.g. title, subject, photographer, rights etc). This will help bring the inscription file, image record and image itself back together, in the event of data scattering after the end of the project. Currently linkage exist between the inscription files and image records. Embedding at least the HD number of the inscription directly into the image metadata will allow us to gradually bring the resources back together, following changes in copyright and licensing.

Out of the three tasks we set out to discuss, one turned out to be impractical and unfeasible, one we accomplished and published the code, one remains to be worked on in the future. Ascertaining the copyright status of all images is physically impossible, so all future experiments will be done on the EDH images in EAGLE MediaWiki. The script for extracting JPGs from the HTML is available on the OEDUc GitHub repo. We have drafted a plan for embedding metadata into the images, following the IPTC standard.

Reflecting on our (first ever) Digital Classicist Wiki Sprint

Wednesday, July 16th, 2014

From (Print) Encyclopedia to (Digital) Wiki

According to Denis Diderot and Jean le Rond d’Alembert the purpose of an encyclopedia in the 18th century was ‘to collect knowledge disseminated around the globe; to set forth its general system to the people with whom we live, and transmit it to those who will come after us, so that the work of preceding centuries will not become useless to the centuries to come’.  Encyclopedias have existed for around 2,000 years; the oldest is in fact a classical text, Naturalis Historia, written ca 77 CE by Pliny the Elder.

Following the (recent) digitalization of raw data, new, digital forms of encyclopedia have emerged. In our very own, digital era, a Wiki is a wider, electronic encyclopedia that is open to contributions and edits by interesting parties. It contains concept analyses, images, media, and so on, and it is freely available, thus making the creation, recording, and dissemination of knowledge a democratised process, open to everyone who wishes to contribute.

 

A Sprint for Digital Classicists

For us, Digital Classicists, scholars and students interested in the application of humanities computing to research in the ancient and Byzantine worlds, the Digital Classicist Wiki is composed and edited by a hub for scholars and students. This wiki collects guidelines and suggestions of major technical issues, and catalogues digital projects and tools of relevance to classicists. The wiki also lists events, bibliographies and publications (print and electronic), and other developments in the field. A discussion group serves as grist for a list of FAQs. As members of the community provide answers and other suggestions, some of these may evolve into independent wiki articles providing work-in-progress guidelines and reports. The scope of the Wiki follows the interests and expertise of collaborators, in general, and of the editors, in particular. The Digital Classicist is hosted by the Department of Digital Humanities at King’s College London, and the Stoa Consortium, University of Kentucky.

So how did we end up editing this massive piece of work? On Tuesday July 1, 2014 and around 16:00 GMT (or 17:00 CET) a group of interested parties gathered up in several digital platforms. The idea was that most of the action will take place in the DigiClass chatroom on IRC, our very own channel called #digiclass. Alongside the traditional chat window, there was also a Skype voice call to get us started and discuss approaches before editing. On the side, we had a GoogleDoc where people simultaneously added what they thought should be improved or created. I was very excited to interact with old members and new. It was a fun break during my mini trip to the Netherlands, and as it proved, very focused on the general attitude of the Digital Classicists team; knowledge is open to everyone who wishes to learn and can be the outcome of a joyful collaborative process.

 

The Technology Factor

As a researcher of digital history, and I suppose most information system scholars would agree, technology is never neutral in the process of ‘making’. The magic of the Wiki consists on the fact that it is a rather simple platform that can be easily tweaked. All users were invited to edit any page to create new pages within the wiki Web site, using only a regular web browser without any extra add-ons. Wiki makes page link creation easy by showing whether an intended target page exists or not. A wiki enables communities to write documents collaboratively, using a simple markup language and a web browser. A single page in a wiki website is referred to as a wiki page, while the entire collection of pages, which are usually well interconnected by hyperlinks, is ‘the wiki’. A wiki is essentially a database for creating, browsing, and searching through information. A wiki allows non-linear, evolving, complex and networked text, argument and interaction. Edits can be made in real time and appear almost instantly online. This can facilitate abuse of the system. Private wiki servers (such as the Digital Classicist one) require user identification to edit pages, thus making the process somewhat mildly controlled. Most importantly, as researchers of the digital we understood in practice that a wiki is not a carefully crafted site for casual visitors. Instead, it seeks to involve the visitor in an ongoing process of creation and collaboration that constantly changes the Web site landscape.

 

Where Technology Shapes the Future of Humanities

In terms of Human resources some with little involvement in the Digital Classicist community before this, got themselves involved in several tasks including correcting pages, suggesting new projects, adding pages to the wiki, helping others with information and background, approaching project-owners and leaders in order to suggest adding or improving information. Collaboration, a practice usually reserved for science scholars, made the process easier and intellectually stimulating.  Moreover, within these overt cyber-spaces of ubiquitous interaction one could identify a strong sense of productive diversity within our own scholarly community; it was visible both in the IRC chat channel as well as over skype. Several different accents and spellings, British, American English, and several continental scholars were gathering up to expand this incredibly fast-pacing process. There was a need to address research projects, categories, and tools found in non-english speaking academic cultures.  As a consequence of this multivocal procedure, more interesting questions arose, not lest methodological. ‘What projects are defined as digital, really’, ‘Isn’t everything a database?’ ‘What is a prototype?’. ‘Shouldn’t there be a special category for dissertations, or visualisations?’.  The beauty of collaboration in all its glory, plus expanding our horizons with technology! And so much fun!

MediaWiki recorded almost 250 changes made in the 1st of July 2014!

The best news, however is that this, first ever wiki sprint was not the last.  In the words of the Organisers, Gabriel Boddard and Simon Mahony,

‘We have recently started a programme of short intensive work-sprints to
improve the content of the Digital Classicist Wiki
(http://wiki.digitalclassicist.org/). A small group of us this week made
about 250 edits in a couple of hours in the afternoon, and added dozens
of new projects, tools, and other information pages.

We would like to invite other members of the Digital Classicist community to
join us for future “sprints” of this kind, which will be held on the
first Tuesday of every month, at 16h00 London time (usually =17:00
Central Europe; =11:00 Eastern US).

To take part in a sprint:

1. Join us in the DigiClass chatroom (instructions at
<http://wiki.digitalclassicist.org/DigiClass_IRC_Channel>) during the
scheduled slot, and we’ll decide what to do there;

2. You will need an account on the Wiki–if you don’t already have one,
please email one of the admins to be invited;

3. You do not need to have taken part before, or to come along every
month; occasional contributors are most welcome!’

The next few sprints are scheduled for:
* August 5th
* September 2nd
* October 7th
* November 4th
* December 2nd

Please, do join us, whenever you can!