PapyGreek

PapyGreek 

How to use the PapyGreek-platform

PapyGreek has been developed by

Erik Henriksson 

for the project Digital Grammar of Greek Documentary Papyri

(ERC-Starting Grant 2017, No. 758481)

Principal Investigator:  Marja Vierros

University of Helsinki

Email addresses follow the format: firstname.lastname@helsinki.fi

1.  Sematia: Annotating papyri or inscriptions

Sign-in

Sign in using a Google ID. Without signing in, you can browse and export texts and the metadata, but not modify or annotate them. If you wish to annotate, you need to be granted access to do so (contact Erik and Marja). When you sign in with the status of an annotator, your work stays under your name (also in the exported files) and you can access your documents by clicking your name on the top right.

Select your document

Under “Collections” you find Documentary papyri (the whole Duke Databank of Documentary Papyri as in papyri.info), Literary Papyri (DCLP) (and Inscriptions [not yet available]). Select the document you want to work with from the list (use Filter for finding your edition; the abbreviations are the same as in papyri.info, DCLP is organised by Trismegistos numbers). Click on the selected text (e.g. p.hels.1.10.xml).

Document window

Includes seven tabs: View, Edit tokens, Import data, Annotate, Edit annotation, Metadata, and Versions. Without signing in, you can only View the text and annotations.

View

In the window that opens, you can see the edition of the text, with shifts into different hands taken into account (m1, m2 etc.); colour codes highlight possible differences between the original and regularised “layers”. Original text is above the line and regularised under the line. Also abbreviations, supplements and underdots are presented in two layers; the brackets will stay in the original layers when the text is uploaded into Arethusa.

Edit tokens

Be sure to check the text carefully before beginning the annotation for possible typos, word division mistakes etc. Check also that corrections suggested in the Berichtigungsliste have been applied in papyri.info). Most preferred way for correcting mistakes in the edition is to insert changes via the Papyrological Editor, so that mistakes are first corrected in the Committed version of DDbDP or DCLP. These will get updated in PapyGreek within a week after they have been committed, and then you can start annotating.

Editing tokens in this tab should therefore be done with great caution. Here you can modify both layers, but not the script-generated versions. Editing tokens creates a new version of the texts (see Versions tab). You can e.g. correct typos (by clicking the word), add or delete words (if e.g. crasis needs to be divided) or sentence divisions.

To edit a word form (or some of its other data, such as language or hand), click on the field you wish to change. A text input will appear where you can make the changes. When finished editing, either press enter or click anywhere outside of the field. Your changes will now show with a green background, but they are not yet saved – to commit your change in the PapyGreek database, click Save. To delete a word, click the minus symbol on the right of the word. To add a word, click the plus symbol. To add a sentence break after a word, click the circle symbol. Remember to add final punctuation before the new sentence break you created. Again, remember that your changes will only be saved if you click the Save button.

If you wish to revert to the script-generated version, you can do so by clicking the Revert to original button on the top left (this button is only visible on user-generated versions). Note that this will irreversibly destroy your manual edits.

Import data

This tab is used when you already have an existing annotation or part of it in XML somewhere (old Sematia, old PapyGreek, or one of the layers annotated first in PG, automatic morphology by Alek Keersmaekers https://github.com/alekkeersmaekers/duke-nlp) etc.), and wish to import that (possibly for using as a basis for a new enhanced one) into PapyGreek.

Importing works as follows. First, choose the layer (Original/Regular) for which you have some pre-annotation data. Then, hover over the word where you wish to begin importing. For example, if you only have data for a single sentence, hover over its initial word. The button will allow you to import from an XML file, and allows you to paste the XML data to a textarea (click on to save). To save your imports to the PapyGreek database, click Import. The other buttons () are useful if the pre-annotated file and the one in PapyGreek have a different number of tokens. For example, if there is a word in PapyGreek that is missing from the imported file, you can correct this by clicking on which will add an empty word above or below the current one, respectively. To skip a token’s data from being imported, click to clear its data. To delete an imported token, click . (This will move up the data for all the following words). Again, note that your changes will only be saved after you click Import.

Annotate

Here you open either the Original or the Regularized version of the text in our local version of Arethusa annotation platform. You can have two windows open and proceed with the Original and Regularised annotations side by side. Another option is to annotate another layer first (usually the regularised is easier to start with) and then use the Import data tab to import the xml for basis. You can leave comments on you annotation decisions within Arethusa Comment tab, and also in PG (used for more general comments).

Annotation Guidelines

• Firstly, we follow the Ancient Greek and Latin Dependency Guidelines

• For Greek:  2.0 by G. Celano (excluding the semantic layer)  treebank_data/Greek_guidelines.md at master · PerseusDL/treebank_data

• For some instances (especially those appearing in papyri or those that aroused questions among the annotators) we have collected our own set of Guidelines.

After annotation

After both layers have been annotated in Arethusa, please submit them to PapyGreek Board for checking. It is not possible to make changes while the text is under review. The Board will either Accept or Reject the text (and comment on needed changes in the Arethusa Comment tab). If rejected, the text will appear under Rejected in the list of Annotated texts and you or someone else can again edit them (following revision suggestions by the Board). When Accepted, the text will appear under Finalized and added to the PapyGreek corpus of treebanked papyri. An xml-file of the annotation can be downloaded separately or as part of the whole corpus. Your Google account name will remain as annotator in the downloaded xml-file and you can list these publications in your CV.

Edit annotation

In this tab you can quickly edit the pos-tags, words, labels of the annotation (like in xml in Arethusa) without having to go to Arethusa (click the lemma, postag, relation or head and write what you wish; then remember to save from the upper left corner). Editing annotations does not work if the text is already finalized, unless an admin releases it back for editing in PG. Note that the window shows the tags that differ from one another between the orig and reg layers highlighted in grey. You can also quickly copy and move all tags of one sentence them from the reg layer to the orig, or vice versa (separately the lemma, postag, relation or head columns) with icons at the sentence row.

Metadata

Here you can add metadata concerning the author, text type, addressee, hand, scribal official and actual writer and link the information to one or more text parts or hands (m1, m2 etc). For people, use Trismegistos Person ID when possible, but in the case of hands whose owner is not (known) in the TM database, you may want to create a PapyGreek ID, especially for later linking with other documents written by the same hand. It is possible to mark the metadata information uncertain in all fields.

        To add a text type, clicknext to TEXT TYPE (the button becomes visible when you move the mouse close to the text). Then, first choose a Hypercategory. If you want to add more fine-grained information about the text type, also select a Category and optionally, a Subcategory as well. When finished, click save. Multiple text types can be added if needed.

        To associate a person with a text, proceed as follows. To start, click on next to PEOPLE. This will create an association template that looks like this:

To add a person ID, click on the icon next to [no person ID]. Here you can either input the linked ID (if you know it), browse IDs that have been added to the system previously, or add a new one. Then, choose a role for the person, again by clickingnext to [no role specified]. If you are not certain of the association, click on CERTAIN, which will toggle the certainty to UNCERTAIN. To add even more fine-grained metadata about the association, click on FIELDS. Here you can describe the handwriting, epithet, ethnicity etc. of the person. To delete the association, click.

Versions

View different versions of the text, in case the text has been edited in the platform, or a new version has been updated from papyri.info Click Open next to a version to view it. Note that old versions are locked from being edited.

2.  Paratypa: Searching orthographic variants

Paratypa can be found under the tab heading Variations. It draws the data from the Duke Databank of Documentary Papyri and performs searches based on the <reg> and <orig> tags within <choice> in the EpiDoc XML encoding of the transcriptions of the papyri.

You can search for replaced letters (like ε written for η, i.e. the editor has corrected ε of the original into η), additions (like ι added by the editor not written in the papyrus), deletions (like ι written in the papyrus but considered superfluous by the editor) as well as unchanged letters in the original.