Texts
Texts in the database are listed. The list can be filtered by text name, provenance, period, genre and creator. Matching is case-insensitive and regular expressions can be used.
After selecting a text, a list of pages is presented. After selecting a page, one may investigate the annotation of that page. The plain view shows the unannotated facsimile or photo. Zooming is by the mouse wheel or by pressing > and <. Navigation is by dragging with the mouse or with arrow keys.
The glyphs view shows annotated signs in different colors. Moving the mouse over a sign shows its normalized form and Gardiner name. The lines view can be used to display all the annotations of a single line at once, by clicking on any sign in that line.
A text can be downloaded as zip file that includes all images and their annotations. Such a zip file can be uploaded to another instance of the application.
Signs
Occurrences of signs are listed, including ligatures (groups of signs) in which signs occur. If no signs are entered, all signs in the database are listed. Signs are entered in the form of Gardiner names, or as mnemonics such as D and nfr. Multiple signs can be entered, separated by spaces. Instead of individual signs, one may also enter ligatures, formed with Unicode control characters. By clicking ?, a new page is opened with a graphical tool for forming groups, which are copied back to the search field after pressing Save.
Sign occurrences can be filtered by text name, provenance, genre, period, and creator; as above, matching is case-insensitive and regular expressions can be used. To keep the browser from crashing due to an excessive number of images, one needs to enter at least one sign name, or (part of) a text name.
There is an option to see sign occurrences in the context of neighbouring signs. By clicking on a sign, the relevant page is shown.
Overview
Signs in the database are listed, with number of occurrences per text. As above, signs and texts can be filtered.
Analysis
Analysis consists in dimensionality reduction, to 1, 2 or 3 dimensions. One or more signs may be entered as Gardiner names; also ligatures may be entered. For 2 or 3 dimensions, each sign occurrence is represented by a dot. Zooming and navigation is by the mouse or by keys, as described before. By moving the mouse over a dot, the corresponding shape is shown. By clicking on a dot, the relevant page is shown.
If the space bar is pressed while the mouse hovers over a dot, the corresponding shape persists in a corner of the graph (for this to work, make sure the graph is under the focus; if not, click on the background of the graph). By repeating this action a second time for the same dot, the shape is removed from the graph. If the space bar is pressed while the mouse is not hovering over a dot, all shapes are removed from the graph.
For 1 dimension, the visualization of values along the x-axis is such that each sign is represented by an area of a fixed size that does not overlap with neighbouring areas. This means that values that are close together appear as thin and tall areas and values that are far apart appear as wide and short areas.
By moving the mouse over the items in the legend on the right, all but the relevant sign occurrences are displayed in yellow.
From among the methods of dimensionality reduction, UMAP tends to be the slowest, due to a known issue with importing from the umap-learn package.
Guesser
Draw an hieratic sign with the paint tool (press the button or type p). Pixels can be erased with the eraser tool (press the button or type e). After pressing or typing Enter, the best matching candidates are listed. The canvas is wiped clean after pressing or typing Delete.