Skip to main content

Users may construct their corpora from scratch or by importing a set of existing files (e.g. recordings and annotations). A corpus is organised in Communications (communicative situations) and Speakers. Each Communication may consist of several Recordings and Annotations. Speakers participate in Communications with specific Roles; and Annotations contain Annotation Levels and Attributes (see next section). The corpus management module of Praaline contains the main Corpus Editor, the tabular version of the Corpus Editor, and the Corpus Structure Editor. Using the Corpus Editor, the user may add, remove and modify items (Communications, Speakers, Recordings and Annotations) in the corpus. The Corpus Editor is shown below: the left pane shows the corpus items organised as a tree. The user may choose to group them in a hierarchy based on the items metadata (for example, in a corpus containing different speaking styles and age groups, the user may opt to show the tree by speaking style first, then by age group). The middle and right pane are metadata editors: when a Communication is selected, the middle pane shows all metadata related to it and its associated Recordings and Annotations, and the right pane all the metadata associated with the Speakers participating in the Communication; conversely, when a Speaker is selected, the primary metadata editor (middle pane) shows the fields relating to the speaker, and the secondary metadata editor (right pane) all the Communications in which this speaker is participating. It is also possible to edit metadata in tabular form (by corpus item type).


Corpus Explorer tree


Unlike most other available systems, the fields available for metadata are not fixed in Praaline. The user may define a metadata structure, using an unlimited number of fields, organized in groups, for each type of corpus item (e.g. a set of fields for Communications, a set of fields for Recordings, a set of fields for Speakers, etc.). The definition of this metadata structure, along with the definition of the annotation structure (which will be explained in the next section) is performed through the Corpus Structure Editor, shown below.


Corpus Structure Editor


Praaline provides standard templates to facilitate the definition of metadata structure and annotation structure when a new corpus is created. It is possible to import data from Exmaralda and Anvil \parencite{Kipp.2001}. It is also possible to import metadata that are stored in tabular files or Excel spreadsheets, using a wizard that interactively establishes the correspondence between the file's columns and the corresponding metadata field in Praaline. By storing the corpus metadata and annotations in a network database server, and corpus media files in a network-accessible location, it is possible to set up collaborative projects.