This git repository hosts the transcription data of the project What is Bedouin-Type Arabic? (ERC Advanced Grant 101020127) (October 2021 to September 2026).
Principal Investigator: Stephan Procházka (University of Vienna)
Cooperation Partner: Charly Mörth (Austrian Academy of Sciences)
See https://wibarab.acdh.oeaw.ac.at/ for more information
Contact us at wibarab@oeaw.ac.at or follow us on Twitter.
THIS IS PRELIMINARY DATA AND COPYRIGHTED MATERIAL!
If you want to use any material in this repository please contact PI Stephan Procházka (University of Vienna).
This will change at the end of the project.
| Directory | Content | Remarks |
|---|---|---|
001_src |
Original sources | Source documents (e.g. raw transcriptions) |
080_scripts_generic |
Conversion Scripts | mostly the ELAN2TEI conversion script (implemented in Python) which generates the initial TEI data prior to tokenization based on the ELAN transcription documents in 122_elan |
082_scripts_xsl |
XSLT scripts | XSLT scripts |
103_tei_w |
TEI-XML with tokens | This is where ELAN2TEI puts its output. Re-running TEI2ELAN will overwrite all content in this directory, so do not do any manual changes here but copy the file to 010_manannot beforehand. |
010_manannot |
manually annotated TEI-XML | Tokenized TEI documents from 103_tei_w which are manually annotated. |
802_tei_odd |
TEI customization (ODD) | This is the source of truth for the SHAWI Schema and the HTML documentation generated from it. |
130_vert_plain |
NoSketch Engine Verticals | NoSketch Engine text verticals |
803_RNG-schematron |
Schemas | derived from the ODD in 802_tei_odd |
804_xsd |
Schemas | derived from the ODD in 802_tei_odd |
850_docs |
Documentation | Further data documentation, esp. the HTML documentation of the ODD |
The directories css, html, js and xsl are used by the TEI Enricher.
- Master files of the audio recordings are stored on the project's network share at the University of Vienna
- the metadata spreadsheet is hosted on Sharepoint.
NB The workflow will be described in due manner but will be very similar to the one described in the SHAWI Data Processing and Curation Document
The following steps happen before data is ingested into this repository:
- fieldwork (recording audio etc.) – The recordings so far cover only material collected in previous campaigns
- campaign metadata is added in the "sources" document in the WIBARAB featuredb repository
- collecting metadata: – This is collected at curated in [the metadata spreadsheet].
Workflow steps reflected in the data in this repository:
- Transcription and translation – Curators segment the audio recordings into sensible sets of "utterances" and transcribe and translate them using ELAN. When transcription has finished, the curator adds the ELAN document(s) to
122_elanand pushes the changes to git. - Tokenization This push triggers the ELAN2TEI conversion workflow which takes all *.eaf files in
122_ELANand transforms them into tokenized standalone TEI documents, storing them under103_tei_w. Additionally, a TEI Corpus file is generated which includes corpus level metadata and controlled vocabularies. - Annotation After transformation to TEI, curators annotate the texts using the TEI_enricher and store the results under
010_manannot. - Conversion to NoSke Verticals During the tokenization process, a NoSke-compatible vertical is created which incorporates the annotations found in ``010_manannot` .
- Deployment The data is automatically deployed to an instance of the NoSketchEngine corpus query engine.