[ Description of multi-level structures ]
ANNODIS_me.zip contains the annotated Glozz files (.aa + .ac + .aam + .as) after normalisation and cleaning:
ANNODIS_me.zip

For downloading the ANNODIS_me corpus without annotations (xml and Glozz files), please see below the Records of ANNODIS_me section.

The following files allow direct access to the annotated structures, which can be selected on the basis of specific properties, and viewed in context:

• ANNODIS_SE.xml contains all SE-type structures (Enumerative Structures) with descriptive properties and annotated cues ( readmeSE.xml describes the annotations contained in this file);
• ANNODIS_CT.xml contains all CT-type structures (Topical Chains) with descriptive properties and annotated cues (readmeCT.xml describes the annotations contained in this file).

To explore these files, click on the following link: ANNODIS_me browser. The exploration and visualisation of the annotated structures calls upon files which are in the ANNODIS_me browser archive.

1. corpus constitution and pre-processing
   • 1_ANNODIS_me_Original.zip: the original documents (WIK1 and WIK2: html format; GEOP and LING: pdf format)..
   • 2_ANNODIS_me_XML.zip: the documents in xml format conforming to TEI-P5 (including TEIP5.dtd) with transcriptTEI.xsl style-sheet designed to allow visualisation with layout respecting the original documents' typographical contrasts and disposition.
   • 3_ANNODIS_me_GlozzFiles.zip: the texts made ready for annotation via the Glozz annotation tool (text files= filename.ac and standoff annotations including layout and premarked features = filename.aa) PLUS the Glozz annotation model (ANNODIS_me.aam) and the Glozz stylesheet (ANNODIS_me.as) required for annotating and highlighting premarked and annotated features and structures (see the annotation manual).
2. corpus annotation and post-processing
   • 4_ ANNODIS_me_RawAnnotations.zip: Glozz files as produced by annotators
   • 5_ANNODIS_me_AfterManualCorrect.aip: Glozz files corrected for errors such as incoherence in cue names
   • 6_ANNODIS_me_AfterAutoClean.zip: Glozz files after 3 automatic procedures: 1) deletion of unattached units; 2) normalisation of feature structures for cue annotation; 3) flagging of unknown annotations for manual categorisation
   • 7_ANNODIS_me_Gold.zip: a Gold standard version for the subset of texts that were multi-annotated.
   • 8_ANNODIS_me_HandCategorisation.zip: Glozz files after manual categorisation of unknown annotations.

ANNODIS_me_history = table listing the records of each file.

The ANNODIS resource is available under Creative Commons By-NC-SA 3.0 (Attribution-NonCommercial-ShareAlike). Please read it carefully.

Lydia-Mai Ho-Dac
Contact :