search

fechan / lestrade-tei-tagger

Mathematica-based TEI Tagger spun off from Audrey Holmes' Historical Markup Tool
GNU General Public License v3.0
0 stars 0 forks source link

readme

Lestrade TEI Tagger

Lestrade TEI Tagger takes plain text and turns it into a TEI-tagged document. It uses Mathematica/Wolfram Language's text entity recognition system to detect entities to be tagged. The end goal is to use Wolfram Language entity data to extract additional information about tagged entities.

Lestrade TEI Tagger is based off of Audrey Holmes' Historical Markup Tool, which used Flair to detect entities.

Entity ref URN schemes

Lestrade will produce produce various URNs (a kind of identifier) for entities, which allows TEI documents to reference other documents/databases with additional information about those entities. These URNs will appear in the @ref attribute of an entity in a text or TEI index.

TEI Index URN Scheme

If Lestrade tags an entity that appears in a TEI index it generates, the tag in the text will have a URN of this format. This URN corresponds to the name of the TEI index as given in setttings.json plus the @xml:id of the corresponding entity in that index. The URN has the following schema:

urn:teiindex:TEI_INDEX_NAME:ENTITY_ID

Example: If Lestrade tags <persName ref="urn:teiindex:myTeiIndex:bob">Bob</persName> in the original text and the TEI index is named "myTeiIndex", then Lestrade will generate a TEI index named "myTeiIndex" that contains <person xml:id="bob">.

Wolfram Language Entity URN Scheme

If Mathematica thinks a text entity has a Wolfram Language entity interpretation, then the text entity's entry in the TEI index will have a @ref attribute containing a URN with the following schema:

urn:WolframEntity:WOLFRAM_ENTITY_TYPE:WOLFRAM_ENTITY_CANONICAL_NAME

Note: WOLFRAM_ENTITY_CANONICAL_NAME is allowed to contain colons! Since it's the last part of the URN scheme, you know that anything that comes after the 3rd colon is part of the canonical name. The WOLFRAM_ENTITY_CANONICAL_NAME is also encoded with HTML entities so that it can safely be used as an XML attribute value.

Example: urn:WolframEntity:Building:&quot;GreatPyramidOfGiza::jbm66&quot;

Using the URN, you should be able to reconstruct the original Wolfram Language Entity with Mathematica. Decode WOLFRAM_ENTITY_CANONICAL_NAME into ASCII characters and use Entity[WOLFRAM_ENTITY_TYPE, WOLFRAM_ENTITY_CANONICAL_NAME].

Example: Entity["Building","GreatPyramidOfGiza::jbm66"]