EMLAP

EMLAP is a dataset documenting the corpus of Early Modern Latin Alchemical Prints and containing (1) the highly curated digital transcriptions of a representative sample of this corpus; (2) the catalogue of metadata describing individual works; (3) the sentence data with automatically generated morphological annotations; (4) the lemmatised sentences extracted from the morphologically annotated sentence data, and; (5) the scripts used for additional automatic cleaning, preprocessing, and preliminary explorations of the transcriptions.

The ontology of metadata is made within the TOME project (http://tome.flu.cas.cz) with the use of GeoNames, CERL, and VIAF authorities, and with the use of alchemical-historical bibliographical catalogues.

The transcriptions were produced with the use of the TRANSKRIBUS tool for automatic character recognition (http://transkribus.org) and then manually curated in two steps: (1) by domain-qualified research collaborators, (2) by a domain expert in the history of alchemy.

We have developed the dataset for the purpose of automatic metaphor detection and other additional text analysis (by hybrid and distant reading through computational analyses); in the future the transcribed corpus will be used as a basis for a database with a dedicated graphical user interface.

The dataset is continuously maintained on GitHub: https://github.com/CCS-ZCU/EMLAP_ETL

You can immediately start to explore the metadata & textual content of the dataset on the cloud using this Google Colaboratory notebook:

https://colab.research.google.com/github/CCS-ZCU/EMLAP_ETL/blob/master/scripts/colab-explorations.ipynb