Extract extent from non-data files

[nuest]

There has been some research on extracting geospatial properties from text and scientific papers. These reange from complex ML models to brute force text extraction. geoextent could wrap other tools and provide the same API for PDF files and DOIs that point to papers, not data. This is a completely new feature angle! Possibly more a research thesis topic..

Resources and related work:

• Extracting and modeling geographic information from scientific articles > https://doi.org/10.1371/journal.pone.0244918 and the code at https://github.com/eacheson/pyscine
• @tnier01's work is also related: when we have point locations, how to get the next relevant administrative region, see http://nbn-resolving.de/urn:nbn:de:hbz:6-69029469735
• M. Mohr thesis extracted coordinates from maps in papers (unpublished)
• Kmoch, A., Uuemaa, E., Klug, H., & Cameron, S. (2018). Enhancing Location-Related Hydrogeological Knowledge. ISPRS International Journal of Geo-Information, 7(4), 132. ​ https://doi.org/10.3390/ijgi7040132
• https://www.journalmap.org/​ ; Karl, J. W., Herrick, J. E., Unnasch, R. S., Gillan, J. K., Ellis, E. C., Lutters, W. G., & Martin, L. J. (2013). Discovering Ecologically Relevant Knowledge from Published Studies through Geosemantic Searching. BioScience, 63(8), 674–682. ​ https://doi.org/10.1525/bio.2013.63.8.10

[nuest]

Maybe there is a way to find all Zenodo records that have a connection with a paper? Can we identify supplemental material this way?

[SbastianGarzon]

Maybe:
https://developers.zenodo.org/#depositions

related_identifiers

relation: Relationship. Controlled vocabulary (isCitedBy, cites, isSupplementTo, isSupplementedBy, isNewVersionOf, isPreviousVersionOf, isPartOf, hasPart, compiles, isCompiledBy, isIdenticalTo, isAlternateIdentifier)