mapling

Mapling finds things, such as place names, in texts. It returns a csv file with a row for each occurence. For each file, it creates an html page with the things highlighted. Just point mapling to a folder full of documents. Mapling uses textract to extract text from many types of files, including csv, doc, docx, pdf, html, txt and many others.

Usage: $ mapling texts/ --gazetteer=gazetteer/gazetteer.txt --model=de_core_news_sm --html To install a spaCy model: $ python -m spacy download de_core_news_sm

• The first approach is to use a gazetteer. Mapling expects a txt file with a row for each place name. Add the --gazetter argument with the path to your file. This approach lets you search for specific terms (not just places) that appear in the text. $ mapling /dir/with/txt_files --gazetteer="/home/me/gazetter.txt"

• The second approach uses a spaCy named entity recognition model. Add the --model argument with the name of an installed spaCy model. If your model is not installed or does not have an ner pipeline, you'll get instructions on how to fix that. This approach will return a large range of entities and places, more than you might list yourself. This is useful for establishing which places, people and organizations appear in a text. $ mapling /dir/with/txt_files --model=de_core_news_md

• Finally, mapling can create visualizations. Add the --html argument $ mapling /dir/with/txt_files --model=de_core_news_md --html

To install:

pip install mapling

In the future, mapling will also work with the Word Historical Gazetteer to rectify, geocode and map your place names.