Explore Datashare as a way to explore indexed documents

[peetucket]

See https://datashare.icij.org/

1. Spin up locally or on a new/existing VM
2. Add police manuals

• This could be useful for the QA as well as for exploration on a clean set
• The demo show me this: We can tag individual documents. That means we can distribute QA easily by letting people tag as keep/delete/whatever
• Fantastic feature # 2 for QA: Filter by format. Which are .doc files, which are pdf, which are OCR/pdf
• Fantastic feature # 3: Search with operators and Regex! If this works, it will be great for us.

not sure it would be an alternative to a custom Blacklight instance - Nicole sees this as a tool that would work really well for us as we QA and later for advanced users (researchers) exploring a clearly defined collection such as the manuals

[jmartin-sul]

we have an instance up on the data analysis VM. a first pass indexing the manuals has been completed. 8734 documents were ingested. this pass just did basic text extraction on PDFs. another couple things TODO:

1. re-run the indexing with OCR enabled (Tasks -> Analyze Your Documents -> Extract Text (_and enable Do you want to extract text also from images and PDFs?, which is not on by default)
2. do an analysis run that does named entity extraction (Tasks -> Analyze Your Documents -> Find people, organizations, locations and email addresses). TODO: does this require first installing an NER plugin? there are options in Settings)