• Get e.g. Topic.txt (a list of all Subject:topics) from JIRA
• Get a report on the encoding file -bi Topic.txt
• Convert to UTF-8: iconv -f ISO-8859-15 -t UTF-8 Topic.txt > topic.txt
• Convert to csv: cat topic.txt > topic.csv
• Cut columns we don't need via csvkit's csvcut: csvcut -t -c ark,topic topic.csv > new_topic.csv
• Convert to tsv: cat new_topic.csv | sed 's/,/\t/g' > topic.tsv or cat new_topic.csv > topic.tsv
• Run script that creates all the csv files: python less_FAST_times.py
• Run the concatenate script: python concatenate.py
• Open result concatenated.csv in OpenRefine
• Check for any obvious anomalies via faceting, clusters
• Run reconciliation service for FAST by cd into the fast-reconcile repo and run: python reconcile.py
• Select options on clean_labels column, then select reconcile: follow prompts
• Review matches. When satisfied, extract label, URI, marc tag into new columns
• Export to Excel for review and splitting off of unmatched terms
• Repeat reconciliation for unmatched terms until all have been attempted
• Once review of the Excel has occurred and sent back for more recon, get back into csv:
  • in2csv topic.xlsx > topic.csv # Requires csvkit
• Repeat earlier steps for recon as necessary

Help from the Refine GitHub Wiki

Some VIAF authorities have Wikidata URIs in them. Once we have those VIAF URIs, we can use Refine's "Create a column by fetching URLs" on this column.

The data is in a JSON version of the authorities, which is appended to each URI as /justlinks.json

Therefore, the expression, using GREL, will be value '/justlinks.json'.

We get back a bunch of JSON in this column, which we will need to parse. To parse out the Wikidata IDs, the GREL will be grel: value.parseJson()["WKP"]

Using jython in OpenRefine's "Create column based on this column" on the VIAF JSON:

import re
g = re.search('.*(https:\/\/en\.wikipedia\.org\/wiki\/\w ). ', value)
return g.group(1)