process.txt

* run software mention service:
cd ~/grobid/software-mentions
./gradlew run

* start a mongo instance if not automatically started 
sudo service mongod start

* when the db is filled (here mongo db called softcite-cord19-scibert), we can check the relevant collection with the mongo console 

mongo
use softcite-cord19-scibert
db.references.count()
db.documents.count()
db.annotations.count()

db.annotations.aggregate( [ {$unwind: "$software-name.normalizedForm"}, {$sortByCount: "$software-name.normalizedForm" } ] )

* install/run the client

source env/bin/activate
python3 software_mentions_client.py --data-path /media/lopez/store/cord-19/data/ --config my_config.json 

* results are stored in mongodb but also written along with the full texts

> find /media/lopez/store/cord-19/data/ -name *.software.json | wc -l

to clean:

> find /media/lopez/store/cord-19/data3/ -name "*.software.json" -type f -delete

* export from mongodb and further ingestion in the KB:

> mongoexport -d softcite-cord19-scibert -c annotations -o /media/lopez/store/cord-19/scibert-db2/softcite-cord19-annotations.json
> mongoexport -d softcite-cord19-scibert -c documents -o /media/lopez/store/cord-19/scibert-db2/softcite-cord19-documents.json
> mongoexport -d softcite-cord19-scibert -c references -o /media/lopez/store/cord-19/scibert-db2/softcite-cord19-references.json