Sinatra app to parse people names from structured biodiversity occurrence data, apply basic regular expressions and heuristics to disambiguate them, and then allow them to be claimed by authenticated users via ORCID. Authenticated users may also help other users that have either ORCID or Wikidata identifiers.
- February 15, 2019: Use Wikidata as optional identifier for the deceased.
- February 3, 2019: Allow download of candidate specimen records then upload of claims.
- January 26, 2019: Present citations of user specimens based on contents of GBIF data packages.
- October 12, 2018: Used Neo4j to store weighted graphs of similarly structured people names as a quick mechanism to expand the list of users' candidate specimens & then present them in greatest to least probable.
- September 29, 2018: The core, parsing component of this project has been split off into a stand-alone ruby gem, dwc_agent.
- ruby 2.5.1+
- Elasticsearch 6.2.4+
- MySQL 14.14+
- Redis 4.0.9+
- Apache Spark 2+
- Unix-based operating system to use GNU parallel
$ git clone https://github.com/dshorthouse/bloodhound.git $ cd bloodhound $ gem install bundler $ bundle install $ mysql -u root bloodhound < db/bloodhound.sql $ cp config.yml.sample config.yml # Adjust content of config.yml $ rackup -p 4567 config.ru
Steps to Import Data & Execute Parsing / Clustering
Step 1: Import Data
See the Apache Spark recipes for quickly importing into MySQL the occurrence csv from a DwC Archive downloaded from GBIF. Apache Spark is used to produce the necessary source csv files for the "Parse & Populate Agents" and "Populate Taxa" steps below.
Step 2: Parse & Populate Agents
$ RACK_ENV=production ./bin/populate_agents.rb --truncate --directory /directory-to-spark-csv-files/ $ RACK_ENV=production sidekiq -c 40 -q agent -r ./environment.rb
Step 3: Populate Taxa
$ RACK_ENV=production ./bin/populate_taxa.rb --truncate --directory /directory-to-spark-csv-files/ $ RACK_ENV=production sidekiq -c 40 -q taxon -r ./environment.rb
Also execute SQL statement at end of /bin/populate_taxa.rb script once queue is finished.
Step 4: Cluster Agents & Store in Neo4j
$ RACK_ENV=production ./bin/cluster_agents.rb --truncate --cluster $ RACK_ENV=production sidekiq -c 40 -q cluster -r ./environment.rb
Truncating a large Neo4j graph.db usually does not work. Instead, it is best to entirely delete the graph.db directory then recreate it.
Example on Mac with homebrew:
$ brew services stop neo4j $ sudo rm -rf /usr/local/opt/neo4j/libexec/data/databases/graph.db $ brew services start neo4j # recreates graph.db $ rake neo4j:migrate # recreate the constraint on graph.db
Step 5: Populate Search in Elasticsearch
$ RACK_ENV=production ./bin/populate_search.rb --rebuild
Neo4j Dump & Restore
Notes to self because I never remember how to dump from my laptop and reload onto the server. Must stop Neo4j before this can be executed.
neo4j-admin dump --database=<database> --to=<destination-path> neo4j-admin load --from=<archive-path> --database=<database> [--force]
brew services stop neo4j neo4j-admin dump --database=graph.db --to=/Users/dshorthouse/Documents/neo4j_backup/ brew services start neo4j service neo4j stop rm -rf /var/lib/neo4j/data/databases/graph.db neo4j-admin load --from=/home/dshorthouse/neo4j_backup/graph.db.dump --database=graph.db #reset permissions chown neo4j:neo4j -R /var/lib/neo4j/data/databases/graph.db service neo4j start
Replacing the database through load requires that the database first be deleted [usually found in /var/lib/neo4j/data/databases on linux machine] and then its permissions be recursively set for the neo4j:adm user:group.
Successive Data Migrations
Unfortunately, gbifIDs are not persistent. These occasionally disappear through processing at GBIF's end. As a result, claims may no longer point to an existing occurrence record and these must then be purged from the user_occurrences table. The following SQL statement can remove these with successive data imports from GBIF:
DELETE uo FROM user_occurrences uo LEFT JOIN occurrences o ON uo.occurrence_id = o.gbifID WHERE o.gbifID IS NULL;
Other considerations are how to get MySQL dump files back on the server. For fastest execution, dump separate tables into each dump file.
mysqlcheck -u root -o bloodhound mysqldump --user root --opt bloodhound agents | gzip > agents.sql.gz mysqldump --user root --opt bloodhound occurrences | gzip > occurrences.sql.gz etc...
Alternatively, mydumper and myloader can be used to accelerate these tasks.
brew install mydumper mydumper --user root --database bloodhound --table-list agents,occurrence,occurrence_recorders,occurrence_determiners,taxa,taxon_occurrences,taxon_determiners --compress --no-schemas --threads 8 --outputdir /Users/dshorthouse/bloodhound_dump apt-get install mydumper myloader --database bloodhound --user bloodhound --password --threads 8 --directory /home/dshorthouse/bloodhound_restore
The MIT License (MIT)
Copyright (c) David P. Shorthouse
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.