-
Notifications
You must be signed in to change notification settings - Fork 0
Home
Bill Baumgartner edited this page Apr 19, 2023
·
5 revisions
Use the following command to see statistics (mainly sentence counts) for an annotation project specified by the biolink_association parameter.
docker run --rm -v [path_to_repo]:/home/input ucdenverccp/elastic2brat:0.4 stats -b [biolink_association]
where,
- [path_to_repo] is the local base path to this
translator-relation-extraction
git repository - [biolink_association] is the Biolink association associated with the annotation project, e.g.
bl_chemical_to_gene
Use the following command to create a new batch of sentences to be annotated. The command will query an Elasticsearch instance to retrieve appropriate sentences for the annotation project based on the specified Biolink association.
docker run --rm -v [path_to_repo]:/home/input ucdenverccp/elastic2brat:0.4 batch -b [biolink_association] -a [annotator] -n [sentence_count] -u [elastic_url] -p [elastic_port] -k [elastic_api_key] -s [subject-idf-threshold] -o [object-idf-threshold] -g [sentences-per-page]
where,
- [path_to_repo] is the local base path to this
translator-relation-extraction
git repository - [biolink_association] is the Biolink association associated with the annotation project, e.g.
bl_chemical_to_gene
- [annotator] is the name/key for the annotator assigned to the batch being created. Note, this will end up as a directory name so it should be a single token (no spaces)
- [sentence_count] is the number of sentences to include in this batch
- [elastic_url] is the URL to the Elasticsearch server -- do not include
http://
- [elastic_port] is the port for the Elasticsearch server
- [elastic_api_key] is the API key for the Elasticsearch server
- [subject-idf-threshold] OPTIONAL threshold for filtering the subject entities based on inverse document frequency. By default, no filtering is performed.
- [object-idf-threshold] OPTIONAL threshold for filtering the object entities based on inverse document frequency. By default, no filtering is performed.
- [sentences-per-page] OPTIONAL parameter that determines how many sentences are included in each BRAT file (default = 20), i.e., how many sentences are presented to the annotator on each page in the BRAT UI.