Its requirede to have our Ontology loaded into GraphDB under the repository name "ogc"
- jupyter nbextension enable --py widgetsnbextension
-Showing Codon Distribution for a species using its SpeciesID with a custom Plot
---> useage: get_codon_graph(SpeciesID)
-Showing Similarities between two Species using a cusotm Plot and overall Similarity
---> useage: compare(SpeciesID1, SpeciesID2)
-Getting an organisms Lineage and its indices using the SpeciesID
---> useage: get_lineage(SpeciesID)
#Automatic Querying and SPARQL-Query generator using a Form
The last part of the Notebook contains a Form that lets a user Query for an organism in a very simple way.
-> after the form is filled do not run the Cell again , otherwise it resets
How to use:
-> Indicate how many statements the query should have (statement = (p,s,o))
-> Indicate how many variables you want to query for
run the next cell
-> Indicate if you want to count something and if yes what
---> this is a dropdown menu where you can select any variable generated
-> For each line indicate the Predicate, Subject and Object (the predicate stays the same for all lines)
---> if you search for a lable the input is not case sensetive and accepts small errors
run next cell
-> Indicate if there should be a limit for the output, if it stays 0 it wont be used as LIMIT
-> Indicate if on which variable the output should be sorted and of which order
--> The query will return the results automatically once the next cell is run and the GraphDB repository has been set-up