usc-isi-i2 / t2wml

Table to Wikidata Mapping Language
MIT License
22 stars 11 forks source link

Generate FIDIL JSON #559

Closed kyao closed 3 years ago

kyao commented 3 years ago

FIDIL organizes dataset at the timeseries level. Assume each FIDIL timeseries can be uniquely specified by the triple (dataset_id, variable_id, country). For the purpose of Causx, except for the time qualifier, ignore qualifiers. Each uploaded Excel/CSV file can generate one or more timeseries.

FIDIL JSON is an array of timeseries objects, each timeseries object has properties

devowit commented 3 years ago

excelFile: array of strings, where each string is the name of an uploaded file sheet: array of strings, where each string is the name of a sheet. If the uploaded file is CSV, then use the CSV filename without the suffix.

in a case with N excel files with M sheets, sheet contains MxN values?

devowit commented 3 years ago

other-user-defined-keys, string

this is a single key whose string value is concatted from multiple keys, or a filler entry for any other user keys, ie

structuredDataMapping, object with properties docID: string factorClass: string normalizer: string relevance:number mykey1: string mykeyt2: string2

devowit commented 3 years ago

attached is an example FIDIL json (from the freedom house press example)

please look at it and let me know if I've built it correctly. Thanks @kyao

example_fidil_freedomhousepress.txt

kyao commented 3 years ago

The file looks good to me. I am sending it to the CX folks to have them take a look.

kyao commented 3 years ago

One thing is all the factor classes should have the prefix: http://ontology.causeex.com/ontology/odps/ICM#

devowit commented 3 years ago

ok, i have pushed the necessary change to add the prefix