openownership / bodskit Goto Github PK

Would be good to have a bit of detail in the docs about the meaning of command outputs where it might not be obvious. For example on the mapping-sheet page, explain what each of the columns in the outputted CSV represents:

• section
• path
• title
• description
• type
• range
• values
• links
• deprecated
• deprecationNotes

Range should indicate the number of values required by a field.

For example, looking at person-statement.json, when the mapping-sheet command is used one would expect that - of the top level fields - only statementID and statementType would have range 1..1 (i.e. a single value is required). But optional like birthDate and deathDate are also given a range of 1..1 .

(Requirements-wise, could we please indicate sub-property ranges with brackets around the range value. So, for example we would end up with:

,identifiers,Identifiers,"One or more official identifiers for this perrson. Where available, official registration numbers should be provided.",array,0..n,,,,
,identifiers,Identifier,An identifier that has been assigned to this entity. The scheme or list from which the identifier is drawn should be declared.,object,,,,,
identifiers,identifiers/id,ID,The identifier for this entity as provided in the declared scheme.,string,(0..1),,,,
identifiers,identifiers/scheme,Scheme,"For entity statements, the scheme should be a entry from the org-id.guide codelist. For person statements, recognised values include 'passport', 'internal' and 'id-card'.",string,(0..1),,http://www.org-id.guide,,
identifiers,identifiers/schemeName,Scheme name,,string,(0..1),,,,
identifiers,identifiers/uri,URI,Where this identifier has a canonical URI this may be included,string,(0..1),uri,https://en.wikipedia.org/wiki/Uniform_Resource_Identifier,,
)

I noticed there were only about six lines different, so I merged those changes into the next version of OCDS Kit, while making changes to support OC4IDS. The command also works with SEDL. We can try to just update one bit of code instead of copying repositories.

(Note: You need to use --order-by path and --infer-required to get the same output as the test fixture. Drop --infer-required to fix #9.)

Similarly, ocdskit schema-report now also supports BODS with only stylistic changes in output, which can easily be changed if needed.

pip install csvkit
cd path/to/data-standard
csvstack --filenames -n codelist schema/codelists/* | sed s/\.csv// | in2csv -f csv 

Produces the same file.

Note: | sed s/\.csv// removes the .csv extension from the first column. | in2csv -f csv ensures all rows have the same number of columns. Neither is strictly necessary for the use case.

Noted by @rhiaro in #6

FYI, IATIKit links back to OCDSKit https://github.com/pwyf/iatikit