A repository of SPARQL queries to be used by LDS-PDI.
The queries are in a custom format where information about the parameters is given in sparql commentaries.
Query files must have the extension .arq
and be formatted as in this example:
#QueryScope=General
#QueryReturnType=Table
#QueryResults=A table containing the Id and matching literal for the given query and language tag with the given limit
#QueryParams=L_NAME,LG_NAME,I_LIM
#QueryUrl=/Res_withFacet?L_NAME=("mkhan chen" AND ("'od zer" OR "ye shes"))&LG_NAME=bo-x-ewts&I_LIM=100
#param.L_NAME.type=string
#param.L_NAME.langTag=LG_NAME
#param.L_NAME.isLucene=true
#param.L_NAME.example=("'od zer" OR "ye shes")
#param.I_LIM.type=int
#param.I_LIM.desc=the maximum number of results
#output.?s.type=URI
#output.?s.desc=the resource URI
#output.?f.type=URI
#output.?f.desc=the resourceType URI of the resource
#output.?lit.type=string
#output.?lit.desc=the label/pref. label of the resource
select distinct ?s ?f ?lit
WHERE {
{ (?s ?sc ?lit) text:query ( skos:prefLabel ?L_NAME ) . }
union
{ (?s ?sc ?lit) text:query ( skos:altLabel ?L_NAME ) . }
union
{ (?t ?sc ?lit) text:query ( rdfs:label ?L_NAME ) . ?s ?p ?t } .
?s a ?f .
}
limit ?I_LIM
These parameters must be defined for all the queries, at the top of the file:
Parameter | Values | Description |
---|---|---|
QueryScope |
the scope of the query | used on the index page to group the queries Scope examples are Work, General, Etexts, Volume, etc... |
QueryReturnType |
Table or Graph |
Indicates if the (resp.) SELECT or CONSTRUCT keyword is used. |
QueryResults |
Example: "all the works of a given author" | a description of the returned results |
QueryParams |
the names of all the params expected by the query | comma separated list of the name of the params expected by the query |
QueryOptParams |
the names of optional params accepted by the query- these param follow the same naming conventions as normal param but they can be ommitted in which case they are not bopund to any value in the query | comma separated list of the name of the optional params expected by the query |
QueryUrl |
any Url using that template | eventually used on the index page to run a valid example |
QueryLimit |
an integer | overrides the default (500) allowed number of results |
Then information must be given on the input variables (variables that are supposed to be changed by the user).
The format is made so that lds-pdi can perform a parameter evaluation in order to prevent SPARQL injection. It understands 3 formats of variables in the SPARQL query itself:
- Literal: each string literal parameter name must be prefixed by
L_
(ex :L_NAME
) - Literal lang tag: each literal parameter can be associated with a lang tag using a parameter prefixed by
LG_
(Ex : if you wantL_FOO
to be associated with thebar
lang tag, you can addLG_FOO=bar
to your request and declare it in theQueryParams
section of your template). - Integer: each integer parameter name must be prefixed by
I_
(ex :I_LIM
) - Resource: each resource URI parameter must be prefixed
R_
(ex :R_RES
) - Boolean: each boolean parameter must be prefixed
B_
- GYear: each resource URI parameter must be prefixed
GY_
The special argument B_RIC
indicates that the query should be geo-restricted.
For optional params only, the variable suffixed _bound
will be replaced with a boolean indicating the presence of the param.
For each input variable, optional or not, there must be definitions in the form:
#{metadata_type}.{variable_name}.{key}={value}
where:
{metadata_type}
isparam
(see output variables for another possible value){variable_name}
is the name of the variable as appearing in the SPARQL query (using the conventions indicated above){key}
and{value}
are the metadata
The possible key / values are:
Key | Value | Note |
---|---|---|
langTag |
a BCP47 lang tag value | only for Literal variables |
isLucene |
true or false |
only for Literal variables, indicates if the value should be parsed as a Lucene query (for text:query Jena extension) |
example |
an example of value | optional |
type |
boolean , integer , string , URI |
gives the type of variable (is it optional?) |
subType |
an indication of subtype | optional, only for Resource variables |
desc |
a description of the value | optional |
langTag |
the value of another variable | this gives the {variable_name} of the lang tag used in the SPARQL query |
Note that you do not have to create entries for the lang tag variables as long as they are indicated as a langTag
value.
Here is an example:
#param.L_NAME.type=string
#param.L_NAME.langTag=LG_NAME
#param.L_NAME.isLucene=true
#param.L_NAME.example=("'od zer" OR "ye shes")
#param.L_NAME.desc=the name to query
For SELECT
queries where the result contains values of SPARQL variables, the template also requires a small set of information about the returned variables. It is in the same format as the input variable, except:
{metadata_type}
isoutput
- there is no need to respect the conventions for variable names
- only
desc
andtype
are valid metadata keys
Additional rule : Filter on variables should be the last ones in a query
ex:
FILTER (contains(?root_name, ?NAME ))
FILTER ((contains(?comment_type, "commentary" ))
will fail because it is subject to an injection attack while :
FILTER ((contains(?comment_type, "commentary" ))
FILTER (contains(?root_name, ?NAME ))
will be considered safe.
#QueryScope=General
#QueryReturnType=Table
#QueryResults=A table containing the Id and matching literal for the given query and language tag with the given limit
#QueryParams=L_NAME,LG_NAME,I_LIM
#QueryUrl=/Res_byName?L_NAME=("mkhan chen" AND ("'od zer" OR "ye shes"))&LG_NAME=bo-x-ewts&I_LIM=100
select distinct ?s ?lit
WHERE {
{ (?s ?sc ?lit) text:query ( skos:prefLabel ?L_NAME ) . }
union
{ (?s ?sc ?lit) text:query ( rdfs:label ?L_NAME ) . }
} limit ?I_LIM
#QueryScope=Person
#QueryReturnType=Table
#QueryResults=All the detailed admin info (notes, status, log entries) about the person data
#QueryParams=R_RES
#QueryUrl=/Person_adminDetails?R_RES=P1583
select distinct
?ID
?preferredName
?y ?noteRef ?note_value ?admin_prop ?admin_ref ?log_value ?git ?status
where {
{
?ID skos:prefLabel ?preferredName ;
adm:gitRevision ?git;
adm:status ?status .
Filter(?ID=?R_RES)
}
UNION {
OPTIONAL{ ?ID :note ?noteRef }.
?noteRef ?y ?note_value .
Filter(?ID=?R_RES)
}
UNION {
OPTIONAL{ ?ID adm:logEntry ?admin_ref }.
?admin_ref ?admin_prop ?log_value .
Filter(?ID=?R_RES)
}
}
You can limit the text search output (i.e the number of matches returned by jena Text) using a LI_XXX param where XXX is the name of the text variable (for instance: L_NAME
)
This naming convention is illustrated by that query string:
/Name_of_the_template?L_XXX=("མཁན་ཆེན་"+AND+("འོད་ཟེར་"+OR+"ཡེ་ཤེས་"))&LG_XXX=bo&LI_XXX=25
if no LI_XXX
is specified, then the limit is set to a default value definied in the application properties:
text_query_limit=1000
These SPARQL queries are Copyright Buddhist Digital Resource Center 2017-2019, and are provided under the Apache License 2.0.