D-Net
/
dhp-graph-dump
4
0
Fork 0
Code Issues Pull Requests 1 Projects Releases Wiki Activity

[eoscDump] updated the schema of the dump adding the fulltext field

This commit is contained in:
Miriam Baglioni 2023-07-13 09:48:35 +02:00
parent e31b2ef6ee
commit 8f960cc5d9
3 changed files with 99 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
dump-schema/src/main/java/eu/dnetlib/dhp/eosc/model/Result.java
View File

@ -36,7 +36,7 @@ public class Result implements Serializable {
description = "Reference to a relevant research infrastructure, initiative or community (RI/RC) among those collaborating with OpenAIRE. Please see https://connect.openaire.eu")
private List<Context> context;
@JsonSchema(description = "Information about the sources from which the record has been collected")
@JsonSchema(description = "Information about the sources from which the result has been collected")
@JsonInclude(JsonInclude.Include.NON_NULL)
protected List<CfHbKeyValue> collectedfrom;
@ -142,6 +142,7 @@ public class Result implements Serializable {
@JsonSchema(description = "The set of relations associated to this result")
private List<Relation> relations;
@JsonSchema(description = "The direct link to the full-text as collected from the data source")
private List<String> fulltext;
public List<String> getFulltext() {

74
dump-schema/src/main/resources/jsonschemas/eosc_result_schema.json
View File

@ -132,6 +132,16 @@
"type": "string",
"description": "Only for results with type 'software': the URL to the repository with the source code"
},
"collectedfrom": {
"description": "Information about the sources from which the result has been collected",
"type": "array",
"items": {
"allOf": [
{"$ref": "#/definitions/CfHbKeyValue"},
{"description": "Information about the sources from which the result has been collected"}
]
}
},
"contactgroup": {
"description": "Only for results with type 'software': Information on the group responsible for providing further information regarding the resource",
"type": "array",
@ -165,9 +175,18 @@
"type": "string",
"description": "Journal issue number"
},
"issnLinking": {"type": "string"},
"issnOnline": {"type": "string"},
"issnPrinted": {"type": "string"},
"issnLinking": {
"type": "string",
"description": "The lissn"
},
"issnOnline": {
"type": "string",
"description": "The eissn"
},
"issnPrinted": {
"type": "string",
"description": "The issn"
},
"name": {
"type": "string",
"description": "Name of the journal or conference"
@ -297,6 +316,14 @@
"type": "array",
"items": {"type": "string"}
},
"fulltext": {
"description": "The direct link to the full-text as collected from the data source",
"type": "array",
"items": {
"type": "string",
"description": "The direct link to the full-text as collected from the data source"
}
},
"geolocation": {
"description": "Geolocation information",
"type": "array",
@ -386,6 +413,13 @@
},
"description": "The money spent to make this book or article available in Open Access. Source for this information is the OpenAPC initiative."
},
"collectedfrom": {
"allOf": [
{"$ref": "#/definitions/CfHbKeyValue"},
{"description": "Information about the source from which the record has been collected"}
]
},
"fulltext": {"type": "string"},
"hostedby": {
"allOf": [
{"$ref": "#/definitions/CfHbKeyValue"},
@ -551,6 +585,38 @@
"type": "string",
"description": "The name of the entity that holds, archives, publishes prints, distributes, releases, issues, or produces the resource."
},
"relations": {
"description": "The set of relations associated to this result",
"type": "array",
"items": {
"type": "object",
"properties": {
"provenance": {
"allOf": [
{"$ref": "#/definitions/Provenance"},
{"description": "The reason why OpenAIRE holds the relation "}
]
},
"reltype": {
"type": "object",
"properties": {
"name": {"type": "string"},
"type": {"type": "string"}
},
"description": "To represent the semantics of a relation between two entities"
},
"source": {
"type": "string",
"description": "The node source in the relation"
},
"target": {
"type": "string",
"description": "The node target in the relation"
}
},
"description": "The set of relations associated to this result"
}
},
"size": {
"type": "string",
"description": "Only for results with type 'dataset': the declared size of the dataset"
@ -588,4 +654,4 @@
"description": "Version of the result"
}
}
}
}

33
dump/src/main/java/eu/dnetlib/dhp/oa/graph/dump/ResultMapper.java
View File

@ -60,7 +60,8 @@ public class ResultMapper implements Serializable {
mapSubject(out, input);
out.setType(input.getResulttype().getClassid());
mapContext(communityMap, out, input);
mapFulltext(out, input);
mapCollectedfrom(out, input);
} catch (ClassCastException cce) {
return null;
}
@ -70,6 +71,18 @@ public class ResultMapper implements Serializable {
}
private static void mapCollectedfrom(Result out, eu.dnetlib.dhp.schema.oaf.Result input) {
out
.setCollectedfrom(
input
.getCollectedfrom()
.stream()
.map(cf -> CfHbKeyValue.newInstance(cf.getKey(), cf.getValue()))
.collect(Collectors.toList()));
}
private static void mapFulltext(Result out, eu.dnetlib.dhp.schema.oaf.Result input) {
if (Optional.ofNullable(input.getFulltext()).isPresent() && !input.getFulltext().isEmpty())
out.setFulltext(input.getFulltext().stream().map(ft -> ft.getValue()).collect(Collectors.toList()));
@ -251,8 +264,9 @@ public class ResultMapper implements Serializable {
input
.getInstance()
.stream()
.map(i -> getCommunityInstance(i))
.map(i -> getCommunityInstance(i, input.getResulttype().getClassid()))
.collect(Collectors.toList()));
}
}
@ -432,6 +446,8 @@ public class ResultMapper implements Serializable {
out.setContainer(c);
out.setType(ModelConstants.PUBLICATION_DEFAULT_RESULTTYPE.getClassname());
}
if (Optional.ofNullable(((Publication) input).getFulltext()).isPresent())
mapFulltext(out, input);
break;
case "dataset":
Dataset id = (Dataset) input;
@ -504,7 +520,8 @@ public class ResultMapper implements Serializable {
.orElse(null));
out.setType(ModelConstants.ORP_DEFAULT_RESULTTYPE.getClassname());
if (Optional.ofNullable(((OtherResearchProduct) input).getFulltext()).isPresent())
mapFulltext(out, input);
break;
default:
throw new NoAvailableEntityTypeException();
@ -520,7 +537,8 @@ public class ResultMapper implements Serializable {
}
private static eu.dnetlib.dhp.eosc.model.Instance getCommunityInstance(eu.dnetlib.dhp.schema.oaf.Instance i) {
private static eu.dnetlib.dhp.eosc.model.Instance getCommunityInstance(eu.dnetlib.dhp.schema.oaf.Instance i,
String resultType) {
eu.dnetlib.dhp.eosc.model.Instance instance = new eu.dnetlib.dhp.eosc.model.Instance();
setCommonValue(i, instance);
@ -529,8 +547,11 @@ public class ResultMapper implements Serializable {
.setHostedby(
CfHbKeyValue.newInstance(i.getHostedby().getKey(), i.getHostedby().getValue()));
instance.setFulltext(i.getFulltext());
if (resultType.equals("publication") ||
resultType.equals("other")) {
if (Optional.ofNullable(i.getFulltext()).isPresent())
instance.setFulltext(i.getFulltext());
}
return instance;
}