wiki-entity 0.5.2 → 0.5.6
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +2 -0
- package/lib/index.js +25 -30
- package/lib/simpleEntity/convertToSimpleEntity.js +2 -0
- package/lib/simpleEntity/simpleEntity.js +1 -1
- package/lib/wikidata/api.js +3 -0
- package/lib/wikidata/get_entity_types.js +9 -7
- package/lib/wikidata/index.js +1 -1
- package/lib/wikidata/simplify_entity.js +7 -10
- package/lib/wikipedia/api.js +4 -4
- package/package.json +1 -1
- package/types/simpleEntity/simpleEntity.d.ts +1 -0
- package/types/types.d.ts +1 -0
package/README.md
CHANGED
|
@@ -59,6 +59,7 @@ type WikiEntity = {
|
|
|
59
59
|
types?: string[];
|
|
60
60
|
redirects?: string[];
|
|
61
61
|
categories?: string[];
|
|
62
|
+
redirectsToId?: string;
|
|
62
63
|
}
|
|
63
64
|
|
|
64
65
|
type WikidataProperty = {
|
|
@@ -113,6 +114,7 @@ export type SimpleEntity = {
|
|
|
113
114
|
countryCodes?: string[]
|
|
114
115
|
data?: SimpleEntityData
|
|
115
116
|
categories?: string[]
|
|
117
|
+
redirectsToId?: string;
|
|
116
118
|
}
|
|
117
119
|
|
|
118
120
|
```
|
package/lib/index.js
CHANGED
|
@@ -1,4 +1,12 @@
|
|
|
1
1
|
"use strict";
|
|
2
|
+
var __awaiter = (this && this.__awaiter) || function (thisArg, _arguments, P, generator) {
|
|
3
|
+
return new (P || (P = Promise))(function (resolve, reject) {
|
|
4
|
+
function fulfilled(value) { try { step(generator.next(value)); } catch (e) { reject(e); } }
|
|
5
|
+
function rejected(value) { try { step(generator["throw"](value)); } catch (e) { reject(e); } }
|
|
6
|
+
function step(result) { result.done ? resolve(result.value) : new P(function (resolve) { resolve(result.value); }).then(fulfilled, rejected); }
|
|
7
|
+
step((generator = generator.apply(thisArg, _arguments || [])).next());
|
|
8
|
+
});
|
|
9
|
+
};
|
|
2
10
|
function __export(m) {
|
|
3
11
|
for (var p in m) if (!exports.hasOwnProperty(p)) exports[p] = m[p];
|
|
4
12
|
}
|
|
@@ -13,30 +21,21 @@ exports.simplifyEntity = simplify_entity_1.simplifyEntity;
|
|
|
13
21
|
__export(require("./simpleEntity"));
|
|
14
22
|
__export(require("./types"));
|
|
15
23
|
function getEntities(params) {
|
|
16
|
-
|
|
17
|
-
|
|
18
|
-
.
|
|
19
|
-
const entities = wikiDataEntities;
|
|
24
|
+
return __awaiter(this, void 0, void 0, function* () {
|
|
25
|
+
const lang = params.language || "en";
|
|
26
|
+
const entities = yield wikidata_1.getEntities(params);
|
|
20
27
|
const ids = Object.keys(entities).filter((id) => utils_1.isValidWikiId(id));
|
|
21
|
-
if (ids.length === 0)
|
|
22
|
-
return
|
|
23
|
-
}
|
|
24
|
-
const tasks = [];
|
|
28
|
+
if (ids.length === 0)
|
|
29
|
+
return [];
|
|
25
30
|
const wikiApi = new api_1.Api({}, params.httpTimeout);
|
|
26
|
-
|
|
27
|
-
if (params.extract) {
|
|
28
|
-
callWikiApi = true;
|
|
31
|
+
if (params.extract)
|
|
29
32
|
wikiApi.extract(params.extract);
|
|
30
|
-
|
|
31
|
-
if (params.redirects) {
|
|
32
|
-
callWikiApi = true;
|
|
33
|
+
if (params.redirects)
|
|
33
34
|
wikiApi.redirects();
|
|
34
|
-
|
|
35
|
-
if (params.categories) {
|
|
36
|
-
callWikiApi = true;
|
|
35
|
+
if (params.categories)
|
|
37
36
|
wikiApi.categories();
|
|
38
|
-
|
|
39
|
-
if (
|
|
37
|
+
const tasks = [];
|
|
38
|
+
if (entities[ids[0]] && entities[ids[0]].sitelinks) {
|
|
40
39
|
const titleIds = ids.reduce((prev, id) => {
|
|
41
40
|
if (entities[id] &&
|
|
42
41
|
entities[id].sitelinks &&
|
|
@@ -57,15 +56,13 @@ function getEntities(params) {
|
|
|
57
56
|
})
|
|
58
57
|
.then((apiResults) => apiResults.forEach((result) => {
|
|
59
58
|
const entity = entities[titleIds[result.title]];
|
|
60
|
-
|
|
59
|
+
entity.pageid = result.pageid;
|
|
60
|
+
if (params.extract)
|
|
61
61
|
entity.extract = result.extract;
|
|
62
|
-
|
|
63
|
-
if (params.redirects) {
|
|
62
|
+
if (params.redirects)
|
|
64
63
|
entity.redirects = result.redirects;
|
|
65
|
-
|
|
66
|
-
if (params.categories) {
|
|
64
|
+
if (params.categories)
|
|
67
65
|
entity.categories = result.categories;
|
|
68
|
-
}
|
|
69
66
|
})));
|
|
70
67
|
}
|
|
71
68
|
if (params.types === true || Array.isArray(params.types)) {
|
|
@@ -83,11 +80,9 @@ function getEntities(params) {
|
|
|
83
80
|
}
|
|
84
81
|
});
|
|
85
82
|
}
|
|
86
|
-
|
|
87
|
-
|
|
88
|
-
|
|
89
|
-
return Object.keys(resultEntities)
|
|
90
|
-
.map((id) => resultEntities[id])
|
|
83
|
+
yield Promise.all(tasks);
|
|
84
|
+
return Object.keys(entities)
|
|
85
|
+
.map((id) => entities[id])
|
|
91
86
|
.filter((it) => !!it);
|
|
92
87
|
});
|
|
93
88
|
}
|
|
@@ -15,6 +15,8 @@ function convertToSimpleEntity(wikiEntity, lang, options) {
|
|
|
15
15
|
entity.description = wikiEntity.description;
|
|
16
16
|
entity.wikiPageId = wikiEntity.pageid;
|
|
17
17
|
entity.about = wikiEntity.extract;
|
|
18
|
+
if (wikiEntity.redirectsToId)
|
|
19
|
+
entity.redirectsToId = wikiEntity.redirectsToId;
|
|
18
20
|
if (wikiEntity.types) {
|
|
19
21
|
entity.types = utils_1.uniq(wikiEntity.types.filter((item) => !/:(Thing|Agent)$/.test(item)));
|
|
20
22
|
}
|
package/lib/wikidata/api.js
CHANGED
|
@@ -46,6 +46,9 @@ function getEntities(params) {
|
|
|
46
46
|
throw getError(data);
|
|
47
47
|
const entities = (data && data.entities) || {};
|
|
48
48
|
for (let id of Object.keys(entities)) {
|
|
49
|
+
const entity = entities[id];
|
|
50
|
+
if (entity.redirects && entity.redirects.to)
|
|
51
|
+
entity.redirectsToId = entity.redirects.to;
|
|
49
52
|
if (!utils_1.isValidWikiId(id) || ~Object.keys(entities[id]).indexOf("missing")) {
|
|
50
53
|
delete entities[id];
|
|
51
54
|
}
|
|
@@ -33,15 +33,17 @@ function parseTypes(types) {
|
|
|
33
33
|
return types.map((item) => item.type.value);
|
|
34
34
|
}
|
|
35
35
|
function dbpediaTypes(name) {
|
|
36
|
-
const
|
|
37
|
-
SELECT ?type
|
|
38
|
-
WHERE {
|
|
39
|
-
<http://dbpedia.org/resource/${encodeURIComponent(name.replace(/\s+/g, "_"))}> rdf:type ?type
|
|
40
|
-
}`;
|
|
36
|
+
const url = `http://dbpedia.org/resource/${name.replace(/\s+/g, "_")}`;
|
|
37
|
+
const query = `SELECT ?type WHERE { <${url}> rdf:type ?type }`;
|
|
41
38
|
return request_1.default("http://dbpedia.org/sparql", {
|
|
42
39
|
params: { query: query },
|
|
43
|
-
timeout:
|
|
44
|
-
})
|
|
40
|
+
timeout: 30 * 1000
|
|
41
|
+
})
|
|
42
|
+
.then((data) => data.results && data.results.bindings)
|
|
43
|
+
.catch((error) => {
|
|
44
|
+
console.error(`${error.message}: ${url}`);
|
|
45
|
+
return [];
|
|
46
|
+
});
|
|
45
47
|
}
|
|
46
48
|
function repairTypes(types) {
|
|
47
49
|
const personIndex = types.findIndex((item) => /:Person$/.test(item));
|
package/lib/wikidata/index.js
CHANGED
|
@@ -100,7 +100,7 @@ function exploreEntityClaims(claims, params) {
|
|
|
100
100
|
pa.forEach((pai) => {
|
|
101
101
|
const val = claims[pai.pid].values[pai.index];
|
|
102
102
|
for (var prop in item) {
|
|
103
|
-
if (~["label", "
|
|
103
|
+
if (~["label", "description"].indexOf(prop)) {
|
|
104
104
|
val[prop] = item[prop];
|
|
105
105
|
}
|
|
106
106
|
}
|
|
@@ -3,31 +3,28 @@ Object.defineProperty(exports, "__esModule", { value: true });
|
|
|
3
3
|
const simplify_claims_1 = require("./simplify_claims");
|
|
4
4
|
function simplifyEntity(lang, data, options = {}) {
|
|
5
5
|
const entity = { id: data.id };
|
|
6
|
-
if (data.pageid)
|
|
6
|
+
if (data.pageid)
|
|
7
7
|
entity.pageid = data.pageid;
|
|
8
|
-
|
|
8
|
+
if (data.redirectsToId)
|
|
9
|
+
entity.redirectsToId = data.redirectsToId;
|
|
9
10
|
if (options.labels !== false && data.labels) {
|
|
10
11
|
entity.labels = simplifyLabels(data.labels);
|
|
11
12
|
entity.label = entity.labels[lang];
|
|
12
13
|
}
|
|
13
14
|
if (options.descriptions !== false && data.descriptions) {
|
|
14
15
|
entity.description = simplifyDescriptions(data.descriptions)[lang];
|
|
15
|
-
if (!entity.description)
|
|
16
|
+
if (!entity.description)
|
|
16
17
|
delete entity.description;
|
|
17
|
-
}
|
|
18
18
|
}
|
|
19
19
|
if (options.aliases !== false && data.aliases) {
|
|
20
20
|
entity.aliases = simplifyAliases(data.aliases)[lang];
|
|
21
|
-
if (!entity.aliases)
|
|
21
|
+
if (!entity.aliases)
|
|
22
22
|
delete entity.aliases;
|
|
23
|
-
}
|
|
24
23
|
}
|
|
25
|
-
if (options.sitelinks !== false && data.sitelinks)
|
|
24
|
+
if (options.sitelinks !== false && data.sitelinks)
|
|
26
25
|
entity.sitelinks = simplifySitelinks(data.sitelinks);
|
|
27
|
-
|
|
28
|
-
if (options.claims !== false && data.claims) {
|
|
26
|
+
if (options.claims !== false && data.claims)
|
|
29
27
|
entity.claims = simplify_claims_1.simplifyClaims(data.claims);
|
|
30
|
-
}
|
|
31
28
|
return entity;
|
|
32
29
|
}
|
|
33
30
|
exports.simplifyEntity = simplifyEntity;
|
package/lib/wikipedia/api.js
CHANGED
|
@@ -33,14 +33,14 @@ class Api {
|
|
|
33
33
|
.map((id) => data.query.pages[id])
|
|
34
34
|
.map((data) => {
|
|
35
35
|
const item = {
|
|
36
|
-
pageid: data
|
|
37
|
-
title: data
|
|
36
|
+
pageid: data.pageid,
|
|
37
|
+
title: data.title
|
|
38
38
|
};
|
|
39
39
|
if (data["categories"]) {
|
|
40
|
-
item.categories = data["categories"].map((
|
|
40
|
+
item.categories = data["categories"].map((it) => it.title);
|
|
41
41
|
}
|
|
42
42
|
if (data["redirects"]) {
|
|
43
|
-
item.redirects = data["redirects"].map((
|
|
43
|
+
item.redirects = data["redirects"].map((it) => it.title);
|
|
44
44
|
}
|
|
45
45
|
if (data["extract"]) {
|
|
46
46
|
item.extract = data["extract"];
|
package/package.json
CHANGED
package/types/types.d.ts
CHANGED
|
@@ -35,6 +35,7 @@ export interface WikiEntity extends WikidataEntity {
|
|
|
35
35
|
types?: string[];
|
|
36
36
|
redirects?: string[];
|
|
37
37
|
categories?: string[];
|
|
38
|
+
redirectsToId?: string;
|
|
38
39
|
}
|
|
39
40
|
export declare type WikiEntities = PlainObject<WikiEntity>;
|
|
40
41
|
export declare type ParamClaimsType = "none" | "all" | "item" | "property";
|