tethys.frontend/src/services/dataset.service.ts

299 lines
14 KiB
TypeScript

import api from "../api/api";
// import { Observable, of } from "rxjs";
import { Observable } from "rxjs";
import { tap, map } from "rxjs/operators";
import { Dataset, DbDataset, Suggestion } from "@/models/dataset";
import { HitHighlight, OpenSearchResponse, SolrResponse } from "@/models/headers";
import { ActiveFilterCategories } from "@/models/solr";
import { VUE_API } from "@/constants";
import { deserialize } from "class-transformer";
class DatasetService {
/**
* Fetch data from the OpenSearch endpoint with fuzzy search enabled.
* This function allows for misspellings in the search term and boosts
* the relevance of matches in the title, author, and subject fields.
*
* @param {string} searchTerm - The search term to query.
*/
/* https://tethys.at/solr/rdr_data/select?&0=fl%3Did%2Clicence%2Cserver_date_published%2Cabstract_output%2Cidentifier%2Ctitle_output%2Ctitle_additional%2Cauthor%2Csubject%2Cdoctype&q=%2A
&q.op=or&defType=edismax&qf=title%5E3%20author%5E2%20subject%5E1&indent=on&wt=json&rows=10&start=0&sort=server_date_published%20desc&facet=on&json.facet.language=%7B%20type%3A%20%22
terms%22%2C%20field%3A%20%22language%22%20%7D&json.facet.subject=%7B%20type%3A%20%22terms%22%2C%20field%3A%20%22subject%22%2C%20limit%3A%20-1%20%7D&json.facet.year=%7B%20type%3A%20%22
terms%22%2C%20field%3A%20%22year%22%20%7D&json.facet.author=%7B%20type%3A%20%22terms%22%2C%20field%3A%20%22author_facet%22%2C%20limit%3A%20-1%20%7D
*/
// private openSearchUrl = "http://opensearch.geoinformation.dev/tethys-records/_search";
// private openSearchUrl = "http://192.168.21.18/tethys-records/_search";
// public searchTerm(term: string): Observable<Dataset[]> {
public searchTerm(term: string, openCore: string, openHost: string): Observable<{ datasets: Dataset[], highlights: HitHighlight[] }> {
// OpenSearch endpoint
const host = "https://" + openHost; // When using geoinformation.dev
// const host = "http://" + openHost; // When using local OpenSearch dev endpoint
const path = "/" + openCore + "/_search";
const base = host + path;
/**
* The match query used for title, author, and subjects fields is case-insensitive by default. The standard analyzer is typically used, which lowercases the terms.
* The wildcard query is case-sensitive by default. To make it case-insensitive, it is needed to use a lowercase filter */
const lowercaseTerm = term.toLowerCase(); // Lowercase the search term
const body = {
query: {
bool: {
should: [
{ match: { title: { query: term, fuzziness: "AUTO", boost: 3 } } },
{ match: { author: { query: term, fuzziness: "AUTO", boost: 2 } } },
{ match: { subjects: { query: term, fuzziness: "AUTO", boost: 1 } } }, // In SOLR is "subject"!
{ wildcard: { title: { value: `${lowercaseTerm}*`, boost: 3 } } },
{ wildcard: { author: { value: `${lowercaseTerm}*`, boost: 2 } } },
{ wildcard: { subjects: { value: `${lowercaseTerm}*`, boost: 1 } } } // In SOLR is "subject"!
],
minimum_should_match: 1
}
},
size: 10,
from: 0,
// sort: [{ server_date_published: { order: "desc" } }],
sort: [{ _score: { order: "desc" } }], // Sort by _score in descending order
track_scores: true, // This ensures "_score" is included even when sorting by other criteria. Otherwise the relevance score is not calculated
aggs: {
language: { terms: { field: "language.keyword" } },
subjects: { terms: { field: "subjects.keyword", size: 10 } } // In SOLR is "subject"!
},
highlight: {
fields: {
title: {},
author: {},
subjects: {}
}
}
};
// Make API call to OpenSearch and return the result
/**
* When a POST request is made to the OpenSearch server using the api.post<OpenSearchResponse> method, the response received from OpenSearch is an object that includes various details about the search results.
* One of the key properties of this response object is _source, which is an array of documents (datasets) that match the search criteria.
* It is used the pipe method to chain RxJS operators to the Observable returned by api.get. The map operator is used to transform the emitted items of the Observable.
*/
return api.post<OpenSearchResponse>(base, body).pipe(
tap(response => console.log("OpenSearchResponse:", response)), // Log the complete response
// tap(response => console.log("Aggre:", response.aggregations?.subjects.buckets[0])), // log the first subject of the array of subjects returned
// tap(response => console.log("Hits:", response.hits)), // log the first subject of the array of subjects returned
// map(response => response.hits.hits.map(hit => hit._source))
map(response => ({
datasets: response.hits.hits.map(hit => hit._source),
highlights: response.hits.hits.map(hit => hit.highlight)
}))
);
}
// For the autocomplete search. Method to perform a search based on a term
public searchTerm_SOLR(term: string, solrCore: string, solrHost: string): Observable<Dataset[]> {
// SOLR endpoint
const host = "https://" + solrHost;
const path = "/solr/" + solrCore + "/select?";
const base = host + path;
//const fields = 'id,server_date_published,abstract_output,title_output,title_additional,author,subject'; // fields we want returned
const fields = [
"id",
"licence",
"server_date_published",
"abstract_output",
"title_output",
"title_additional",
"author",
"subject",
"doctype",
].toString();
const qfFields = "title^3 author^2 subject^1";
const q_params = {
"0": "fl=" + fields,
q: term + "*",
defType: "edismax",
qf: qfFields,
indent: "on",
wt: "json",
};
// Make API call to Solr and return the result
/**
* When a GET request is made to the Solr server using the api.get<SolrResponse> method, the response received from Solr is an object that includes various details about the search results.
* One of the key properties of this response object is docs, which is an array of documents (datasets) that match the search criteria.
* It is used the pipe method to chain RxJS operators to the Observable returned by api.get. The map operator is used to transform the emitted items of the Observable.
*/
const stations = api.get<SolrResponse>(base, q_params).pipe(map((res: SolrResponse) => res.response.docs));
return stations;
}
/* E.g. Only one facet => Author: Coric, Stjepan (16)
https://tethys.at/solr/rdr_data/select?&0=fl%3Did%2Clicence%2Cserver_date_published%2Cabstract_output%2Cidentifier%2Ctitle_output%2Ctitle_additional%2Cauthor%2Csubject%2Cdoctype&q=%2A
&q.op=or&defType=edismax&qf=title%5E3%20author%5E2%20subject%5E1&indent=on&wt=json&rows=10&fq=author%3A%28%22Coric%2C%20Stjepan%22%29&start=0&sort=server_date_published%20desc&facet=on
&json.facet.language=%7B%20type%3A%20%22terms%22%2C%20field%3A%20%22language%22%20%7D
&json.facet.subject=%7B%20type%3A%20%22terms%22%2C%20field%3A%20%22subject%22%2C%20limit%3A%20-1%20%7D
&json.facet.year=%7B%20type%3A%20%22terms%22%2C%20field%3A%20%22year%22%20%7D
&json.facet.author=%7B%20type%3A%20%22terms%22%2C%20field%3A%20%22author_facet%22%2C%20limit%3A%20-1%20%7D */
/**
* This method performs a faceted search on a Solr core. Faceted search allows the user to filter search results based on various categories (facets)
*/
public facetedSearch(
suggestion: Suggestion | string,
activeFilterCategories: ActiveFilterCategories,
solrCore: string,
solrHost: string,
start?: string, // Starting page
): Observable<SolrResponse> {
// console.log("face:", suggestion);
// console.log(activeFilterCategories);
// console.log(solrCore);
// console.log(solrHost);
// console.log(start);
// Construct Solr query parameters
const host = "https://" + solrHost;
const path = "/solr/" + solrCore + "/select?";
const base = host + path;
const fields = [
"id",
"licence",
"server_date_published",
"abstract_output",
"identifier",
"title_output",
"title_additional",
"author",
"subject",
"doctype",
].toString();
// Determine search term, query operator, and query fields based on the suggestion type. Depending on whether suggestion is a string or a Suggestion object, it constructs the search term and query fields differently.
let term, queryOperator, qfFields;
if (typeof suggestion === "string") { // f suggestion is a string, it appends a wildcard (*) for partial matches.
term = suggestion + "*";
queryOperator = "or";
qfFields = "title^3 author^2 subject^1";
} else if (suggestion instanceof Suggestion) { // If suggestion is a Suggestion object, it forms a more specific query based on the type and value of the suggestion.
term = suggestion.type + ':"' + suggestion.value + '"';
queryOperator = "and";
qfFields = undefined;
}
// Set default value for start if not provided
if (start === undefined) start = "0";
// Construct filter fields based on active filter categories
const filterFields = new Array<string>();
if (Object.keys(activeFilterCategories).length > 0) {
/* Declare variable prop with a type that is a key of the activeFilterCategories. The 'keyof typeof' activeFilterCategories type represents all possible keys
that can exist on the activeFilterCategories --> prop can only be assigned a value that is a key of the activeFilterCategories object */
let prop: keyof typeof activeFilterCategories;
for (prop in activeFilterCategories) {
const filterItems = activeFilterCategories[prop];
filterItems.forEach(function (value: string) {
filterFields.push(prop + ':("' + value + '")');
// e.g. Array [ 'subject:("Vektordaten")', 'author:("GeoSphere Austria, ")' ]
});
}
}
// https://solr.apache.org/guide/8_4/json-request-api.html
// Construct Solr query parameters
const q_params = {
"0": "fl=" + fields,
q: term,
"q.op": queryOperator,
defType: "edismax",
qf: qfFields,
// df: "title",
indent: "on",
wt: "json",
rows: 10,
// fq: ["subject:Steiermark", "language:de"],
fq: filterFields,
start: start,
sort: "server_date_published desc",
facet: "on",
// "facet.field": "language",
"json.facet.language": '{ type: "terms", field: "language" }',
"json.facet.subject": '{ type: "terms", field: "subject", limit: -1 }',
"json.facet.year": '{ type: "terms", field: "year" }',
"json.facet.author": '{ type: "terms", field: "author_facet", limit: -1 }',
};
/* E.g.
{"0":"fl=id,licence,server_date_published,abstract_output,identifier,title_output,title_additional,author,subject,doctype","q":"*","q.op":"or","defType":"edismax",
"qf":"title^3 author^2 subject^1",
"indent":"on","wt":"json","rows":10,
"fq":["subject:(\"Vektordaten\")","author:(\"GeoSphere Austria, \")"],
"start":"0","sort":"server_date_published desc","facet":"on",
"json.facet.language":"{ type: \"terms\", field: \"language\" }",
"json.facet.subject":"{ type: \"terms\", field: \"subject\", limit: -1 }",
"json.facet.year":"{ type: \"terms\", field: \"year\" }",
"json.facet.author":"{ type: \"terms\", field: \"author_facet\", limit: -1 }"}
*/
// console.log(JSON.stringify(q_params));
// Make API call to Solr and return the result
const stations = api.get<SolrResponse>(base, q_params);
return stations;
}
// Method to fetch years
public getYears(): Observable<string[]> {
const host = VUE_API;
const path = "/api/years";
const base = host + path;
const years = api.get<string[]>(base);
return years;
}
// Method to fetch documents for a specific year
public getDocuments(year: string): Observable<Array<DbDataset>> {
const host = VUE_API;
const path = "/api/sitelinks/" + year;
const base = host + path;
const documents: Observable<DbDataset[]> = api.get<Array<DbDataset>>(base);
return documents;
}
// Method to fetch a dataset by its ID
public getDataset(id: number): Observable<DbDataset> {
const host = VUE_API;
const path = "/api/dataset/" + id;
const apiUrl = host + path;
const dataset = api.get<DbDataset>(apiUrl).pipe(map((res) => this.prepareDataset(res)));
return dataset;
}
// Method to fetch a dataset by its DOI
public getDatasetByDoi(doi: string): Observable<DbDataset> {
const host = VUE_API;
const path = "/api/dataset/10.24341/tethys." + doi;
const apiUrl = host + path;
const dataset = api.get<DbDataset>(apiUrl).pipe(map((res) => this.prepareDataset(res)));
return dataset;
}
// Method to prepare dataset object
private prepareDataset(datasetObj: DbDataset): DbDataset {
const dataset = deserialize<DbDataset>(DbDataset, JSON.stringify(datasetObj));
dataset.url = document.documentURI;
return dataset;
}
}
export default new DatasetService();