Parsing strict again #188 isaacs/sax-js#206

UUDigitalHumanitieslab · Dec 7, 2018 · e2950bb · e2950bb
1 parent edb7c55
commit e2950bb
Show file tree

Hide file tree

Showing 5 changed files with 80 additions and 18 deletions.
diff --git a/web-ui/package-lock.json b/web-ui/package-lock.json
diff --git a/web-ui/package.json b/web-ui/package.json
@@ -41,7 +41,7 @@
         "hammerjs": "^2.0.8",
         "jquery": "^3.3.1",
         "jquery-ui": "^1.12.1",
-        "lassy-xpath": "^0.3.8",
+        "lassy-xpath": "^0.3.9",
         "lodash": "^4.17.10",
         "ngx-clipboard": "^10.0.0",
         "node-sass": "^4.9.2",

diff --git a/web-ui/src/app/components/step/xpath-input/xpath-input.component.ts b/web-ui/src/app/components/step/xpath-input/xpath-input.component.ts
@@ -1,7 +1,7 @@
 import { Component, EventEmitter, Input, OnChanges, Output } from '@angular/core';
 
 import { Subject, Subscription } from 'rxjs';
-import { debounceTime } from 'rxjs/operators';
+import { debounceTime, distinctUntilChanged } from 'rxjs/operators';
 
 import { StepComponent } from '../step.component';
 import { MacroService, ValueEvent, ReconstructorService, ExtractinatorService, PathVariable } from 'lassy-xpath/ng';
@@ -36,16 +36,18 @@ export class XpathInputComponent extends StepComponent implements OnChanges {
         super();
 
         macroService.loadDefault();
-        this.subscriptions.push(this.valueSubject.pipe(debounceTime(500)).subscribe(value => {
-            let paths: PathVariable[] = null;
-            try {
-                paths = extractinatorService.extract(value);
-                this.treeXml = reconstructorService.construct(paths, value);
-            } catch (err) {
-                // probably some malformed input
-                console.error(err);
-            }
-        }));
+        this.subscriptions.push(this.valueSubject.pipe(
+            debounceTime(500),
+            distinctUntilChanged()).subscribe(value => {
+                let paths: PathVariable[] = null;
+                try {
+                    paths = extractinatorService.extract(value);
+                    this.treeXml = reconstructorService.construct(paths, value);
+                } catch (err) {
+                    // probably some malformed input
+                    console.error(err);
+                }
+            }));
     }
 
     valid = false;

diff --git a/web-ui/src/app/services/xml-parse.service.spec.ts b/web-ui/src/app/services/xml-parse.service.spec.ts
@@ -0,0 +1,62 @@
+import { TestBed, inject, async } from '@angular/core/testing';
+
+import { XmlParseService } from './xml-parse.service';
+
+describe('XmlParseService', () => {
+    beforeEach(() => {
+        TestBed.configureTestingModule({
+            providers: [
+                XmlParseService
+            ]
+        });
+    });
+
+    it('should be created', inject([XmlParseService], (service: XmlParseService) => {
+        expect(service).toBeTruthy();
+    }));
+
+    it('should parse metadata', async(inject([XmlParseService], async (service: XmlParseService) => {
+        const parsed = await service.parse(`<metadata>
+        <meta type="text" name="session" value="306_Busverhaalwithout"/>
+        <meta type="text" name="charencoding" value="UTF8"/>
+        <meta type="text" name="media" value="306_Busverhaal, video"/>
+        <meta type="text" name="transcriber" value="Margriet"/>
+        <meta type="text" name="speaker" value="CHI"/>
+        <meta type="text" name="origutt" value="en eind goed &amp;+alt al goed is hij nu gewoon een bus."/>
+        <meta type="text" name="language" value="nld"/>
+        <meta type="text" name="corpus" value="change_corpus_later"/>
+        <meta type="text" name="age" value="10;11."/>
+        <meta type="int" name="months" value="131"/>
+        <meta type="text" name="sex" value="male"/>
+        <meta type="text" name="group" value="Controle"/>
+        <meta type="text" name="role" value="Target_Child"/>
+        <meta type="int" name="uttid" value="19"/>
+    </metadata>`);
+        expect(parsed.metadata).toBeTruthy();
+        expect(parsed.metadata.meta).toBeTruthy();
+        expect(parsed.metadata.meta.length).toBe(14);
+        expect(parsed.metadata.meta[0].$.name).toBe('session');
+        expect(parsed.metadata.meta[4].$.value).toBe('CHI');
+    })));
+
+    it('should parse attribute nodes', async(inject([XmlParseService], async (service: XmlParseService) => {
+        const parsed = await service.parse(`<node cat="top"><node varName="$node" cat="">
+        <node varName="$node1" rel="su" pt="vnw">
+        </node>
+        <node varName="$node2" rel="hd" pt="ww">
+        </node>
+        <node varName="$node3" rel="predc" cat="np">
+            <node varName="$node4" rel="det" pt="lid">
+            </node>
+            <node varName="$node5" rel="hd" pt="n">
+            </node>
+        </node>
+    </node></node>`);
+        expect(parsed.node).toBeTruthy();
+        expect(parsed.node.node).toBeTruthy();
+        expect(parsed.node.node.length).toBe(1);
+        expect(parsed.node.node[0].$).toEqual({ varName: '$node', cat: '' });
+        expect(parsed.node.node[0].node.length).toBe(3);
+        expect(parsed.node.node[0].node[2].$).toEqual({ varName: '$node3', rel: 'predc', cat: 'np' });
+    })));
+});
diff --git a/web-ui/src/app/services/xml-parse.service.ts b/web-ui/src/app/services/xml-parse.service.ts
@@ -4,9 +4,7 @@ import { Parser } from 'xml2js';
 
 @Injectable()
 export class XmlParseService {
-    private parser = new Parser({
-        strict: false
-    });
+    private parser = new Parser();
 
     parse(xml: string): Promise<any> {
         return new Promise<any>((resolve, reject) => {