get list of movie links

ratanparai · Apr 25, 2019 · 33aba4f · 33aba4f
1 parent 5498731
commit 33aba4f
Show file tree

Hide file tree

Showing 3 changed files with 1,189 additions and 0 deletions.
diff --git a/app/src/main/java/com/ratanparai/moviedog/scrapper/DekhvhaiScrapper.kt b/app/src/main/java/com/ratanparai/moviedog/scrapper/DekhvhaiScrapper.kt
@@ -2,9 +2,11 @@ package com.ratanparai.moviedog.scrapper
 
 import com.ratanparai.moviedog.db.entity.Movie
 import org.jsoup.Jsoup
+import org.jsoup.nodes.Document
 import java.io.IOException
 import java.util.*
 import java.util.concurrent.TimeUnit
+import kotlin.collections.ArrayList
 
 class DekhvhaiScrapper {
 
@@ -81,6 +83,16 @@ class DekhvhaiScrapper {
 
     }
 
+    fun getListOfMovieLinksFromSearchResult(document: Document): List<String>{
+        val result = ArrayList<String>()
+        val elements = document.select("#tabs_i2-pane1 > div > a")
+        for (elem in elements) {
+            result.add(elem.attr("href"))
+        }
+
+        return result
+    }
+
 
     fun getOnlyTitleFromTitleAndYear(titleWithyear: String): String {
         return titleWithyear.substring(0, titleWithyear.indexOf("(")).trim { it <= ' ' }

diff --git a/app/src/test/java/com/ratanparai/moviedog/scrapper/DekhvhaiScrapperTest.kt b/app/src/test/java/com/ratanparai/moviedog/scrapper/DekhvhaiScrapperTest.kt
@@ -1,9 +1,23 @@
 package com.ratanparai.moviedog.scrapper
 
 import com.google.common.truth.Truth.assertThat
+import org.jsoup.Jsoup
+import org.jsoup.nodes.Document
+import org.junit.Before
 import org.junit.Test
+import java.io.File
 
 class DekhvhaiScrapperTest {
+
+    private lateinit var dekhvhai: Document
+
+    @Before
+    fun loadHtml() {
+        val file = ClassLoader.getSystemResource("HarryPotter_Dekhvhai.html").readText()
+
+        dekhvhai = Jsoup.parse(file, "UTF-8")
+    }
+
     @Test
     fun shouldGetCorrectTitleFromTitleAndYear() {
         // Arrange
@@ -51,4 +65,25 @@ class DekhvhaiScrapperTest {
         assertThat(actual).isEqualTo(expected)
 
     }
+
+    @Test
+    fun shouldLoadHtmlFileAsJsoupDocument() {
+        //val file = File("html/HarryPotter_Dekhvhai.html")
+        val file = ClassLoader.getSystemResource("HarryPotter_Dekhvhai.html").readText()
+
+        val document = Jsoup.parse(file, "UTF-8")
+
+        assertThat(document).isNotNull()
+    }
+
+    @Test
+    fun getMovieUrlsToBrowse() {
+        val scrapper = DekhvhaiScrapper()
+        val movieLinks = scrapper.getListOfMovieLinksFromSearchResult(dekhvhai)
+        assertThat(movieLinks.size).isEqualTo(4)
+
+        assertThat(movieLinks[0]).isEqualTo("http://dekhvhai.com/movie.php?imdbid=tt1201607&cat=English%20Movie")
+
+    }
+
 }