scrape more events from pages

app/src/main/java/com/akdev/nofbeventscraper/FbPageScraper.java

@@ -1,7 +1,10 @@
 package com.akdev.nofbeventscraper;
 
+import android.content.SharedPreferences;
 import android.os.AsyncTask;
 
+import androidx.preference.PreferenceManager;
+
 import org.jsoup.Jsoup;
 import org.jsoup.nodes.Document;
 
@@ -44,32 +47,50 @@ public class FbPageScraper extends AsyncTask<Void, Void, Void> {
     @Override
     protected Void doInBackground(Void... voids) {
 
-        try {
-            // use default android user agent
-            String user_agent = "Mozilla/5.0 (X11; Linux x86_64)";
-            Document document = Jsoup.connect(url).userAgent(user_agent).get();
 
-            if (document == null) {
+        do {
-                throw new IOException();
+            try {
+                // use default android user agent
+                String user_agent = "Mozilla/5.0 (X11; Linux x86_64)";
+                Document document = Jsoup.connect(url).userAgent(user_agent).get();
+
+                if (document == null) {
+                    throw new IOException();
+                }
+
+                String regex = "(/events/[0-9]*)(/\\?event_time_id=[0-9]*)?";
+
+                List<String> event_links_href = document
+                        .getElementsByAttributeValueMatching("href", Pattern.compile(regex))
+                        .eachAttr("href");
+
+                for (String link : event_links_href) {
+                    this.event_links.add("https://www.facebook.com" + link);
+                }
+
+                SharedPreferences shared_prefs = PreferenceManager
+                        .getDefaultSharedPreferences(scraper.main.get());
+
+                int max = shared_prefs.getInt("page_event_max", 5);
+
+                if (event_links.size() < max) {
+                    String next_url = document
+                            .getElementsByAttributeValueMatching("href", "has_more=1")
+                            .first().attr("href");
+
+                    this.url = "https://mbasic.facebook.com" + next_url;
+                } else {
+                    url = null;
+                }
+
+            } catch (IOException e) {
+                e.printStackTrace();
+                this.error = R.string.error_connection;
+            } catch (Exception e) {
+                e.printStackTrace();
+                this.error = R.string.error_unknown;
             }
-
+        } while (url != null);
-            String regex = "(/events/[0-9]*)(/\\?event_time_id=[0-9]*)?";
-
-            List<String> event_links_href = document
-                    .getElementsByAttributeValueMatching("href", Pattern.compile(regex))
-                    .eachAttr("href");
-
-            for (String link : event_links_href) {
-                this.event_links.add("https://www.facebook.com" + link);
-            }
-
-        } catch (IOException e) {
-            e.printStackTrace();
-            this.error = R.string.error_connection;
-        } catch (Exception e) {
-            e.printStackTrace();
-            this.error = R.string.error_unknown;
-        }
 
         return null;
     }

app/src/main/java/com/akdev/nofbeventscraper/FbScraper.java

@@ -22,7 +22,7 @@ public class FbScraper {
     protected List<AsyncTask> tasks;
     url_type_enum url_type = url_type_enum.EVENT;
     private String input_url;
-    private WeakReference<MainActivity> main; // no context leak with WeakReference
+    protected WeakReference<MainActivity> main; // no context leak with WeakReference
 
     /**
      * Constructor with WeakReference to the main activity, to add events.

app/src/main/res/values-de/strings.xml

@@ -20,4 +20,6 @@
     <string name="preferences_event_snackbar">"Veranstaltungen gelöscht "</string>
     <string name="done">Fertig</string>
     <string name="undo">Rückgängig</string>
+    <string name="preferences_page_event_max_summary">Maximale Anzahl Events, die von einer einzelnen Seite geladen werden sollen.</string>
+    <string name="preferences_page_event_max">Veranstaltungslimit für Seiten</string>
 </resources>

app/src/main/res/values/strings.xml

@@ -30,9 +30,13 @@
     <string name="preferences_event_setting">Clear event list</string>
     <string name="preferences_event_snackbar">Events list cleared</string>
 
+    <string name="preferences_page_event_max_summary">Maximum amount of events scraped from a single page link.</string>
+    <string name="preferences_page_event_max">Page event limit</string>
+
     <!-- others -->
     <string name="event_placeholder" translatable="false">Placeholder</string>
     <string name="done">Done</string>
     <string name="undo">Undo</string>
 
+
 </resources>

app/src/main/res/xml/root_preferences.xml

@@ -16,6 +16,15 @@
     <PreferenceCategory app:title="@string/preferences_events_header">
 
 
+        <SeekBarPreference
+            android:defaultValue="5"
+            app:showSeekBarValue="true"
+            app:min="5"
+            android:max="30"
+            android:summary="@string/preferences_page_event_max_summary"
+            android:key="page_event_max"
+            android:title="@string/preferences_page_event_max" />
+
         <Preference
             android:key="event_reset"
             android:title="@string/preferences_event_setting" />