s434730/RandomSec
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Move to Guava's cache for ColumnAdditionByFetchingURLsOperation

Browse Source
This commit is contained in:
Antonin Delpeuch 2017-03-08 09:32:34 +00:00
parent a9c4b0af16
commit 32c232c2d6
2 changed files with 32 additions and 31 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

57
main/src/com/google/refine/operations/column/ColumnAdditionByFetchingURLsOperation.java
View File

@ -43,8 +43,8 @@ import java.net.URLConnection;
import java.util.ArrayList; import java.util.ArrayList;
import java.util.List; import java.util.List;
import java.util.Properties; import java.util.Properties;
import java.util.Map; import java.util.concurrent.TimeUnit;
import java.util.HashMap; import java.util.concurrent.ExecutionException;
import org.json.JSONException; import org.json.JSONException;
import org.json.JSONObject; import org.json.JSONObject;
@ -73,6 +73,9 @@ import com.google.refine.operations.cell.TextTransformOperation;
import com.google.refine.process.LongRunningProcess; import com.google.refine.process.LongRunningProcess;
import com.google.refine.process.Process; import com.google.refine.process.Process;
import com.google.refine.util.ParsingUtilities; import com.google.refine.util.ParsingUtilities;
import com.google.common.cache.CacheBuilder;
import com.google.common.cache.LoadingCache;
import com.google.common.cache.CacheLoader;
public class ColumnAdditionByFetchingURLsOperation extends EngineDependentOperation { public class ColumnAdditionByFetchingURLsOperation extends EngineDependentOperation {
final protected String _baseColumnName; final protected String _baseColumnName;
@ -172,7 +175,7 @@ public class ColumnAdditionByFetchingURLsOperation extends EngineDependentOperat
final protected Evaluable _eval; final protected Evaluable _eval;
final protected long _historyEntryID; final protected long _historyEntryID;
protected int _cellIndex; protected int _cellIndex;
protected Map<String, Serializable> _urlCache; protected LoadingCache<String, Serializable> _urlCache;
public ColumnAdditionByFetchingURLsProcess( public ColumnAdditionByFetchingURLsProcess(
Project project, Project project,
@ -185,7 +188,28 @@ public class ColumnAdditionByFetchingURLsOperation extends EngineDependentOperat
_engine = engine; _engine = engine;
_eval = eval; _eval = eval;
_historyEntryID = HistoryEntry.allocateID(); _historyEntryID = HistoryEntry.allocateID();
_urlCache = new HashMap<String, Serializable>(); _urlCache = CacheBuilder.newBuilder()
.maximumSize(2048)
.expireAfterWrite(10, TimeUnit.MINUTES)
.build(
new CacheLoader<String, Serializable>() {
public Serializable load(String urlString) {
Serializable result = fetch(urlString);
try {
// Always sleep for the delay, no matter how long the
// request took. This is more responsible than substracting
// the time spend requesting the URL, because it naturally
// slows us down if the server is busy and takes a long time
// to reply.
if (_delay > 0) {
Thread.sleep(_delay);
}
} catch (InterruptedException e) {
return null;
}
return result;
}
});
} }
@Override @Override
@ -238,8 +262,6 @@ public class ColumnAdditionByFetchingURLsOperation extends EngineDependentOperat
} }
} }
_urlCache.clear();
if (!_canceled) { if (!_canceled) {
HistoryEntry historyEntry = new HistoryEntry( HistoryEntry historyEntry = new HistoryEntry(
_historyEntryID, _historyEntryID,
@ -260,25 +282,10 @@ public class ColumnAdditionByFetchingURLsOperation extends EngineDependentOperat
CellAtRow cachedFetch(CellAtRow urlData) { CellAtRow cachedFetch(CellAtRow urlData) {
String urlString = urlData.cell.value.toString(); String urlString = urlData.cell.value.toString();
Serializable cellResult = _urlCache.get(urlString); Serializable cellResult = null;
if (cellResult == null) { try {
cellResult = fetch(urlString); cellResult = _urlCache.get(urlString);
if (cellResult != null) { } catch(ExecutionException e) {
_urlCache.put(urlString, cellResult);
}
try {
// Always sleep for the delay, no matter how long the
// request took. This is more responsible than substracting
// the time spend requesting the URL, because it naturally
// slows us down if the server is busy and takes a long time
// to reply.
if (_delay > 0) {
Thread.sleep(_delay);
}
} catch (InterruptedException e) {
return null;
}
} }
if (cellResult != null) { if (cellResult != null) {

6
main/tests/server/src/com/google/refine/tests/model/UrlFetchingTests.java
View File

@ -128,12 +128,6 @@ public class UrlFetchingTests extends RefineTest {
row.setCell(0, new Cell(i < 5 ? "apple":"orange", null)); row.setCell(0, new Cell(i < 5 ? "apple":"orange", null));
project.rows.add(row); project.rows.add(row);
} }
//engine.getAllRows().accept(project, new CountingRowVisitor(5)) ;
//engine.getAllFilteredRows().accept(project, new CountingRowVisitor(4));
// Function fc = new FacetCount();
// Integer count = (Integer) fc.call(bindings, new Object[] {"a", "value", "Column A"});
// Assert.assertEquals(count.intValue(), 4);
// EngineDependentOperation op = new RowRemovalOperation(engine_config);
EngineDependentOperation op = new ColumnAdditionByFetchingURLsOperation(engine_config, EngineDependentOperation op = new ColumnAdditionByFetchingURLsOperation(engine_config,
"fruits", "fruits",
"\"https://www.random.org/integers/?num=1&min=1&max=100&col=1&base=10&format=plain&rnd=new&city=\"+value", "\"https://www.random.org/integers/?num=1&min=1&max=100&col=1&base=10&format=plain&rnd=new&city=\"+value",