This is an automated email from the ASF dual-hosted git repository.
tallison pushed a commit to branch branch_3x
in repository https://gitbox.apache.org/repos/asf/tika.git
The following commit(s) were added to refs/heads/branch_3x by this push:
new 3fc3476f9 TIKA-4385 : fix : GDAL process output read in another thread
(#2126)
3fc3476f9 is described below
commit 3fc3476f997d559713fa823c42716141545c5283
Author: Leszek Sliwko <[email protected]>
AuthorDate: Sun Feb 16 15:38:49 2025 +0000
TIKA-4385 : fix : GDAL process output read in another thread (#2126)
* TIKA-4385 : fix : GDAL process output read in another thread
Co-authored-by: Tilman Hausherr <[email protected]>
(cherry picked from commit 05500bbd9a2840934652e5b647e0a85506a0325b)
---
.../org/apache/tika/parser/gdal/GDALParser.java | 90 +++++++++++-----------
1 file changed, 45 insertions(+), 45 deletions(-)
diff --git
a/tika-parsers/tika-parsers-extended/tika-parser-scientific-module/src/main/java/org/apache/tika/parser/gdal/GDALParser.java
b/tika-parsers/tika-parsers-extended/tika-parser-scientific-module/src/main/java/org/apache/tika/parser/gdal/GDALParser.java
index e82f9b6b7..1ba196942 100644
---
a/tika-parsers/tika-parsers-extended/tika-parser-scientific-module/src/main/java/org/apache/tika/parser/gdal/GDALParser.java
+++
b/tika-parsers/tika-parsers-extended/tika-parser-scientific-module/src/main/java/org/apache/tika/parser/gdal/GDALParser.java
@@ -19,12 +19,10 @@ package org.apache.tika.parser.gdal;
//JDK imports
-import static java.nio.charset.StandardCharsets.UTF_8;
import static org.apache.tika.parser.external.ExternalParser.INPUT_FILE_TOKEN;
import java.io.IOException;
import java.io.InputStream;
-import java.io.InputStreamReader;
import java.io.Reader;
import java.io.StringReader;
import java.util.Arrays;
@@ -42,15 +40,20 @@ import org.slf4j.LoggerFactory;
import org.xml.sax.ContentHandler;
import org.xml.sax.SAXException;
+import org.apache.tika.config.Field;
+import org.apache.tika.config.TikaTaskTimeout;
import org.apache.tika.exception.TikaException;
import org.apache.tika.io.TemporaryResources;
import org.apache.tika.io.TikaInputStream;
+import org.apache.tika.metadata.ExternalProcess;
import org.apache.tika.metadata.Metadata;
import org.apache.tika.mime.MediaType;
import org.apache.tika.parser.ParseContext;
import org.apache.tika.parser.Parser;
import org.apache.tika.parser.external.ExternalParser;
import org.apache.tika.sax.XHTMLContentHandler;
+import org.apache.tika.utils.FileProcessResult;
+import org.apache.tika.utils.ProcessUtils;
//Tika imports
//SAX imports
@@ -76,6 +79,8 @@ public class GDALParser implements Parser {
private static final long serialVersionUID = -3869130527323941401L;
private static final Logger LOG =
LoggerFactory.getLogger(GDALParser.class);
+ public static final long DEFAULT_TIMEOUT_MS = 60000;
+
private static final Set<MediaType> SUPPORTED_TYPES =
Collections.unmodifiableSet(new HashSet<>(
Arrays.asList(MediaType.application("x-netcdf"),
MediaType.application("vrt"),
MediaType.image("geotiff"), MediaType.image("nitf"),
@@ -140,6 +145,12 @@ public class GDALParser implements Parser {
private String command;
+ private int maxStdErr = 100000;
+
+ private int maxStdOut = 100000;
+
+ private long timeoutMs = DEFAULT_TIMEOUT_MS;
+
public GDALParser() {
setCommand("gdalinfo ${INPUT}");
}
@@ -184,8 +195,23 @@ public class GDALParser implements Parser {
TemporaryResources tmp = new TemporaryResources();
TikaInputStream tis = TikaInputStream.get(stream, tmp, metadata);
- String runCommand = processCommand(tis);
- String output = execCommand(new String[]{runCommand});
+ String[] runCommand = processCommand(tis).split("\\s+", -1);
+
+ long localTimeoutMillis = TikaTaskTimeout.getTimeoutMillis(context,
timeoutMs);
+ FileProcessResult result = ProcessUtils.execute(new
ProcessBuilder(runCommand),
+ localTimeoutMillis, maxStdOut, maxStdErr);
+
+ metadata.set(ExternalProcess.IS_TIMEOUT, result.isTimeout());
+ metadata.set(ExternalProcess.EXIT_VALUE, result.getExitValue());
+ metadata.set(ExternalProcess.STD_OUT_LENGTH, result.getStdoutLength());
+ metadata.set(ExternalProcess.STD_OUT_IS_TRUNCATED,
result.isStdoutTruncated());
+ metadata.set(ExternalProcess.STD_ERR_LENGTH, result.getStderrLength());
+ metadata.set(ExternalProcess.STD_ERR_IS_TRUNCATED,
result.isStderrTruncated());
+
+ metadata.set(ExternalProcess.STD_OUT, result.getStdout());
+ metadata.set(ExternalProcess.STD_ERR, result.getStderr());
+
+ String output = result.getStdout();
// now extract the actual metadata params
// from the GDAL output in the content stream
@@ -290,47 +316,6 @@ public class GDALParser implements Parser {
}
}
- private String execCommand(String[] cmd) throws IOException {
- // Execute
- Process process;
- String output = null;
- if (cmd.length == 1) {
- process = Runtime.getRuntime().exec(cmd[0]);
- } else {
- process = Runtime.getRuntime().exec(cmd);
- }
-
- try {
- InputStream out = process.getInputStream();
-
- try {
- output = extractOutput(out);
- } catch (Exception e) {
- LOG.warn("Exception extracting output", e);
- output = "";
- }
-
- } finally {
- try {
- process.waitFor();
- } catch (InterruptedException ignore) {
- }
- }
- return output;
-
- }
-
- private String extractOutput(InputStream stream) throws SAXException,
IOException {
- StringBuilder sb = new StringBuilder();
- try (Reader reader = new InputStreamReader(stream, UTF_8)) {
- char[] buffer = new char[1024];
- for (int n = reader.read(buffer); n != -1; n =
reader.read(buffer)) {
- sb.append(buffer, 0, n);
- }
- }
- return sb.toString();
- }
-
private void processOutput(ContentHandler handler, Metadata metadata,
String output)
throws SAXException, IOException {
XHTMLContentHandler xhtml = new XHTMLContentHandler(handler, metadata);
@@ -349,4 +334,19 @@ public class GDALParser implements Parser {
}
+ @Field
+ public void setTimeoutMs(long timeoutMs) {
+ this.timeoutMs = timeoutMs;
+ }
+
+ @Field
+ public void setMaxStdErr(int maxStdErr) {
+ this.maxStdErr = maxStdErr;
+ }
+
+ @Field
+ public void setMaxStdOut(int maxStdOut) {
+ this.maxStdOut = maxStdOut;
+ }
+
}
