improve prescan for schemaVersion attribute

openepcis · Oct 31, 2023 · 1fed64c · 1fed64c
1 parent 0942263
commit 1fed64c
Show file tree

Hide file tree

Showing 2 changed files with 40 additions and 31 deletions.
diff --git a/src/main/java/io/openepcis/convert/AttributePreScanUtil.java b/src/main/java/io/openepcis/convert/AttributePreScanUtil.java
@@ -0,0 +1,36 @@
+package io.openepcis.convert;
+
+import java.io.BufferedInputStream;
+import java.io.IOException;
+import java.nio.charset.StandardCharsets;
+import java.util.regex.Matcher;
+import java.util.regex.Pattern;
+
+public class AttributePreScanUtil {
+
+  private static final String SCHEMA_VERSION_REGEX ="schemaVersion\"?'?\\s*[=:]\\s*([\"'])?([^\"']*)";
+  private static final Pattern SCHEMA_VERSION_PATTERN = Pattern.compile(SCHEMA_VERSION_REGEX);
+  private static final int READ_LIMIT = 4096;
+  public static final String scanSchemaVersion(final BufferedInputStream input) throws IOException {
+    input.mark(READ_LIMIT);
+    try  {
+      final StringBuilder sb = new StringBuilder();
+      final byte[] buffer = new byte[64];
+      int len = -1;
+      int bytesReceived = 0;
+      Matcher matcher = SCHEMA_VERSION_PATTERN.matcher(sb.toString());
+      while (!matcher.find(0) && bytesReceived < READ_LIMIT && (len = input.read(buffer)) != -1) {
+        sb.append(new String(buffer, 0, len, StandardCharsets.UTF_8));
+        bytesReceived += len;
+        matcher = SCHEMA_VERSION_PATTERN.matcher(sb.toString());
+      }
+      if (matcher.find(0)) {
+        return matcher.group(2);
+      }
+      return "";
+    } finally {
+      input.reset();
+    }
+  }
+
+}
diff --git a/src/main/java/io/openepcis/convert/VersionTransformer.java b/src/main/java/io/openepcis/convert/VersionTransformer.java
@@ -32,13 +32,11 @@
 import io.openepcis.convert.xml.XmlToJsonConverter;
 import io.openepcis.convert.xml.XmlVersionTransformer;
 import io.openepcis.model.rest.ProblemResponseBody;
-import io.openepcis.validation.xml.PreScanUtil;
 import jakarta.xml.bind.JAXBContext;
 import jakarta.xml.bind.JAXBException;
 import lombok.extern.slf4j.Slf4j;
 
 import java.io.*;
-import java.nio.charset.StandardCharsets;
 import java.util.List;
 import java.util.Optional;
 import java.util.concurrent.ExecutorService;
@@ -181,40 +179,15 @@ public final EPCISVersion versionDetector(final BufferedInputStream epcisDocumen
             return conversion.fromVersion();
         }
 
-        // TODO: optimize prescan with regex matcher
-        String preScanVersion = null;
-        if (EPCISFormat.XML.equals(conversion.fromMediaType())) {
-            preScanVersion = PreScanUtil.scanFirstTag(epcisDocument);
-        } else {
-            epcisDocument.mark(1024);
-            // pre scan 1024 bytes to detect version
-            final byte[] preScan = new byte[1024];
-            epcisDocument.read(preScan, 0, preScan.length);
-            epcisDocument.reset();
-            preScanVersion = new String(preScan, StandardCharsets.UTF_8);
-        }
+        final String preScanVersion = AttributePreScanUtil.scanSchemaVersion(epcisDocument);
 
-        if (!preScanVersion.contains(EPCIS.SCHEMA_VERSION)) {
+        if (preScanVersion.isEmpty()) {
             throw new FormatConverterException(
                     "Unable to detect EPCIS schemaVersion for given document, please check the document again");
         }
 
-        EPCISVersion fromVersion;
-
-        if (preScanVersion.contains(EPCIS.SCHEMA_VERSION + "=\"1.2\"")
-                || preScanVersion.contains(EPCIS.SCHEMA_VERSION + "='1.2'")
-                || preScanVersion.replace(" ", "").contains("\"" + EPCIS.SCHEMA_VERSION + "\":\"1.2\"")) {
-            fromVersion = EPCISVersion.VERSION_1_2_0;
-        } else if (preScanVersion.contains(EPCIS.SCHEMA_VERSION + "=\"2.0\"")
-                || preScanVersion.contains(EPCIS.SCHEMA_VERSION + "='2.0'")
-                || preScanVersion.replace(" ", "").contains("\"" + EPCIS.SCHEMA_VERSION + "\":\"2.0\"")) {
-            fromVersion = EPCISVersion.VERSION_2_0_0;
-        } else {
-            throw new FormatConverterException(
-                    "Provided document contains unsupported EPCIS document version");
-        }
-
-        return fromVersion;
+        return EPCISVersion.fromString(preScanVersion).orElseThrow(() -> new FormatConverterException(
+                String.format("Provided document contains unsupported EPCIS document version %s", preScanVersion)));
     }
 
     /**