(sample) Ensure we flush the log before adding it to the tar file

We now sometimes get 206 when good.

code/common/model/java/nu/marginalia/model/EdgeUrl.java

@@ -121,6 +121,28 @@ public class EdgeUrl implements Serializable {
         return sb.toString();
     }
 
+
+    public String toDisplayString() {
+        StringBuilder sb = new StringBuilder(256);
+
+        sb.append(proto);
+        sb.append("://");
+        sb.append(domain);
+
+        if (port != null) {
+            sb.append(':');
+            sb.append(port);
+        }
+
+        sb.append(path);
+
+        if (param != null) {
+            sb.append('?').append(param);
+        }
+
+        return sb.toString();
+    }
+
     public String dir() {
         return path.replaceAll("/[^/]+$", "/");
     }
@@ -303,7 +325,14 @@ class EdgeUriFactory {
             }
 
             if (shouldUrlEncode) {
-                sb.append(URLEncoder.encode(queryPart, StandardCharsets.UTF_8));
+                int idx = queryPart.indexOf('=');
+                if (idx < 0) {
+                    sb.append(URLEncoder.encode(queryPart, StandardCharsets.UTF_8));
+                } else {
+                    sb.append(URLEncoder.encode(queryPart.substring(0, idx), StandardCharsets.UTF_8));
+                    sb.append('=');
+                    sb.append(URLEncoder.encode(queryPart.substring(idx + 1), StandardCharsets.UTF_8));
+                }
             } else {
                 sb.append(queryPart);
             }

code/common/model/test/nu/marginalia/model/EdgeUrlTest.java

@@ -58,16 +58,32 @@ class EdgeUrlTest {
         Assertions.assertEquals("/wiki/Sámi", EdgeUriFactory.parseURILenient("https://en.wikipedia.org/wiki/Sámi").getPath());
         Assertions.assertEquals("https://en.wikipedia.org/wiki/S%C3%A1mi", EdgeUriFactory.parseURILenient("https://en.wikipedia.org/wiki/Sámi").toString());
         Assertions.assertEquals("https://en.wikipedia.org/wiki/S%C3%A1mi", new EdgeUrl("https://en.wikipedia.org/wiki/Sámi").toString());
+
+        Assertions.assertEquals("https://www.prijatelji-zivotinja.hr/index.en.php?id=2301k", new EdgeUrl("https://www.prijatelji-zivotinja.hr/index.en.php?id=2301k").toString());
     }
 
     @Test
     void testParms() throws URISyntaxException {
         Assertions.assertEquals("id=123", new EdgeUrl("https://search.marginalia.nu/?id=123").param);
+        Assertions.assertEquals("https://search.marginalia.nu/?id=123", new EdgeUrl("https://search.marginalia.nu/?id=123").toString());
+
         Assertions.assertEquals("t=123", new EdgeUrl("https://search.marginalia.nu/?t=123").param);
+        Assertions.assertEquals("https://search.marginalia.nu/?t=123", new EdgeUrl("https://search.marginalia.nu/?t=123").toString());
+
         Assertions.assertEquals("v=123", new EdgeUrl("https://search.marginalia.nu/?v=123").param);
+        Assertions.assertEquals("https://search.marginalia.nu/?v=123", new EdgeUrl("https://search.marginalia.nu/?v=123").toString());
+
         Assertions.assertEquals("id=1", new EdgeUrl("https://memex.marginalia.nu/showthread.php?id=1&count=5&tracking=123").param);
+        Assertions.assertEquals("https://memex.marginalia.nu/showthread.php?id=1",
+                new EdgeUrl("https://memex.marginalia.nu/showthread.php?id=1&count=5&tracking=123").toString());
+
+
         Assertions.assertEquals("id=1&t=5", new EdgeUrl("https://memex.marginalia.nu/shöwthrëad.php?id=1&t=5&tracking=123").param);
+        Assertions.assertEquals("https://memex.marginalia.nu/sh%C3%B6wthr%C3%ABad.php?id=1&t=5", new EdgeUrl("https://memex.marginalia.nu/shöwthrëad.php?id=1&t=5&tracking=123").toString());
+
         Assertions.assertEquals("id=1&t=5", new EdgeUrl("https://memex.marginalia.nu/shöwthrëad.php?trëaking=123&id=1&t=5&").param);
+        Assertions.assertEquals("https://memex.marginalia.nu/sh%C3%B6wthr%C3%ABad.php?id=1&t=5", new EdgeUrl("https://memex.marginalia.nu/shöwthrëad.php?trëaking=123&id=1&t=5&").toString());
+
         Assertions.assertNull(new EdgeUrl("https://search.marginalia.nu/?m=123").param);
         Assertions.assertNull(new EdgeUrl("https://search.marginalia.nu/?follow=123").param);
     }

code/common/service/java/nu/marginalia/service/server/JoobyService.java

@@ -122,6 +122,11 @@ public class JoobyService {
         // single digit percentage difference since HTML already compresses very well with level = 1.
         options.setCompressionLevel(1);
 
+        // Set a cap on the number of worker threads, as Jooby's default value does not seem to consider
+        // multi-tenant servers with high thread counts, and spins up an exorbitant number of threads in that
+        // scenario
+        options.setWorkerThreads(Math.min(128, options.getWorkerThreads()));
+
 
         jooby.setServerOptions(options);
 

code/processes/crawling-process/build.gradle

@@ -67,8 +67,6 @@ dependencies {
     testImplementation libs.mockito
     testImplementation libs.wiremock
 
-
-
     testImplementation project(':code:processes:test-data')
 }
 

code/processes/crawling-process/java/nu/marginalia/crawl/CrawlerMain.java

@@ -448,13 +448,7 @@ public class CrawlerMain extends ProcessMainClass {
             // We don't have a lock, so we can't run this task
             // we return to avoid blocking the pool for too long
             if (lock.isEmpty()) {
-                if (retryQueue.remainingCapacity() > 0) {
+                pendingCrawlTasks.remove(domain);
-                    // Sleep a moment to avoid busy looping via the retry queue
-                    // in the case when few tasks remain and almost all are ineligible for
-                    // immediate restart
-                    Thread.sleep(5);
-                }
-
                 retryQueue.put(this);
                 return;
             }

code/processes/crawling-process/java/nu/marginalia/crawl/retreival/revisit/CrawlerRevisitor.java

@@ -74,7 +74,7 @@ public class CrawlerRevisitor {
 
             // If the reference document is empty or the HTTP status is not 200, we'll skip it since it's
             // unlikely to produce anything meaningful for us.
-            if (doc.httpStatus != 200)
+            if (doc.httpStatus != 200 && doc.httpStatus != 206)
                 continue;
             if (!doc.hasBody())
                 continue;

code/processes/crawling-process/java/nu/marginalia/crawl/retreival/revisit/DocumentWithReference.java

@@ -58,7 +58,7 @@ public record DocumentWithReference(
         if (null == doc)
             return ContentTags.empty();
 
-        if (doc.documentBodyBytes.length == 0 || doc.httpStatus != 200)
+        if (doc.documentBodyBytes.length == 0 || (doc.httpStatus != 200 && doc.httpStatus != 206))
             return ContentTags.empty();
 
         String lastmod = doc.getLastModified();

code/processes/crawling-process/model/java/nu/marginalia/ContentTypes.java

@@ -1,5 +1,7 @@
 package nu.marginalia;
 
+import org.apache.commons.lang3.StringUtils;
+
 import java.util.Set;
 
 public class ContentTypes {
@@ -11,9 +13,9 @@ public class ContentTypes {
             "text/plain");
 
     public static boolean isAccepted(String contentTypeHeader) {
-        String lcHeader = contentTypeHeader.toLowerCase();
+        String lcHeader = StringUtils.substringBefore(contentTypeHeader.toLowerCase(), ';');
         for (var type : acceptedContentTypes) {
-            if (lcHeader.startsWith(type)) {
+            if (lcHeader.equals(type)) {
                 return true;
             }
         }
@@ -21,7 +23,7 @@ public class ContentTypes {
     }
 
     public static boolean isBinary(String contentTypeHeader) {
-        String lcHeader = contentTypeHeader.toLowerCase();
+        String lcHeader = StringUtils.substringBefore(contentTypeHeader.toLowerCase(), ';');
         return lcHeader.startsWith("application/pdf");
     }
 

code/processes/crawling-process/model/java/nu/marginalia/slop/SlopCrawlDataRecord.java

@@ -277,7 +277,8 @@ public record SlopCrawlDataRecord(String domain,
         try (var table = new SlopTable(path)) {
             ShortColumn.Reader statusReader = statusColumn.open(table);
             while (statusReader.hasRemaining()) {
-                if (statusReader.get() == 200) {
+                int status = statusReader.get();
+                if (status == 200 || status == 206) {
                     cnt++;
                 }
             }

code/processes/crawling-process/test/nu/marginalia/crawling/retreival/CrawlerRetreiverTest.java

@@ -117,6 +117,86 @@ class CrawlerRetreiverTest {
         }
     }
 
+
+    @Test
+    public void testWarcOutputPDF() throws IOException {
+        var specs = CrawlerMain.CrawlSpecRecord
+                .builder()
+                .crawlDepth(5)
+                .domain("www.marginalia.nu")
+                .urls(List.of("https://www.marginalia.nu/junk/test.pdf"))
+                .build();
+        Path tempFile = null;
+        Path slopFile = null;
+        try {
+            tempFile = Files.createTempFile("crawling-process", "warc");
+            slopFile = Files.createTempFile("crawling-process", ".slop.zip");
+
+            doCrawl(tempFile, specs);
+
+            Set<String> requests = new HashSet<>();
+            Set<String> responses = new HashSet<>();
+
+            // Inspect the WARC file
+            try (var reader = new WarcReader(tempFile)) {
+                reader.forEach(record -> {
+                    if (record instanceof WarcRequest req) {
+                        requests.add(req.target());
+                        System.out.println(req.type() + ":" + req.target());
+                    }
+                    else if (record instanceof WarcResponse rsp) {
+                        responses.add(rsp.target());
+                        System.out.println(rsp.type() + ":" + rsp.target());
+                    }
+                    else {
+                        System.out.println(record.type());
+                    }
+                });
+            }
+
+            assertTrue(requests.contains("https://www.marginalia.nu/junk/test.pdf"));
+            assertEquals(requests, responses);
+
+            // Convert the WARC file to a Slop file
+            SlopCrawlDataRecord
+                    .convertWarc("www.marginalia.nu", new UserAgent("test.marginalia.nu", "test.marginalia.nu"), tempFile, slopFile);
+
+            CrawledDomain domain = null;
+            Map<String, CrawledDocument> documents = new HashMap<>();
+
+            // Extract the contents of the Slop file
+            try (var stream = SerializableCrawlDataStream.openDataStream(slopFile)) {
+                while (stream.hasNext()) {
+                    var doc = stream.next();
+                    if (doc instanceof CrawledDomain dr) {
+                        assertNull(domain);
+                        domain = dr;
+                    }
+                    else if (doc instanceof CrawledDocument dc) {
+                        System.out.println(dc.url + "\t" + dc.crawlerStatus + "\t" + dc.httpStatus);
+                        documents.put(dc.url, dc);
+                    }
+                }
+            } catch (Exception e) {
+                throw new RuntimeException(e);
+            }
+
+            // Verify we have a PDF in the Slop file
+            assertNotNull(domain);
+            var pdfDoc = documents.get("https://www.marginalia.nu/junk/test.pdf");
+            assertNotNull(pdfDoc);
+            assertEquals("https://www.marginalia.nu/junk/test.pdf", pdfDoc.url);
+            assertEquals(206, pdfDoc.httpStatus);
+            assertTrue(pdfDoc.documentBodyBytes.length > 100);
+        }
+        finally {
+            if (tempFile != null)
+                Files.deleteIfExists(tempFile);
+            if (slopFile != null)
+                Files.deleteIfExists(slopFile);
+        }
+    }
+
     @Test
     public void testWarcOutputNoKnownUrls() throws IOException {
         var specs = CrawlerMain.CrawlSpecRecord

code/processes/export-task-process/java/nu/marginalia/extractor/SampleDataExporter.java

@@ -1,6 +1,7 @@
 package nu.marginalia.extractor;
 
 import com.google.inject.Inject;
+import nu.marginalia.process.control.ProcessHeartbeat;
 import nu.marginalia.process.log.WorkLog;
 import nu.marginalia.process.log.WorkLogEntry;
 import nu.marginalia.slop.SlopCrawlDataRecord;
@@ -20,17 +21,18 @@ import java.nio.file.StandardCopyOption;
 import java.nio.file.StandardOpenOption;
 import java.nio.file.attribute.PosixFilePermissions;
 import java.sql.SQLException;
-import java.util.ArrayList;
+import java.util.*;
-import java.util.Collections;
-import java.util.List;
 
 public class SampleDataExporter {
     private final FileStorageService storageService;
+    private final ProcessHeartbeat processHeartbeat;
 
     @Inject
-    public SampleDataExporter(FileStorageService storageService) {
+    public SampleDataExporter(FileStorageService storageService, ProcessHeartbeat processHeartbeat) {
         this.storageService = storageService;
+        this.processHeartbeat = processHeartbeat;
     }
+
     public void export(FileStorageId crawlId, FileStorageId destId, int size, String ctFilter, String name) throws SQLException, IOException {
         FileStorage destStorage = storageService.getStorage(destId);
         Path inputDir = storageService.getStorage(crawlId).asPath();
@@ -59,12 +61,6 @@ public class SampleDataExporter {
         Path newCrawlerLogFile = Files.createTempFile(destStorage.asPath(), "crawler", ".log",
                 PosixFilePermissions.asFileAttribute(PosixFilePermissions.fromString("rw-r--r--")));
 
-        try (var bw = Files.newBufferedWriter(newCrawlerLogFile, StandardOpenOption.CREATE, StandardOpenOption.TRUNCATE_EXISTING)) {
-            for (var item : entriesAll) {
-                bw.write(item.id() + " " + item.ts() + " " + item.relPath() + " " + item.cnt() + "\n");
-            }
-        }
-
         Path newManifestJsonFile = Files.createTempFile(destStorage.asPath(), "manifest", ".json",
                 PosixFilePermissions.asFileAttribute(PosixFilePermissions.fromString("rw-r--r--")));
         Files.writeString(newManifestJsonFile, " { \"description\": \"" + name.replace("[\"\\]", "_") + "\",\n      \"type\": \"CRAWL_DATA\" }\n");
@@ -72,29 +68,38 @@ public class SampleDataExporter {
         var tmpTarFile = Files.createTempFile(destStorage.asPath(), "data", ".tar",
                 PosixFilePermissions.asFileAttribute(PosixFilePermissions.fromString("rw-r--r--")));
 
-        try (var stream = new TarArchiveOutputStream(Files.newOutputStream(tmpTarFile, StandardOpenOption.CREATE, StandardOpenOption.TRUNCATE_EXISTING))) {
+        try (var stream = new TarArchiveOutputStream(Files.newOutputStream(tmpTarFile, StandardOpenOption.CREATE, StandardOpenOption.TRUNCATE_EXISTING));
-            for (var item : entriesAll) {
+             var logWriter = Files.newBufferedWriter(newCrawlerLogFile, StandardOpenOption.CREATE, StandardOpenOption.TRUNCATE_EXISTING);
+             var hb = processHeartbeat.createAdHocTaskHeartbeat("Generating Sample")
+        ) {
+            for (var item : hb.wrap("Scanning", entriesAll)) {
                 Path crawlDataPath = inputDir.resolve(item.relPath());
                 if (!Files.exists(crawlDataPath)) continue;
 
                 if (StringUtils.isBlank(ctFilter)) {
                     addFileToTar(stream, crawlDataPath, item.relPath());
+                    logWriter.write(item.id() + " " + item.ts() + " " + item.relPath() + " " + item.cnt() + "\n");
                 }
                 else /* filter != null */ {
-                    boolean didFilterData = false;
+                    Path filteredData = null;
                     try {
-                        crawlDataPath = filterEntries(crawlDataPath, ctFilter);
+                        filteredData = filterEntries(crawlDataPath, ctFilter);
-                        didFilterData = true;
+                        addFileToTar(stream, filteredData, item.relPath());
-                        addFileToTar(stream, crawlDataPath, item.relPath());
+                        logWriter.write(item.id() + " " + item.ts() + " " + item.relPath() + " " + item.cnt() + "\n");
+                    }
+                    catch (NoSuchElementException ex) {
+                        // Ignore
                     }
                     finally {
-                        if (didFilterData) {
+                        if (filteredData != null) {
-                            Files.deleteIfExists(crawlDataPath);
+                            Files.deleteIfExists(filteredData);
                         }
                     }
                 }
             }
 
+            logWriter.flush();
+
             addFileToTar(stream, newCrawlerLogFile, "crawler.log");
             addFileToTar(stream, newManifestJsonFile, "marginalia-manifest.json");
         }
@@ -106,34 +111,36 @@ public class SampleDataExporter {
         Files.move(tmpTarFile, destStorage.asPath().resolve("crawl-data.tar"), StandardCopyOption.ATOMIC_MOVE, StandardCopyOption.REPLACE_EXISTING);
     }
 
-    /** Filters the entries in the crawl data file based on the content type.
+    /** Filters the entries in the crawl data file based on the content type. */
-     * @param crawlDataPath The path to the crawl data file.
+    private Path filterEntries(Path crawlDataPath, String contentTypeFilter) throws IOException, NoSuchElementException {
-     * @param contentTypeFilter The content type to filter by.
-     * @return The path to the filtered crawl data file, or null if an error occurred.
-     */
-    private Path filterEntries(Path crawlDataPath, String contentTypeFilter) throws IOException {
         Path tempDir = crawlDataPath.resolveSibling(crawlDataPath.getFileName() + ".filtered");
         Path tempFile = crawlDataPath.resolveSibling(crawlDataPath.getFileName() + ".filtered.slop.zip");
 
+        // We may have debris from a previous run, so let's clean it up
+        if (Files.isDirectory(tempDir)) {
+            FileUtils.deleteDirectory(tempDir.toFile());
+        }
         Files.createDirectory(tempDir);
 
         try (var writer = new SlopCrawlDataRecord.Writer(tempDir);
              var reader = new SlopCrawlDataRecord.FilteringReader(crawlDataPath) {
                  @Override
                  public boolean filter(String url, int status, String contentType) {
-                     if (contentTypeFilter.equals(contentType))
+                     return Objects.equals(StringUtils.substringBefore(contentType, ';'), contentTypeFilter)
-                         return true;
+                                || contentType.startsWith("x-marginalia/"); // metadata records
-                     else if (contentType.startsWith("x-marginalia/"))
-                         // This is a metadata entry, typically domain or redirect information
-                         // let's keep those to not confuse the consumer of the data, which might
-                         // expect at least the domain summary
-                         return true;
-                     return false;
                  }
              }
         ) {
+            boolean wroteEntry = false;
             while (reader.hasRemaining()) {
-                writer.write(reader.get());
+                var entry = reader.get();
+                writer.write(entry);
+
+                wroteEntry = wroteEntry || Objects.equals(StringUtils.substringBefore(entry.contentType(), ';'), contentTypeFilter);
+            }
+
+            if (!wroteEntry) {
+                throw new NoSuchElementException("No relevant entries");
             }
 
             SlopTablePacker.packToSlopZip(tempDir, tempFile);

code/services-application/search-service/java/nu/marginalia/search/model/UrlDetails.java

@@ -180,7 +180,7 @@ public class UrlDetails implements Comparable<UrlDetails> {
      * semantically meaningful codepoints into entity codes */
     public String displayUrl() {
         StringBuilder sb = new StringBuilder();
-        String urlStr = url.toString();
+        String urlStr = url.toDisplayString();
         for (int i = 0; i < urlStr.length(); i++) {
             char c = urlStr.charAt(i);
 

deploy.txt

@@ -4,4 +4,6 @@
 2025-01-07:  Deploy executor.
 2025-04-24:  Deploy executor.
 2025-04-24:  Deploy assistant.
-2025-04-24:  Deploy qs, search and api-services.
+2025-05-04:  Deploy qs, search and api-services.
+2025-05-05:  Deploy executor partition 4.
+2025-05-05:  Deploy control.