(index-client) Clean up index client code

This should have the rpc stream reception be performed in parallel in separate threads, rather blocking sequentially in the main thread, hopefully giving a slight performance boost.

code/common/db/java/nu/marginalia/db/DbDomainQueries.java

@@ -20,7 +20,10 @@ public class DbDomainQueries {
     private final HikariDataSource dataSource;
 
     private static final Logger logger = LoggerFactory.getLogger(DbDomainQueries.class);
+
     private final Cache<EdgeDomain, Integer> domainIdCache = CacheBuilder.newBuilder().maximumSize(10_000).build();
+    private final Cache<Integer, EdgeDomain> domainNameCache = CacheBuilder.newBuilder().maximumSize(10_000).build();
+    private final Cache<String, List<DomainWithNode>> siblingsCache = CacheBuilder.newBuilder().maximumSize(10_000).build();
 
     @Inject
     public DbDomainQueries(HikariDataSource dataSource)
@@ -30,16 +33,21 @@ public class DbDomainQueries {
 
 
     public Integer getDomainId(EdgeDomain domain) throws NoSuchElementException {
-        try (var connection = dataSource.getConnection()) {
+        try {
-
             return domainIdCache.get(domain, () -> {
-                try (var stmt = connection.prepareStatement("SELECT ID FROM EC_DOMAIN WHERE DOMAIN_NAME=?")) {
+                try (var connection = dataSource.getConnection();
+                     var stmt = connection.prepareStatement("SELECT ID FROM EC_DOMAIN WHERE DOMAIN_NAME=?")) {
+
                     stmt.setString(1, domain.toString());
                     var rsp = stmt.executeQuery();
                     if (rsp.next()) {
                         return rsp.getInt(1);
                     }
                 }
+                catch (SQLException ex) {
+                    throw new RuntimeException(ex);
+                }
+
                 throw new NoSuchElementException();
             });
         }
@@ -49,9 +57,6 @@ public class DbDomainQueries {
         catch (ExecutionException ex) {
             throw new RuntimeException(ex.getCause());
         }
-        catch (SQLException ex) {
-            throw new RuntimeException(ex);
-        }
     }
 
     public OptionalInt tryGetDomainId(EdgeDomain domain) {
@@ -84,47 +89,55 @@ public class DbDomainQueries {
     }
 
     public Optional<EdgeDomain> getDomain(int id) {
-        try (var connection = dataSource.getConnection()) {
 
+        EdgeDomain existing = domainNameCache.getIfPresent(id);
+        if (existing != null) {
+            return Optional.of(existing);
+        }
+
+        try (var connection = dataSource.getConnection()) {
             try (var stmt = connection.prepareStatement("SELECT DOMAIN_NAME FROM EC_DOMAIN WHERE ID=?")) {
                 stmt.setInt(1, id);
                 var rsp = stmt.executeQuery();
                 if (rsp.next()) {
-                    return Optional.of(new EdgeDomain(rsp.getString(1)));
+                    var val = new EdgeDomain(rsp.getString(1));
+                    domainNameCache.put(id, val);
+                    return Optional.of(val);
                 }
                 return Optional.empty();
             }
         }
-        catch (UncheckedExecutionException ex) {
-            throw new RuntimeException(ex.getCause());
-        }
         catch (SQLException ex) {
             throw new RuntimeException(ex);
         }
     }
 
-    public List<DomainWithNode> otherSubdomains(EdgeDomain domain, int cnt) {
+    public List<DomainWithNode> otherSubdomains(EdgeDomain domain, int cnt) throws ExecutionException {
-        List<DomainWithNode> ret = new ArrayList<>();
+        String topDomain = domain.topDomain;
 
-        try (var conn = dataSource.getConnection();
+        return siblingsCache.get(topDomain, () -> {
-             var stmt = conn.prepareStatement("SELECT DOMAIN_NAME, NODE_AFFINITY FROM EC_DOMAIN WHERE DOMAIN_TOP = ? LIMIT ?")) {
+            List<DomainWithNode> ret = new ArrayList<>();
-            stmt.setString(1, domain.topDomain);
-            stmt.setInt(2, cnt);
 
-            var rs = stmt.executeQuery();
+            try (var conn = dataSource.getConnection();
-            while (rs.next()) {
+                 var stmt = conn.prepareStatement("SELECT DOMAIN_NAME, NODE_AFFINITY FROM EC_DOMAIN WHERE DOMAIN_TOP = ? LIMIT ?")) {
-                var sibling = new EdgeDomain(rs.getString(1));
+                stmt.setString(1, topDomain);
+                stmt.setInt(2, cnt);
 
-                if (sibling.equals(domain))
+                var rs = stmt.executeQuery();
-                    continue;
+                while (rs.next()) {
+                    var sibling = new EdgeDomain(rs.getString(1));
 
-                ret.add(new DomainWithNode(sibling, rs.getInt(2)));
+                    if (sibling.equals(domain))
+                        continue;
+
+                    ret.add(new DomainWithNode(sibling, rs.getInt(2)));
+                }
+            } catch (SQLException e) {
+                logger.error("Failed to get domain neighbors");
             }
-        } catch (SQLException e) {
+            return ret;
-            logger.error("Failed to get domain neighbors");
+        });
-        }
 
-        return ret;
     }
 
     public record DomainWithNode (EdgeDomain domain, int nodeAffinity) {

code/common/db/java/nu/marginalia/db/DbDomainStatsExportMultitool.java

@@ -1,118 +0,0 @@
-package nu.marginalia.db;
-
-import com.zaxxer.hikari.HikariDataSource;
-
-import java.sql.Connection;
-import java.sql.PreparedStatement;
-import java.sql.SQLException;
-import java.util.ArrayList;
-import java.util.List;
-import java.util.OptionalInt;
-
-/** Class used in exporting data.  This is intended to be used for a brief time
- * and then discarded, not kept around as a service.
- */
-public class DbDomainStatsExportMultitool implements AutoCloseable {
-    private final Connection connection;
-    private final int nodeId;
-    private final PreparedStatement knownUrlsQuery;
-    private final PreparedStatement visitedUrlsQuery;
-    private final PreparedStatement goodUrlsQuery;
-    private final PreparedStatement domainNameToId;
-
-    private final PreparedStatement allDomainsQuery;
-    private final PreparedStatement crawlQueueDomains;
-    private final PreparedStatement indexedDomainsQuery;
-
-    public DbDomainStatsExportMultitool(HikariDataSource dataSource, int nodeId) throws SQLException {
-        this.connection = dataSource.getConnection();
-        this.nodeId = nodeId;
-
-        knownUrlsQuery = connection.prepareStatement("""
-                SELECT KNOWN_URLS
-                FROM EC_DOMAIN INNER JOIN DOMAIN_METADATA
-                    ON EC_DOMAIN.ID=DOMAIN_METADATA.ID
-                WHERE DOMAIN_NAME=?
-                """);
-        visitedUrlsQuery = connection.prepareStatement("""
-                SELECT VISITED_URLS
-                FROM EC_DOMAIN INNER JOIN DOMAIN_METADATA
-                    ON EC_DOMAIN.ID=DOMAIN_METADATA.ID
-                WHERE DOMAIN_NAME=?
-                """);
-        goodUrlsQuery = connection.prepareStatement("""
-                SELECT GOOD_URLS
-                FROM EC_DOMAIN INNER JOIN DOMAIN_METADATA
-                    ON EC_DOMAIN.ID=DOMAIN_METADATA.ID
-                WHERE DOMAIN_NAME=?
-                """);
-        domainNameToId = connection.prepareStatement("""
-                SELECT ID
-                FROM EC_DOMAIN
-                WHERE DOMAIN_NAME=?
-                """);
-        allDomainsQuery = connection.prepareStatement("""
-                SELECT DOMAIN_NAME
-                FROM EC_DOMAIN
-                """);
-        crawlQueueDomains = connection.prepareStatement("""
-                SELECT DOMAIN_NAME
-                FROM CRAWL_QUEUE
-                """);
-        indexedDomainsQuery = connection.prepareStatement("""
-                SELECT DOMAIN_NAME
-                FROM EC_DOMAIN
-                WHERE INDEXED > 0
-                """);
-    }
-
-    public OptionalInt getVisitedUrls(String domainName) throws SQLException {
-        return executeNameToIntQuery(domainName, visitedUrlsQuery);
-    }
-
-    public OptionalInt getDomainId(String domainName) throws SQLException {
-        return executeNameToIntQuery(domainName, domainNameToId);
-    }
-
-    public List<String> getCrawlQueueDomains() throws SQLException {
-        return executeListQuery(crawlQueueDomains, 100);
-    }
-    public List<String> getAllIndexedDomains() throws SQLException {
-        return executeListQuery(indexedDomainsQuery, 100_000);
-    }
-
-    private OptionalInt executeNameToIntQuery(String domainName, PreparedStatement statement)
-            throws SQLException {
-        statement.setString(1, domainName);
-        var rs = statement.executeQuery();
-
-        if (rs.next()) {
-            return OptionalInt.of(rs.getInt(1));
-        }
-
-        return OptionalInt.empty();
-    }
-
-    private List<String> executeListQuery(PreparedStatement statement, int sizeHint) throws SQLException {
-        List<String> ret = new ArrayList<>(sizeHint);
-
-        var rs = statement.executeQuery();
-
-        while (rs.next()) {
-            ret.add(rs.getString(1));
-        }
-
-        return ret;
-    }
-
-    @Override
-    public void close() throws SQLException {
-        knownUrlsQuery.close();
-        goodUrlsQuery.close();
-        visitedUrlsQuery.close();
-        allDomainsQuery.close();
-        crawlQueueDomains.close();
-        domainNameToId.close();
-        connection.close();
-    }
-}

code/index/api/java/nu/marginalia/index/api/IndexClient.java

@@ -16,20 +16,18 @@ import org.slf4j.LoggerFactory;
 
 import java.util.ArrayList;
 import java.util.Comparator;
-import java.util.Iterator;
 import java.util.List;
 import java.util.concurrent.CompletableFuture;
 import java.util.concurrent.ExecutorService;
 import java.util.concurrent.Executors;
-
+import java.util.concurrent.atomic.AtomicInteger;
-import static java.lang.Math.clamp;
 
 @Singleton
 public class IndexClient {
     private static final Logger logger = LoggerFactory.getLogger(IndexClient.class);
     private final GrpcMultiNodeChannelPool<IndexApiGrpc.IndexApiBlockingStub> channelPool;
     private final DomainBlacklistImpl blacklist;
-    private static final ExecutorService executor = Executors.newVirtualThreadPerTaskExecutor();
+    private static final ExecutorService executor = Executors.newCachedThreadPool();
 
     @Inject
     public IndexClient(GrpcChannelPoolFactory channelPoolFactory, DomainBlacklistImpl blacklist) {
@@ -51,40 +49,31 @@ public class IndexClient {
 
     /** Execute a query on the index partitions and return the combined results. */
     public AggregateQueryResponse executeQueries(RpcIndexQuery indexRequest, Pagination pagination) {
-        List<CompletableFuture<Iterator<RpcDecoratedResultItem>>> futures =
-                channelPool.call(IndexApiGrpc.IndexApiBlockingStub::query)
-                        .async(executor)
-                        .runEach(indexRequest);
 
         final int requestedMaxResults = indexRequest.getQueryLimits().getResultsTotal();
-        final int resultsUpperBound = requestedMaxResults * channelPool.getNumNodes();
 
-        List<RpcDecoratedResultItem> results = new ArrayList<>(resultsUpperBound);
+        AtomicInteger totalNumResults = new AtomicInteger(0);
 
-        for (var future : futures) {
+        List<RpcDecoratedResultItem> results =
-            try {
+                channelPool.call(IndexApiGrpc.IndexApiBlockingStub::query)
-                future.get().forEachRemaining(results::add);
+                        .async(executor)
-            }
+                        .runEach(indexRequest)
-            catch (Exception e) {
+                        .stream()
-                logger.error("Downstream exception", e);
+                        .map(future -> future.thenApply(iterator -> {
-            }
+                            List<RpcDecoratedResultItem> ret = new ArrayList<>(requestedMaxResults);
-        }
+                            iterator.forEachRemaining(ret::add);
+                            totalNumResults.addAndGet(ret.size());
+                            return ret;
+                        }))
+                        .map(CompletableFuture::join)
+                        .flatMap(List::stream)
+                        .filter(item -> !isBlacklisted(item))
+                        .sorted(comparator)
+                        .skip(Math.max(0, (pagination.page - 1) * pagination.pageSize))
+                        .limit(pagination.pageSize)
+                        .toList();
 
-        // Sort the results by ranking score and remove blacklisted domains
+        return new AggregateQueryResponse(results, pagination.page(), totalNumResults.get());
-        results.sort(comparator);
-        results.removeIf(this::isBlacklisted);
-
-        int numReceivedResults = results.size();
-
-        // pagination is typically 1-indexed, so we need to adjust the start and end indices
-        int indexStart = (pagination.page - 1) * pagination.pageSize;
-        int indexEnd = (pagination.page) * pagination.pageSize;
-
-        results = results.subList(
-                clamp(indexStart, 0, Math.max(0, results.size() - 1)), // from is inclusive, so subtract 1 from size()
-                clamp(indexEnd, 0, results.size()));
-
-        return new AggregateQueryResponse(results, pagination.page(), numReceivedResults);
     }
 
     private boolean isBlacklisted(RpcDecoratedResultItem item) {

code/services-application/search-service/java/nu/marginalia/search/svc/SearchSiteInfoService.java

@@ -28,6 +28,7 @@ import org.slf4j.LoggerFactory;
 import java.sql.SQLException;
 import java.util.*;
 import java.util.concurrent.CompletableFuture;
+import java.util.concurrent.ExecutionException;
 import java.util.concurrent.Future;
 import java.util.concurrent.TimeUnit;
 import java.util.function.Supplier;
@@ -67,8 +68,12 @@ public class SearchSiteInfoService {
         this.screenshotService = screenshotService;
         this.dataSource = dataSource;
         this.searchSiteSubscriptions = searchSiteSubscriptions;
+
+        Thread.ofPlatform().name("Recently Added Domains Model Updater").start(this::modelUpdater);
     }
 
+    private volatile SiteOverviewModel cachedOverviewModel = new SiteOverviewModel(List.of());
+
     @GET
     @Path("/site")
     public ModelAndView<?> handleOverview(@QueryParam String domain) {
@@ -77,23 +82,48 @@ public class SearchSiteInfoService {
             return new MapModelAndView("redirect.jte", Map.of("url", "/site/"+domain));
         }
 
-        List<SiteOverviewModel.DiscoveredDomain> domains = new ArrayList<>();
-
-        try (var conn = dataSource.getConnection();
-             var stmt = conn.prepareStatement("SELECT DOMAIN_NAME, DISCOVER_DATE FROM EC_DOMAIN WHERE NODE_AFFINITY = 0 ORDER BY ID DESC LIMIT 10")) {
-
-            var rs = stmt.executeQuery();
-            while (rs.next()) {
-                domains.add(new SiteOverviewModel.DiscoveredDomain(rs.getString("DOMAIN_NAME"), rs.getString("DISCOVER_DATE")));
-            }
-        }
-        catch (SQLException ex) {
-            throw new RuntimeException();
-        }
-
         return new MapModelAndView("siteinfo/start.jte",
                 Map.of("navbar", NavbarModel.SITEINFO,
-                        "model", new SiteOverviewModel(domains)));
+                        "model", cachedOverviewModel));
+    }
+
+    private void modelUpdater() {
+        while (!Thread.interrupted()) {
+            List<SiteOverviewModel.DiscoveredDomain> domains = new ArrayList<>();
+
+            // This query can be quite expensive, so we can't run it on demand
+            // for every request. Instead, we run it every 15 minutes and cache
+            // the result.
+
+            try (var conn = dataSource.getConnection();
+                 var stmt = conn.prepareStatement("""
+                    SELECT DOMAIN_NAME, DISCOVER_DATE
+                    FROM EC_DOMAIN
+                    WHERE NODE_AFFINITY = 0
+                    ORDER BY ID DESC
+                    LIMIT 10
+                    """))
+            {
+                var rs = stmt.executeQuery();
+                while (rs.next()) {
+                    domains.add(new SiteOverviewModel.DiscoveredDomain(
+                            rs.getString("DOMAIN_NAME"),
+                            rs.getString("DISCOVER_DATE"))
+                    );
+                }
+            } catch (SQLException ex) {
+                logger.warn("Failed to get recently added domains: {}", ex.getMessage());
+            }
+
+            cachedOverviewModel = new SiteOverviewModel(domains);
+
+            try {
+                TimeUnit.MINUTES.sleep(15);
+            } catch (InterruptedException e) {
+                Thread.currentThread().interrupt();
+                break;
+            }
+        }
     }
 
     public record SiteOverviewModel(List<DiscoveredDomain> domains) {
@@ -107,7 +137,7 @@ public class SearchSiteInfoService {
             @PathParam String domainName,
             @QueryParam String view,
             @QueryParam Integer page
-    ) throws SQLException {
+    ) throws SQLException, ExecutionException {
 
         if (null == domainName || domainName.isBlank()) {
             return null;
@@ -193,7 +223,7 @@ public class SearchSiteInfoService {
         );
     }
 
-    private SiteInfoWithContext listInfo(Context context, String domainName) {
+    private SiteInfoWithContext listInfo(Context context, String domainName) throws ExecutionException {
 
         var domain = new EdgeDomain(domainName);
         final int domainId = domainQueries.tryGetDomainId(domain).orElse(-1);

code/services-application/search-service/resources/jte/siteinfo/view/overview.jte

@@ -1,5 +1,4 @@
 @import nu.marginalia.db.DbDomainQueries
-@import nu.marginalia.model.EdgeDomain
 @import nu.marginalia.search.svc.SearchSiteInfoService
 @import nu.marginalia.search.svc.SearchSiteInfoService.*
 @import nu.marginalia.search.model.UrlDetails
@@ -81,35 +80,6 @@
 
         @endif
 
-
-        @if (!siteInfo.siblingDomains().isEmpty())
-            <div class="mx-3 flex place-items-baseline space-x-2 p-2 bg-gray-100 dark:bg-gray-600 rounded">
-                <i class="fas fa-globe"></i>
-                <span>Related Subdomains</span>
-            </div>
-
-            <table class="min-w-full divide-y divide-gray-200 dark:divide-gray-600 mx-4">
-                <thead>
-                <tr class="bg-gray-50 dark:bg-gray-700">
-                    <th scope="col" class="px-2 py-2 text-left text-xs font-medium text-gray-500 dark:text-gray-100 uppercase tracking-wider">Domain Name</th>
-                </tr>
-                </thead>
-                <tbody class="bg-white dark:bg-gray-800 divide-y divide-gray-200 dark:divide-gray-600 text-xs">
-                @for (DbDomainQueries.DomainWithNode sibling : siteInfo.siblingDomains())
-                    <tr>
-                        <td class="px-3 py-6 md:py-3 whitespace-nowrap">
-                            <a class="text-liteblue dark:text-blue-200" href="/site/${sibling.domain().toString()}">${sibling.domain().toString()}</a>
-
-                            @if (!sibling.isIndexed())
-                                <i class="ml-1 fa-regular fa-question-circle text-gray-400 dark:text-gray-600 text-xs" title="Not indexed"></i>
-                            @endif
-                        </td>
-                    </tr>
-                @endfor
-                </tbody>
-            </table>
-        @endif
-
         @if (siteInfo.domainInformation().isUnknownDomain())
             <div class="mx-3 flex place-items-baseline space-x-2 p-2 bg-gray-100 dark:bg-gray-600 rounded">
                 <i class="fa-regular fa-circle-question"></i>
@@ -178,6 +148,36 @@
             </form>
         @endif
 
+
+        @if (!siteInfo.siblingDomains().isEmpty())
+            <div class="mx-3 flex place-items-baseline space-x-2 p-2 bg-gray-100 dark:bg-gray-600 rounded">
+                <i class="fas fa-globe"></i>
+                <span>Related Subdomains</span>
+            </div>
+
+            <table class="min-w-full divide-y divide-gray-200 dark:divide-gray-600 mx-4">
+                <thead>
+                <tr class="bg-gray-50 dark:bg-gray-700">
+                    <th scope="col" class="px-2 py-2 text-left text-xs font-medium text-gray-500 dark:text-gray-100 uppercase tracking-wider">Domain Name</th>
+                </tr>
+                </thead>
+                <tbody class="bg-white dark:bg-gray-800 divide-y divide-gray-200 dark:divide-gray-600 text-xs">
+                @for (DbDomainQueries.DomainWithNode sibling : siteInfo.siblingDomains())
+                    <tr>
+                        <td class="px-3 py-6 md:py-3 whitespace-nowrap">
+                            <a class="text-liteblue dark:text-blue-200" href="/site/${sibling.domain().toString()}">${sibling.domain().toString()}</a>
+
+                            @if (!sibling.isIndexed())
+                                <i class="ml-1 fa-regular fa-question-circle text-gray-400 dark:text-gray-600 text-xs" title="Not indexed"></i>
+                            @endif
+                        </td>
+                    </tr>
+                @endfor
+                </tbody>
+            </table>
+        @endif
+
+
         @if (siteInfo.isKnown())
             <div class="mx-3 flex place-items-baseline space-x-2 p-2 bg-gray-100 dark:bg-gray-600 rounded">
                 <i class="fas fa-chart-simple"></i>