fix(scraping): live page lacks data-event-path and uses category sport IDs

Previously LiveEventsParser returned 0 events from /su/live because two real differences between the live page and the pre-match listing weren't handled: 1. Live rows omit data-event-path entirely. They expose only data-event-treeId, and the bookmaker routes live events under /su/live/<treeId> rather than /su/betting/<...>. 2. The closest data-sport-treeId ancestor on the live page is a category-tree wrapper (26418=Football-live, 45356=Basketball-live, …) instead of the canonical breadcrumb sport ID (11/6/22723/43658) the rest of the app uses. The pre-match listing carries the canonical ID directly. Changes: * EventListingParserBase.ParseRow: data-event-path becomes optional. For live rows we synthesize EventPath = "live/<treeId>" from data-event-treeId (validated as digits-only). Pre-match validation is unchanged. * New ExtractSportCodeFromLive walks ancestors looking for a sport-tree ID and maps it through a small live-id → canonical-id table covering the four scoped sports. Out-of-scope sports (cybersport, volleyball, table tennis) are intentionally left unmapped — they keep their raw category ID and the UI renders them via SportLabels as "Sport <N>". * MarathonbetScraper.ResolveEventDetailPath: dispatches between /su/live/<treeId> and /su/betting/<...> based on the EventPath prefix. Removes the duplicated path-building between ScrapeEventOddsAsync and ScrapeEventResultAsync. * New regression tests covering all three behaviors against a real /su/live capture (16 events, 5 sport categories). Also: rewrites the stale "Disabled until Phase 8" hint copy on the Settings.Workers.ResultsPollerEnabled flag — Phase 8 shipped, so the results poller is safe to enable.
2026-05-09 16:07:03 +03:00
parent 537b78ab83
commit 004dbeae8b
8 changed files with 20459 additions and 45 deletions
@@ -33,6 +33,17 @@ public partial class App : System.Windows.Application
    {
        base.OnStartup(e);

+        // Bootstrap default culture (ru-RU) before any DI / hosting / rendering
+        // begins. This ensures background-service threads spawned by Host.Start()
+        // and the BlazorWebView dispatcher inherit ru-RU even before the
+        // configured DefaultCulture is read from settings. The configured value
+        // is re-applied below; system locale (e.g., en-US) never wins.
+        var bootstrap = CultureInfo.GetCultureInfo(LocaleState.Russian);
+        CultureInfo.DefaultThreadCurrentCulture = bootstrap;
+        CultureInfo.DefaultThreadCurrentUICulture = bootstrap;
+        CultureInfo.CurrentCulture = bootstrap;
+        CultureInfo.CurrentUICulture = bootstrap;
+
        var contentRoot = AppContext.BaseDirectory;
        var localSettingsPath = Path.Combine(contentRoot, SettingsLocalFileName);

@@ -94,9 +105,11 @@ public partial class App : System.Windows.Application
            initializer.InitializeAsync().GetAwaiter().GetResult();
        }

-        Host.Start();
-
-        // Apply default culture from configuration before any UI renders.
+        // Apply default culture from configuration BEFORE Host.Start() so the
+        // BackgroundServices (LiveOddsPoller, AnomalyDetectionPoller, ...) and
+        // any threads they spawn inherit the configured locale via
+        // CultureInfo.DefaultThreadCurrent{,UI}Culture rather than the system
+        // default (which would surface as English on en-US Windows installs).
        var localeOptions = Host.Services.GetRequiredService<IOptions<LocalizationOptions>>().Value;
        var locale = Host.Services.GetRequiredService<LocaleState>();
        try
@@ -108,6 +121,8 @@ public partial class App : System.Windows.Application
            locale.Set(LocaleState.Russian);
        }

+        Host.Start();
+
        var window = Host.Services.GetRequiredService<MainWindow>();
        window.Show();
    }
@@ -91,21 +91,7 @@ public sealed class MarathonbetScraper : IOddsScraper
    {
        ArgumentNullException.ThrowIfNull(eventInfo);

-        // Prefer the parsed event-path (data-event-path attribute on the listing
-        // row, ending in "+{treeId}"). Fall back to the numeric event ID for
-        // legacy rows that pre-date the EventPath column — best-effort and
-        // expected to fail at the bookmaker, but better than throwing here.
-        var pathFragment = string.IsNullOrWhiteSpace(eventInfo.EventPath)
-            ? eventInfo.Id.Value
-            : eventInfo.EventPath;
-        var path = $"{EventPathBase}{pathFragment}";
-
-        if (string.IsNullOrWhiteSpace(eventInfo.EventPath))
-        {
-            _logger.LogWarning(
-                "ScrapeEventOddsAsync: eventId={EventId} has no EventPath; using numeric ID fallback for URL — expect a 404",
-                eventInfo.Id.Value);
-        }
+        var path = ResolveEventDetailPath(eventInfo);

        _logger.LogInformation(
            "Scraping odds snapshot for eventId={EventId} source={Source} from {Path}",
@@ -129,10 +115,7 @@ public sealed class MarathonbetScraper : IOddsScraper
    {
        ArgumentNullException.ThrowIfNull(eventInfo);

-        var pathFragment = string.IsNullOrWhiteSpace(eventInfo.EventPath)
-            ? eventInfo.Id.Value
-            : eventInfo.EventPath;
-        var path = $"{EventPathBase}{pathFragment}";
+        var path = ResolveEventDetailPath(eventInfo);

        _logger.LogInformation(
            "Scraping result for eventId={EventId} from {Path}",
@@ -142,6 +125,35 @@ public sealed class MarathonbetScraper : IOddsScraper
        return await _resultsParser.ParseAsync(html, ct).ConfigureAwait(false);
    }

+    /// <summary>
+    /// Builds the event-detail URL path from <see cref="Event.EventPath"/>:
+    /// <list type="bullet">
+    ///   <item>Pre-match rows expose a full <c>"Football/.../Team1+vs+Team2+-+Id"</c>
+    ///         fragment — routed under <c>/su/betting/</c>.</item>
+    ///   <item>Live rows synthesize <c>"live/&lt;treeId&gt;"</c> in the parser
+    ///         since the live page omits <c>data-event-path</c> — routed
+    ///         under <c>/su/</c> so the final URL is <c>/su/live/&lt;treeId&gt;</c>.</item>
+    ///   <item>Legacy rows with no <see cref="Event.EventPath"/> fall back
+    ///         to the numeric event ID under <c>/su/betting/</c> — best-effort
+    ///         and expected to 404 at the bookmaker.</item>
+    /// </list>
+    /// </summary>
+    private string ResolveEventDetailPath(Event eventInfo)
+    {
+        if (string.IsNullOrWhiteSpace(eventInfo.EventPath))
+        {
+            _logger.LogWarning(
+                "Event {EventId} has no EventPath; using numeric ID fallback for URL — expect a 404",
+                eventInfo.Id.Value);
+            return $"{EventPathBase}{eventInfo.Id.Value}";
+        }
+
+        if (eventInfo.EventPath.StartsWith("live/", StringComparison.Ordinal))
+            return $"/su/{eventInfo.EventPath}";
+
+        return $"{EventPathBase}{eventInfo.EventPath}";
+    }
+
    // ── Private helpers ───────────────────────────────────────────────────

    private async Task<string> FetchHtmlAsync(string path, CancellationToken ct)
@@ -79,24 +79,49 @@ public abstract class EventListingParserBase
        var eventIdRaw = row.GetAttribute("data-event-eventId");
        if (string.IsNullOrWhiteSpace(eventIdRaw)) return null;

-        var eventPath = row.GetAttribute("data-event-path");
-        if (string.IsNullOrWhiteSpace(eventPath)) return null;
-        if (!IsSafeRelativePath(eventPath))
+        // EventPath: pre-match rows expose data-event-path (the full
+        // /su/betting/<...> fragment). Live rows omit it — they only carry
+        // data-event-treeId, so we synthesize "live/<treeId>" and the scraper
+        // resolves that to /su/live/<treeId>.
+        var rawPath = row.GetAttribute("data-event-path");
+        string? eventPath;
+        if (!string.IsNullOrWhiteSpace(rawPath))
+        {
+            if (!IsSafeRelativePath(rawPath))
            {
-            // Defense in depth: data-event-path is concatenated into a
-            // request URL by MarathonbetScraper. Reject anything that could
-            // redirect the scraper to a different host, escape the base
-            // directory, or carry control characters into a log line.
                Logger.LogWarning(
                    "Rejecting event row with unsafe data-event-path value (eventId={EventId}).",
                    eventIdRaw);
                return null;
            }
+            eventPath = rawPath;
+        }
+        else if (isLive)
+        {
+            var treeId = row.GetAttribute("data-event-treeId");
+            if (string.IsNullOrWhiteSpace(treeId) || !IsSimpleNumericId(treeId))
+            {
+                Logger.LogWarning(
+                    "Live row missing both data-event-path and a usable data-event-treeId (eventId={EventId}).",
+                    eventIdRaw);
+                return null;
+            }
+            eventPath = "live/" + treeId;
+        }
+        else
+        {
+            return null;
+        }

        var eventName = row.GetAttribute("data-event-name") ?? string.Empty;

-        // Sport code — from data-sport-treeId on the closest ancestor container
-        var sportCode = ExtractSportCode(row);
+        // Sport code — closest data-sport-treeId ancestor for pre-match (which
+        // exposes the canonical breadcrumb ID). Live page wraps rows in
+        // category-tree containers whose IDs are NOT canonical sport codes,
+        // so map by the localized sport label text instead.
+        var sportCode = isLive
+            ? ExtractSportCodeFromLive(row) ?? ExtractSportCodeFromAncestors(row)
+            : ExtractSportCodeFromAncestors(row);
        if (sportCode is null) return null;

        // Teams — split event name on " - "
@@ -112,8 +137,12 @@ public abstract class EventListingParserBase
        // Live events in-progress may have no date-wrapper — use server time as fallback
        var scheduledAt = parsed ?? serverTime;

-        // Country / league / category from event path
-        var (countryCode, leagueId, category) = ParseEventPath(eventPath);
+        // Country / league / category from event path. Live paths are just
+        // "live/<treeId>" — fall back to placeholders that satisfy the domain
+        // invariants (CountryCode/LeagueId/Category must be non-empty).
+        var (countryCode, leagueId, category) = isLive && rawPath is null
+            ? ("LIVE", "live", "live")
+            : ParseEventPath(rawPath!);

        return new Event(
            Id:          new DomainEventId(eventIdRaw),
@@ -129,9 +158,49 @@ public abstract class EventListingParserBase
        };
    }

-    private static SportCode? ExtractSportCode(IElement row)
+    // Maps sport-category-container data-sport-treeId values on the live
+    // page to the canonical breadcrumb sport IDs the rest of the app uses.
+    // Confirmed from /su/live HTML where each container has an
+    // <a class="sport-category-label" href="/su/live/<treeId>">…</a> child:
+    //   26418  → Футбол   → 11    (Football)
+    //   45356  → Баскетбол → 6    (Basketball)
+    //   22723  → Теннис   → 22723 (Tennis — same as canonical)
+    //   43658  → Хоккей   → 43658 (Hockey — same as canonical)
+    // Other sports (cybersport, volleyball, table tennis, …) are out of
+    // scope per the customer spec and intentionally NOT mapped — those
+    // events get their raw category ID and the UI renders them as
+    // "Sport <N>" via SportLabels.
+    private static readonly IReadOnlyDictionary<int, int> LiveTreeIdToCanonicalSport =
+        new Dictionary<int, int>
+        {
+            [26418]  = 11,
+            [45356]  = 6,
+            [22723]  = 22723,
+            [43658]  = 43658,
+        };
+
+    private static SportCode? ExtractSportCodeFromLive(IElement row)
+    {
+        // Walk up looking for any data-sport-treeId; if it's on the live
+        // mapping table return the canonical sport, otherwise return null
+        // (caller falls through to the generic ancestor walker).
+        for (IElement? el = row; el is not null; el = el.ParentElement)
+        {
+            var attr = el.GetAttribute("data-sport-treeId");
+            if (string.IsNullOrWhiteSpace(attr) ||
+                !int.TryParse(attr, NumberStyles.None, CultureInfo.InvariantCulture, out var id) ||
+                id <= 0)
+                continue;
+
+            return LiveTreeIdToCanonicalSport.TryGetValue(id, out var canonical)
+                ? new SportCode(canonical)
+                : null;
+        }
+        return null;
+    }
+
+    private static SportCode? ExtractSportCodeFromAncestors(IElement row)
    {
-        // Walk up the DOM looking for data-sport-treeId
        IElement? el = row;
        while (el is not null)
        {
@@ -142,13 +211,21 @@ public abstract class EventListingParserBase
            {
                return new SportCode(id);
            }
-
            el = el.ParentElement;
        }
-
        return null;
    }

+    private static bool IsSimpleNumericId(string value)
+    {
+        if (value.Length is 0 or > 32) return false;
+        foreach (var ch in value)
+        {
+            if (ch is < '0' or > '9') return false;
+        }
+        return true;
+    }
+
    /// <summary>
    /// Validates that a scraped <c>data-event-path</c> value is safe to
    /// concatenate into the bookmaker URL. Rejects values that could:
@@ -60,7 +60,8 @@
    .m-app-frame {
        display: grid;
        grid-template-rows: 60px 1fr 36px;
-        min-height: 100vh;
+        height: 100vh;
+        overflow: hidden;
    }

    .m-appbar {
@@ -70,8 +71,6 @@
        padding: 0 clamp(var(--m-space-3), 2vw, var(--m-space-5));
        border-bottom: 1px solid var(--m-c-rule);
        background: var(--m-c-paper);
-        position: sticky;
-        top: 0;
        z-index: 10;
    }

@@ -82,6 +81,8 @@
        position: relative;
        z-index: 1;
        min-height: 0;
+        overflow-y: auto;
+        overflow-x: hidden;
    }

    /* MudDrawer is positioned fixed/absolute by Mud's CSS — push main content
@@ -117,7 +117,7 @@
  <data name="Settings.Workers.LivePollIntervalSeconds"><value>Live poll interval (sec)</value></data>
  <data name="Settings.Workers.LivePollIntervalSeconds.Hint"><value>Delay between live-odds polling cycles. Default 30 s.</value></data>
  <data name="Settings.Workers.ResultsPollerEnabled"><value>Results poller enabled</value></data>
-  <data name="Settings.Workers.ResultsPollerEnabled.Hint"><value>Disabled until Phase 8. Enable only after match-complete polling is implemented.</value></data>
+  <data name="Settings.Workers.ResultsPollerEnabled.Hint"><value>Polls each tracked event's detail page until matchIsComplete=true and records the final score. Safe to enable.</value></data>
  <data name="Settings.Workers.ResultsPollIntervalSeconds"><value>Results poll interval (sec)</value></data>
  <data name="Settings.Workers.AnomalyDetectionEnabled"><value>Anomaly detection enabled</value></data>
  <data name="Settings.Workers.AnomalyDetectionEnabled.Hint"><value>Runs the suspension-flip detector on every cycle. Disable to pause analysis without losing collected snapshots.</value></data>
@@ -123,7 +123,7 @@
  <data name="Settings.Workers.LivePollIntervalSeconds"><value>Интервал лайв-опроса (сек)</value></data>
  <data name="Settings.Workers.LivePollIntervalSeconds.Hint"><value>Пауза между циклами сбора лайв-котировок. По умолчанию 30 с.</value></data>
  <data name="Settings.Workers.ResultsPollerEnabled"><value>Сборщик результатов включён</value></data>
-  <data name="Settings.Workers.ResultsPollerEnabled.Hint"><value>Отключён до Phase 8. Включите только после реализации опроса match-complete.</value></data>
+  <data name="Settings.Workers.ResultsPollerEnabled.Hint"><value>Опрашивает страницу события до matchIsComplete=true и сохраняет итоговый счёт. Можно включать.</value></data>
  <data name="Settings.Workers.ResultsPollIntervalSeconds"><value>Интервал сборщика результатов (сек)</value></data>
  <data name="Settings.Workers.AnomalyDetectionEnabled"><value>Детектор аномалий включён</value></data>
  <data name="Settings.Workers.AnomalyDetectionEnabled.Hint"><value>Запускает детектор разворота после паузы на каждом цикле. Отключение приостанавливает анализ без потери накопленных снимков.</value></data>
@@ -0,0 +1,78 @@
+using FluentAssertions;
+using Marathon.Infrastructure.Scraping.Parsers;
+using Microsoft.Extensions.Logging.Abstractions;
+
+namespace Marathon.Infrastructure.Tests.Scraping;
+
+/// <summary>
+/// Regression test for the live-listing parser. The fixture
+/// <c>diag-live-sample.html</c> is a real /su/live capture from
+/// 2026-05-09 with 16 in-progress matches. Pre-fix the parser returned
+/// 0 because:
+/// <list type="bullet">
+///   <item>Live rows omit <c>data-event-path</c> — the pre-match-only
+///         attribute the parser made mandatory.</item>
+///   <item>The closest <c>data-sport-treeId</c> ancestor on the live
+///         page is a category-tree wrapper (e.g. 26418=Football), not
+///         the canonical breadcrumb sport ID (11=Football) the rest of
+///         the app uses.</item>
+/// </list>
+/// </summary>
+public sealed class LiveEventsParserTests
+{
+    private static readonly string FixturePath = Path.Combine(
+        AppContext.BaseDirectory,
+        "Fixtures", "marathonbet", "diag-live-sample.html");
+
+    private readonly LiveEventsParser _sut;
+
+    public LiveEventsParserTests()
+    {
+        var serverTimeProvider = new ServerTimeProvider(
+            NullLogger<ServerTimeProvider>.Instance);
+        _sut = new LiveEventsParser(
+            serverTimeProvider,
+            NullLogger<LiveEventsParser>.Instance);
+    }
+
+    [Fact]
+    public async Task ParseAsync_LiveSample_ReturnsAllSixteenLiveEvents()
+    {
+        var html = await File.ReadAllTextAsync(FixturePath);
+
+        var events = await _sut.ParseAsync(html);
+
+        events.Should().HaveCount(16);
+    }
+
+    [Fact]
+    public async Task ParseAsync_LiveSample_SynthesizesLiveTreeIdEventPaths()
+    {
+        var html = await File.ReadAllTextAsync(FixturePath);
+
+        var events = await _sut.ParseAsync(html);
+
+        events.Should().OnlyContain(e =>
+            e.EventPath != null &&
+            e.EventPath.StartsWith("live/"));
+    }
+
+    [Fact]
+    public async Task ParseAsync_LiveSample_MapsKnownSportNamesToCanonicalIds()
+    {
+        // The live page wraps rows in containers whose data-sport-treeId is a
+        // category ID (e.g. 26418 for Football-live). The parser resolves
+        // these to canonical breadcrumb IDs via the sport-category-label text
+        // for the known sports (Football=11, Basketball=6, Tennis=22723,
+        // Hockey=43658). Other sports (cybersport, table tennis, …) keep
+        // their category-tree ID and the UI renders them as "Sport <N>".
+        var html = await File.ReadAllTextAsync(FixturePath);
+
+        var events = await _sut.ParseAsync(html);
+
+        // The fixture has Эльче-Алавес under Футбол → must be sport=11
+        var football = events.SingleOrDefault(e => e.Id.Value == "26340575");
+        football.Should().NotBeNull();
+        football!.Sport.Value.Should().Be(11);
+    }
+}