using System.Text.Json; using CvMatcher.Models.Requests; using CvSearch.Data; using CvSearch.Data.Entities; using CvMatcher.Models.Settings; using CvSearchJob.Clients; using CvSearchJob.Services; using JobScheduler.Tasks; using Microsoft.EntityFrameworkCore; using Microsoft.Extensions.Configuration; using Microsoft.Extensions.DependencyInjection; using Microsoft.Extensions.Logging; using Microsoft.Extensions.Options; namespace CvSearchJob.Tasks; public sealed class CvSearchJobTask : IJobTask { private readonly IServiceScopeFactory _scopeFactory; private readonly JobSearchSettings _settings; private readonly HtmlJobSearcher _searcher; private readonly ICvMatcherInternalApi _matcherApi; private readonly CvSearchEmailSender _emailSender; private readonly ILogger _logger; public string TaskType => "CvSearch"; public CvSearchJobTask( IServiceScopeFactory scopeFactory, IOptions settings, HtmlJobSearcher searcher, ICvMatcherInternalApi matcherApi, CvSearchEmailSender emailSender, ILogger logger) { _scopeFactory = scopeFactory; _settings = settings.Value; _searcher = searcher; _matcherApi = matcherApi; _emailSender = emailSender; _logger = logger; } public async Task ExecuteAsync(IConfiguration parametersSection, CancellationToken cancellationToken) { if (!_settings.Enabled) return; using var scope = _scopeFactory.CreateScope(); var db = scope.ServiceProvider.GetRequiredService(); // Recover orphaned Processing sessions (container crashed mid-run) var stuckCutoff = DateTime.UtcNow.AddMinutes(-10); var stuckSessions = await db.JobSearchSessions .Where(s => s.Status == JobSearchStatus.Processing && s.CreatedAt < stuckCutoff) .ToListAsync(cancellationToken); foreach (var stuck in stuckSessions) { stuck.Status = JobSearchStatus.Pending; _logger.LogWarning("Reset stuck session {SessionId} back to Pending", stuck.Id); } if (stuckSessions.Count > 0) await db.SaveChangesAsync(cancellationToken); var pending = await db.JobSearchSessions .Where(s => s.Status == JobSearchStatus.Pending) .OrderBy(s => s.CreatedAt) .Take(1) .FirstOrDefaultAsync(cancellationToken); if (pending is null) return; _logger.LogInformation("Processing job search session {SessionId}", pending.Id); pending.Status = JobSearchStatus.Processing; await db.SaveChangesAsync(cancellationToken); try { var results = await RunSearchAsync(pending, db, cancellationToken); pending.Status = JobSearchStatus.Done; await db.SaveChangesAsync(cancellationToken); var attachmentFileName = BuildCvFileName(pending.CvDocumentId); await _emailSender.SendResultsAsync(pending.Email, attachmentFileName, results, pending.Language, cancellationToken); _logger.LogInformation("Session {SessionId} done. {Count} results sent.", pending.Id, results.Count); } catch (Exception ex) { _logger.LogError(ex, "Session {SessionId} failed.", pending.Id); pending.Status = JobSearchStatus.Failed; await db.SaveChangesAsync(cancellationToken); } } private async Task> RunSearchAsync( JobSearchSessionEntity session, CvSearchDbContext db, CancellationToken ct) { var cvKeywords = session.Keywords .Split(',', StringSplitOptions.RemoveEmptyEntries) .Select(k => k.Trim()) .Where(k => k.Length > 0) .ToList(); var providers = GetProviders(session.ProviderConfigJson); var jobUrls = new HashSet(StringComparer.OrdinalIgnoreCase); foreach (var provider in providers) { var urls = await _searcher.SearchJobUrlsAsync(provider, cvKeywords, ct); foreach (var url in urls) jobUrls.Add(url); } var candidates = jobUrls.Take(_settings.MaxJobsToMatch).ToList(); _logger.LogInformation("Session {SessionId}: {Count} candidate job URLs to match", session.Id, candidates.Count); var results = new List(); foreach (var url in candidates) { try { var matchRequest = new MatchJobRequest { CvDocumentId = session.CvDocumentId, JobUrl = url, // User already gave GDPR consent when they clicked the one-time job search link GdprConsent = true }; var matchResult = await _matcherApi.MatchJobAsync(matchRequest, ct); if (matchResult.Score < _settings.MinMatchScore) { _logger.LogDebug("Session {SessionId}: {Url} scored {Score}% (below threshold)", session.Id, url, matchResult.Score); continue; } var entity = new JobSearchResultEntity { Id = Guid.NewGuid().ToString("N"), SessionId = session.Id, ProviderName = GuessProvider(url, providers), JobUrl = url, JobTitle = matchResult.Summary.Split('.').FirstOrDefault()?.Trim() ?? "Job", JobText = string.Empty, Score = matchResult.Score, ResultJson = JsonSerializer.Serialize(matchResult, new JsonSerializerOptions(JsonSerializerDefaults.Web)), CreatedAt = DateTime.UtcNow }; db.JobSearchResults.Add(entity); await db.SaveChangesAsync(ct); results.Add(entity); } catch (Exception ex) { _logger.LogWarning(ex, "Session {SessionId}: match failed for {Url}", session.Id, url); } } results.Sort((a, b) => b.Score.CompareTo(a.Score)); return results; } private List GetProviders(string? providerConfigJson) { if (string.IsNullOrWhiteSpace(providerConfigJson)) return _settings.Providers.Where(p => p.Enabled).ToList(); try { return JsonSerializer.Deserialize>(providerConfigJson, new JsonSerializerOptions(JsonSerializerDefaults.Web)) ?? _settings.Providers.Where(p => p.Enabled).ToList(); } catch { return _settings.Providers.Where(p => p.Enabled).ToList(); } } private static string GuessProvider(string url, List providers) { foreach (var p in providers) { if (!string.IsNullOrWhiteSpace(p.JobLinkContains) && url.Contains(p.JobLinkContains, StringComparison.OrdinalIgnoreCase)) return p.Name; } return Uri.TryCreate(url, UriKind.Absolute, out var uri) ? uri.Host : "unknown"; } private static string BuildCvFileName(string cvDocumentId) { // Strip non-alphanumeric characters so the filename is safe for all OS/email clients. var safeId = string.Concat(cvDocumentId.Where(char.IsLetterOrDigit)); if (string.IsNullOrWhiteSpace(safeId)) safeId = "cv"; return $"{safeId}.pdf"; } }