hamkadr/src/JobsMedical.Web/Pages/Admin/Ingested.cshtml.cs

using JobsMedical.Web.Data;
using JobsMedical.Web.Models;
using JobsMedical.Web.Services;
using Microsoft.AspNetCore.Authorization;
using Microsoft.AspNetCore.Mvc;
using Microsoft.AspNetCore.Mvc.RazorPages;
using Microsoft.EntityFrameworkCore;

namespace JobsMedical.Web.Pages.Admin;

/// <summary>Every crawled item with its outcome (queued / published / flagged / discarded),
/// filterable by status and source — the full audit trail of ingestion.</summary>
[Authorize(Roles = "Admin")]
public class IngestedModel : PageModel
{
    private readonly AppDbContext _db;
    public IngestedModel(AppDbContext db) => _db = db;

    public List<RawListing> Items { get; private set; } = new();
    public int Total { get; private set; }
    public Dictionary<RawListingStatus, int> Counts { get; private set; } = new();
    public List<SourceStat> SourceBreakdown { get; private set; } = new();
    [TempData] public string? Message { get; set; }

    /// <summary>Per-source tally: how many crawled vs how many actually published.</summary>
    public record SourceStat(string Source, int Total, int Published);

    [BindProperty(SupportsGet = true)] public string? Status { get; set; }   // new|flagged|published|discarded|all
    [BindProperty(SupportsGet = true)] public string? Source { get; set; }

    public async Task OnGetAsync()
    {
        Counts = await _db.RawListings.GroupBy(r => r.Status)
            .Select(g => new { g.Key, C = g.Count() }).ToDictionaryAsync(x => x.Key, x => x.C);

        // Per-source breakdown — group exact SourceChannel rows then fold into source "families"
        // (تلگرام/ch → تلگرام, وب‌سایت (host) → وب‌سایت) so the table reads one row per source.
        var bySource = await _db.RawListings.GroupBy(r => r.SourceChannel)
            .Select(g => new
            {
                Source = g.Key,
                Total = g.Count(),
                Published = g.Count(x => x.Status == RawListingStatus.Normalized),
            })
            .ToListAsync();
        SourceBreakdown = bySource
            .GroupBy(x => SourceFamily(x.Source))
            .Select(g => new SourceStat(g.Key, g.Sum(x => x.Total), g.Sum(x => x.Published)))
            .OrderByDescending(s => s.Published).ThenByDescending(s => s.Total)
            .ToList();

        var q = _db.RawListings.AsNoTracking().AsQueryable();

        var st = Status?.ToLowerInvariant() switch
        {
            "new" => (RawListingStatus?)RawListingStatus.New,
            "flagged" => RawListingStatus.Flagged,
            "published" => RawListingStatus.Normalized,
            "discarded" => RawListingStatus.Discarded,
            _ => null,
        };
        if (st is not null) q = q.Where(r => r.Status == st);
        if (!string.IsNullOrWhiteSpace(Source)) q = q.Where(r => r.SourceChannel.Contains(Source));

        Total = await q.CountAsync();
        Items = await q.OrderByDescending(r => r.FetchedAt).Take(200).ToListAsync();
    }

    /// <summary>Collapse a channel label to its source family: "تلگرام/nurses" → "تلگرام",
    /// "وب‌سایت (medjobs.ir)" → "وب‌سایت". Divar/Bale/Medjobs already have no suffix.</summary>
    private static string SourceFamily(string? channel)
    {
        if (string.IsNullOrWhiteSpace(channel)) return "نامشخص";
        var cut = channel.IndexOfAny(new[] { '/', '(' });
        return (cut > 0 ? channel[..cut] : channel).Trim();
    }

    /// <summary>
    /// ARCHIVE (never delete) everything published from ingestion: the aggregated Shift/Job/Talent
    /// posts are flipped to Archived (hidden from the site but kept for analytics); the raw crawl
    /// rows are retained untouched as the permanent archive.
    /// </summary>
    public async Task<IActionResult> OnPostArchivePublishedAsync()
    {
        var shifts = await _db.Shifts
            .Where(s => s.Source == ShiftSource.Aggregated && s.Status != ShiftStatus.Archived)
            .ExecuteUpdateAsync(u => u.SetProperty(s => s.Status, ShiftStatus.Archived));
        var jobs = await _db.JobOpenings
            .Where(j => j.Source == ShiftSource.Aggregated && j.Status != ShiftStatus.Archived)
            .ExecuteUpdateAsync(u => u.SetProperty(j => j.Status, ShiftStatus.Archived));
        var talent = await _db.TalentListings
            .Where(t => t.Source == ShiftSource.Aggregated && t.Status != ShiftStatus.Archived)
            .ExecuteUpdateAsync(u => u.SetProperty(t => t.Status, ShiftStatus.Archived));

        string P(int n) => JalaliDate.ToPersianDigits(n.ToString());
        Message = $"بایگانی شد (از سایت پنهان، در پایگاه‌داده نگه‌داری شد): {P(shifts)} شیفت، {P(jobs)} استخدام، {P(talent)} آماده‌به‌کار.";
        return RedirectToPage(new { Status });
    }
}