hamkadr/src/JobsMedical.Web/Services/Scraping/AiAuditor.cs

using System.Net.Http.Headers;
using System.Text;
using System.Text.Json;
using JobsMedical.Web.Models;

namespace JobsMedical.Web.Services.Scraping;

public record AiStructured(
    string? Kind, string? Role, string? City, string? District, string? ShiftType,
    string? EmploymentType, long? PayAmount, int? SharePercent, string? Title, string? FacilityName);

/// <summary>An AI verdict on a raw listing.</summary>
public record AiAuditResult(string Decision, int Confidence, string? Reason, AiStructured? Data)
{
    public bool Approve => Decision.Equals("approve", StringComparison.OrdinalIgnoreCase);
    public bool Reject => Decision.Equals("reject", StringComparison.OrdinalIgnoreCase);
}

public interface IAiAuditor
{
    /// <summary>Audit a raw post. Returns null when AI is off or the call fails (fail safe → manual).</summary>
    Task<AiAuditResult?> AuditAsync(string rawText, AppSetting settings, CancellationToken ct = default);
}

/// <summary>
/// Calls any OpenAI-compatible chat-completions endpoint (self-hosted vLLM/Ollama, or an Iranian
/// provider — OpenAI/Anthropic are blocked from Iran). The admin-set system prompt is the
/// "framework" that tells the model how to approve/reject/structure. We ask for strict JSON and
/// parse it. Any failure returns null so ingestion falls back to the rule-based path.
/// </summary>
public class OpenAiCompatibleAuditor : IAiAuditor
{
    private readonly IHttpClientFactory _http;
    private readonly ILogger<OpenAiCompatibleAuditor> _log;

    public OpenAiCompatibleAuditor(IHttpClientFactory http, ILogger<OpenAiCompatibleAuditor> log)
    {
        _http = http;
        _log = log;
    }

    public async Task<AiAuditResult?> AuditAsync(string rawText, AppSetting s, CancellationToken ct = default)
    {
        if (!s.AiEnabled || string.IsNullOrWhiteSpace(s.AiEndpoint)) return null;

        try
        {
            var payload = new
            {
                model = string.IsNullOrWhiteSpace(s.AiModel) ? "gpt-4o-mini" : s.AiModel,
                temperature = 0,
                response_format = new { type = "json_object" },
                messages = new object[]
                {
                    new { role = "system", content = s.AiSystemPrompt },
                    new { role = "user", content = "آگهی خام:\n" + rawText + "\n\nفقط با JSON پاسخ بده." },
                },
            };

            var client = _http.CreateClient("ai");
            client.Timeout = TimeSpan.FromSeconds(30);
            using var req = new HttpRequestMessage(HttpMethod.Post, s.AiEndpoint)
            {
                Content = new StringContent(JsonSerializer.Serialize(payload), Encoding.UTF8, "application/json"),
            };
            if (!string.IsNullOrWhiteSpace(s.AiApiKey))
                req.Headers.Authorization = new AuthenticationHeaderValue("Bearer", s.AiApiKey);

            using var resp = await client.SendAsync(req, ct);
            resp.EnsureSuccessStatusCode();
            var body = await resp.Content.ReadAsStringAsync(ct);

            using var doc = JsonDocument.Parse(body);
            var content = doc.RootElement
                .GetProperty("choices")[0].GetProperty("message").GetProperty("content").GetString();
            if (string.IsNullOrWhiteSpace(content)) return null;

            return ParseVerdict(content);
        }
        catch (Exception ex)
        {
            _log.LogWarning(ex, "AI audit failed — falling back to rule-based decision.");
            return null;
        }
    }

    private static AiAuditResult? ParseVerdict(string json)
    {
        // The content itself should be a JSON object; tolerate code fences.
        json = json.Trim().Trim('`');
        var start = json.IndexOf('{');
        var end = json.LastIndexOf('}');
        if (start < 0 || end <= start) return null;
        json = json.Substring(start, end - start + 1);

        using var doc = JsonDocument.Parse(json);
        var r = doc.RootElement;
        string? S(string k) => r.TryGetProperty(k, out var v) && v.ValueKind == JsonValueKind.String ? v.GetString() : null;
        int I(string k, int d) => r.TryGetProperty(k, out var v) && v.TryGetInt32(out var n) ? n : d;
        long? L(string k) => r.TryGetProperty(k, out var v) && v.TryGetInt64(out var n) ? n : null;
        int? NI(string k) => r.TryGetProperty(k, out var v) && v.TryGetInt32(out var n) ? n : null;

        var decision = (S("decision") ?? "review").ToLowerInvariant();
        var data = new AiStructured(S("kind"), S("role"), S("city"), S("district"), S("shiftType"),
            S("employmentType"), L("payAmount"), NI("sharePercent"), S("title"), S("facilityName"));
        return new AiAuditResult(decision, Math.Clamp(I("confidence", 50), 0, 100), S("reason"), data);
    }
}