SocialPub/PrivaPub/Federation/Objects/ObjectShapes.cs

using PrivaPub.Models.Post;

using System.Globalization;
using System.Net;
using System.Text.Json.Nodes;
using System.Text.RegularExpressions;
using System.Xml;

using static PrivaPub.Federation.Objects.ActivityJson;

namespace PrivaPub.Federation.Objects
{
	public static partial class ObjectShapes
	{
		const int MaxVariants = 20;
		const int MaxCaptions = 20;
		const int MaxPlaces = 4;
		const int CoverMaxWidth = 1280;
		const int MaxEmojis = 64;
		const int MaxPollOptions = 20;
		const int MaxAudience = 32;

		static readonly Dictionary<string, string> MediaTypesByExtension = new(StringComparer.OrdinalIgnoreCase)
		{
			[".jpg"] = "image/jpeg", [".jpeg"] = "image/jpeg", [".png"] = "image/png", [".gif"] = "image/gif", [".webp"] = "image/webp",
			[".avif"] = "image/avif", [".heic"] = "image/heic", [".svg"] = "image/svg+xml",
			[".mp4"] = "video/mp4", [".m4v"] = "video/mp4", [".webm"] = "video/webm", [".mov"] = "video/quicktime", [".mkv"] = "video/x-matroska",
			[".mp3"] = "audio/mpeg", [".ogg"] = "audio/ogg", [".oga"] = "audio/ogg", [".opus"] = "audio/opus", [".m4a"] = "audio/mp4",
			[".flac"] = "audio/flac", [".wav"] = "audio/wav", [".m3u8"] = "application/x-mpegURL"
		};

		public static IEnumerable<JsonObject> Objects(JsonNode node) => node switch
		{
			JsonArray array => array.OfType<JsonObject>(),
			JsonObject single => new[] { single },
			_ => Enumerable.Empty<JsonObject>()
		};

		public static (string Url, string MediaType) MediaLink(JsonNode node)
		{
			switch (node)
			{
				case JsonValue value when value.TryGetValue<string>(out var text):
					return (text, default);
				case JsonObject link when Value(link, "href") is { } href:
					return (href, Value(link, "mediaType"));
				case JsonObject wrapper when wrapper["url"] is { } inner:
					var (url, mediaType) = MediaLink(inner);
					return (url, mediaType ?? Value(wrapper, "mediaType"));
				case JsonArray array:
					var links = array.Select(MediaLink).Where(l => l.Url != default).ToList();
					return links.FirstOrDefault(l => l.MediaType?.StartsWith("text/html", StringComparison.OrdinalIgnoreCase) != true);
				default:
					return default;
			}
		}

		public static string InferMediaType(string declared, string url, string objectType)
		{
			if (!string.IsNullOrWhiteSpace(declared) && declared.Contains('/'))
				return declared.Trim();
			if (Uri.TryCreate(url, UriKind.Absolute, out var uri) && MediaTypesByExtension.TryGetValue(Path.GetExtension(uri.AbsolutePath), out var byExtension))
				return byExtension;
			return objectType switch
			{
				"Image" => "image/*",
				"Video" => "video/*",
				"Audio" => "audio/*",
				_ => default
			};
		}

		public static string Image(JsonNode node)
		{
			var candidates = node switch
			{
				JsonValue value when value.TryGetValue<string>(out var text) => new[] { (Url: text, Width: (int?)default) },
				JsonArray array => array.Select(item => (Url: MediaLink(item).Url, Width: Int(item as JsonObject, "width"))).ToArray(),
				JsonObject single => new[] { (Url: MediaLink(single).Url, Width: Int(single, "width")) },
				_ => Array.Empty<(string Url, int? Width)>()
			};
			var usable = candidates.Where(c => Origin.Of(c.Url) != default).ToList();
			return usable.Where(c => c.Width <= CoverMaxWidth).OrderByDescending(c => c.Width).Select(c => c.Url).FirstOrDefault()
				?? usable.Select(c => c.Url).FirstOrDefault();
		}

		public static List<CustomEmoji> Emojis(JsonNode tags) =>
			Objects(tags)
				.Where(t => Value(t, "type") == "Emoji")
				.Select(t => new CustomEmoji
				{
					Shortcode = Value(t, "name")?.Trim(':'),
					URL = MediaLink(t["icon"]).Url,
					Updated = Moment(Value(t, "updated"))
				})
				.Where(e => !string.IsNullOrEmpty(e.Shortcode) && e.Shortcode.Length <= 100 && Shortcode().IsMatch(e.Shortcode) && Origin.Of(e.URL) != default)
				.DistinctBy(e => e.Shortcode)
				.Take(MaxEmojis)
				.ToList();

		public static InteractionRule QuotePolicy(JsonObject note)
		{
			if (note["interactionPolicy"] is not JsonObject policy || policy["canQuote"] is not JsonObject canQuote)
				return default;
			static List<string> Who(JsonNode node) => node switch
			{
				JsonArray array => array.Select(Id).Where(id => id != default).Take(MaxAudience).ToList(),
				JsonNode single when Id(single) is { } id => new List<string> { id },
				_ => new List<string>()
			};
			return new InteractionRule
			{
				Automatic = Who(canQuote["automaticApproval"] ?? canQuote["always"]),
				Manual = Who(canQuote["manualApproval"] ?? canQuote["approvalRequired"])
			};
		}

		public static PostPoll Poll(JsonObject note)
		{
			if (Value(note, "type") != "Question")
				return default;
			var multiple = note["anyOf"] is JsonArray;
			var options = Objects(note[multiple ? "anyOf" : "oneOf"])
				.Select(o => new PollOption
				{
					Title = Text(Value(o, "name"), 200),
					Votes = o["replies"] is JsonObject replies && Int(replies, "totalItems") is { } total ? total : Int(o, "_misskey_votes") ?? 0
				})
				.Where(o => o.Title != default)
				.Take(MaxPollOptions)
				.ToList();
			if (options.Count < 2)
				return default;
			var closedAt = note["closed"] switch
			{
				JsonValue value when value.TryGetValue<bool>(out var closed) => closed ? Moment(Value(note, "endTime")) ?? DateTime.UtcNow : default(DateTime?),
				JsonValue value when value.TryGetValue<string>(out var at) => Moment(at),
				_ => default
			};
			var expiresAt = Moment(Value(note, "endTime"));
			// Pleroma and Akkoma put an open poll's end in `closed`, often without `endTime`: it is closed once that has passed
			if (closedAt > DateTime.UtcNow)
			{
				expiresAt ??= closedAt;
				closedAt = default;
			}
			return new PostPoll
			{
				Options = options,
				Multiple = multiple,
				ExpiresAt = expiresAt,
				ClosedAt = closedAt,
				VotersCount = Int(note, "votersCount")
			};
		}

		public static string Cover(JsonObject note) =>
			Image(note["image"]) ?? Image(note["icon"]) ?? (note["preview"] is JsonObject preview && Value(preview, "type") == "Image" ? Image(preview) : default);

		public static double? Seconds(JsonNode node)
		{
			switch (node)
			{
				case JsonValue value when value.TryGetValue<double>(out var number):
					return number >= 0 && number < 1e7 ? number : default;
				case JsonValue value when value.TryGetValue<string>(out var text) && !string.IsNullOrWhiteSpace(text):
					try
					{
						var span = XmlConvert.ToTimeSpan(text.Trim());
						return span >= TimeSpan.Zero ? span.TotalSeconds : default;
					}
					catch (FormatException)
					{
						return double.TryParse(text, NumberStyles.Float, CultureInfo.InvariantCulture, out var seconds) && seconds >= 0 ? seconds : default;
					}
				default:
					return default;
			}
		}

		public static string Text(string html, int max)
		{
			if (string.IsNullOrWhiteSpace(html))
				return default;
			var text = WebUtility.HtmlDecode(Tags().Replace(html, " "));
			text = Spaces().Replace(text, " ").Trim();
			if (text.Length == 0)
				return default;
			return text.Length <= max ? text : text[..max];
		}

		public static string Language(JsonObject note) =>
			note["@context"] is JsonArray context
				? context.OfType<JsonObject>().Select(entry => Value(entry, "@language")).FirstOrDefault(language => !string.IsNullOrWhiteSpace(language))
				: default;

		public static PostSource Source(JsonObject note) => note["source"] is JsonObject source && Value(source, "content") is { } content
			? new PostSource { Content = content.Length <= 64 * 1024 ? content : content[..(64 * 1024)], MediaType = Value(source, "mediaType") }
			: default;

		public static PostLink Link(JsonObject note)
		{
			var attached = Objects(note["attachment"]).FirstOrDefault(a => Value(a, "type") == "Link" && IsWeb(Value(a, "href")));
			if (attached != default)
			{
				var preview = attached["preview"] as JsonObject;
				return new PostLink
				{
					Href = Value(attached, "href"),
					MediaType = Value(attached, "mediaType"),
					Title = Text(Value(preview, "name") ?? Value(attached, "name"), 300),
					Description = Text(Value(preview, "summary") ?? Value(preview, "content"), 1000),
					ImageURL = preview == default ? Image(note["image"]) : Image(preview["image"]) ?? Image(preview["icon"]),
					FromPublisher = preview != default
				};
			}
			if (note["source"] is JsonValue source && source.TryGetValue<string>(out var sourceUrl) && IsWeb(sourceUrl))
				return new PostLink { Href = sourceUrl, ImageURL = Image(note["image"]) };
			return default;
		}

		public static VideoDetails Video(JsonObject note)
		{
			if (Value(note, "type") != "Video")
				return default;
			var links = Objects(note["url"]).ToList();
			var nested = links.Where(l => Value(l, "mediaType")?.Equals("application/x-mpegURL", StringComparison.OrdinalIgnoreCase) == true)
				.SelectMany(l => Objects(l["tag"]))
				.ToList();
			return new VideoDetails
			{
				DurationSeconds = Seconds(note["duration"]),
				Variants = Variants(links.Concat(nested).ToList(), "video/", "audio/"),
				PlaylistURL = links.Where(l => Value(l, "mediaType")?.Equals("application/x-mpegURL", StringComparison.OrdinalIgnoreCase) == true)
					.Select(l => Value(l, "href")).FirstOrDefault(IsWeb),
				IsLive = Bool(note, "isLiveBroadcast") == true,
				State = Int(note, "state"),
				Views = Long(note, "views"),
				CommentsPolicy = Int(note, "commentsPolicy"),
				DownloadEnabled = Bool(note, "downloadEnabled"),
				Licence = Named(note["licence"] ?? note["license"]),
				Category = Named(note["category"]),
				ChannelURI = Objects(note["attributedTo"]).Where(a => Value(a, "type") == "Group").Select(a => Id(a)).FirstOrDefault(),
				EmbedURL = Value(note, "embedUrl") is { } embed && IsWeb(embed) ? embed : default,
				ChaptersURL = Value(note, "hasParts") is { } chapters && IsWeb(chapters) ? chapters : default,
				Captions = Objects(note["subtitleLanguage"])
					.Select(c => new Caption
					{
						Language = Value(c, "identifier"),
						Url = Objects(c["url"]).Where(u => Value(u, "mediaType") == "text/vtt").Select(u => Value(u, "href")).FirstOrDefault(IsWeb)
							?? (Value(c, "url") is { } plain && IsWeb(plain) ? plain : default)
					})
					.Where(c => c.Url != default)
					.Take(MaxCaptions)
					.ToList()
			};
		}

		public static AudioDetails Audio(JsonObject note)
		{
			if (Value(note, "type") != "Audio")
				return default;
			return new AudioDetails
			{
				DurationSeconds = Seconds(note["duration"]),
				Variants = Variants(Objects(note["url"]).ToList(), "audio/"),
				Album = Named(note["album"]),
				Position = Int(note, "position"),
				Licence = Named(note["license"] ?? note["licence"])
			};
		}

		public static EventDetails Event(JsonObject note)
		{
			if (Value(note, "type") != "Event")
				return default;
			var start = Value(note, "startTime");
			var places = Objects(note["location"]).Select(Place).Where(p => p != default).Take(MaxPlaces).ToList();
			return new EventDetails
			{
				Start = Moment(start),
				End = Moment(Value(note, "endTime")),
				TimeZone = Value(note, "timezone"),
				FloatingTime = start?.EndsWith("-00:00", StringComparison.Ordinal) == true,
				Places = places,
				OnlineURL = places.Select(p => p.Url).FirstOrDefault(u => u != default)
					?? Objects(note["attachment"]).Where(a => Value(a, "type") == "Link" && Value(a, "name") == "Website").Select(a => Value(a, "href")).FirstOrDefault(IsWeb),
				JoinMode = Value(note, "joinMode"),
				ParticipantCount = Int(note, "participantCount"),
				Capacity = Int(note, "maximumAttendeeCapacity"),
				Status = Value(note, "ical:status") ?? Value(note, "eventStatus") ?? Value(note, "status"),
				Category = Named(note["category"])
			};
		}

		static EventPlace Place(JsonObject location)
		{
			var type = Value(location, "type");
			if (type == "VirtualLocation")
				return Value(location, "url") is { } url && IsWeb(url) ? new EventPlace { Name = Text(Value(location, "name"), 200), Url = url } : default;
			var address = location["address"] switch
			{
				JsonValue value when value.TryGetValue<string>(out var text) => text,
				JsonObject postal => string.Join(", ", new[] { "streetAddress", "postalCode", "addressLocality", "addressRegion", "addressCountry" }
					.Select(field => Value(postal, field)).Where(part => !string.IsNullOrWhiteSpace(part))),
				_ => default
			};
			var place = new EventPlace
			{
				Name = Text(Value(location, "name") ?? Value(location, "content"), 200),
				Address = Text(address, 300),
				Latitude = Double(location, "latitude"),
				Longitude = Double(location, "longitude")
			};
			return place.Name == default && place.Address == default && place.Latitude == default ? default : place;
		}

		static List<MediaVariant> Variants(List<JsonObject> links, params string[] prefixes) =>
			links.Where(l => Value(l, "mediaType") is { } type && prefixes.Any(p => type.StartsWith(p, StringComparison.OrdinalIgnoreCase)) && IsWeb(Value(l, "href")))
				.Select(l => new MediaVariant
				{
					Url = Value(l, "href"),
					MediaType = Value(l, "mediaType"),
					Width = Int(l, "width"),
					Height = Int(l, "height"),
					Fps = Double(l, "fps"),
					Size = Long(l, "size"),
					Streams = Objects(l["attachment"])
						.Where(p => Value(p, "name") == "ffprobe_codec_type")
						.Select(p => Value(p, "value"))
						.Where(v => v is "audio" or "video")
						.Distinct()
						.ToList()
				})
				.Take(MaxVariants)
				.ToList();

		static string Named(JsonNode node) => node switch
		{
			JsonValue value when value.TryGetValue<string>(out var text) => Text(text, 200),
			JsonObject named => Text(Value(named, "name") ?? Value(named, "identifier"), 200),
			_ => default
		};

		static DateTime? Moment(string value) =>
			DateTimeOffset.TryParse(value, CultureInfo.InvariantCulture, DateTimeStyles.AssumeUniversal, out var parsed) ? parsed.UtcDateTime : (DateTime?)null;

		static bool IsWeb(string url) =>
			Uri.TryCreate(url, UriKind.Absolute, out var uri) && (uri.Scheme == Uri.UriSchemeHttps || uri.Scheme == Uri.UriSchemeHttp);

		static int? Int(JsonObject node, string property) =>
			node?[property] is JsonValue value && value.TryGetValue<double>(out var number) && number is >= 0 and < int.MaxValue ? (int)number : (int?)null;

		static long? Long(JsonObject node, string property) =>
			node?[property] is JsonValue value && value.TryGetValue<double>(out var number) && number is >= 0 and < long.MaxValue ? (long)number : (long?)null;

		static double? Double(JsonObject node, string property) =>
			node?[property] is JsonValue value && value.TryGetValue<double>(out var number) && double.IsFinite(number) ? number : (double?)null;

		static bool? Bool(JsonObject node, string property) =>
			node?[property] is JsonValue value && value.TryGetValue<bool>(out var flag) ? flag : (bool?)null;

		[GeneratedRegex(@"^[\w+\-.@]+$")]
		private static partial Regex Shortcode();

		[GeneratedRegex("<[^>]*>")]
		private static partial Regex Tags();

		[GeneratedRegex(@"\s+")]
		private static partial Regex Spaces();
	}
}