rust/forkmessager
1
0
Fork 0
Code Issues 2 Pull Requests Actions Packages Projects 3 Releases Wiki Activity

Правка импорта

Browse Source
This commit is contained in:
Халимов Рустам
2026-04-06 22:35:54 +03:00
parent 1558b20470
commit d96e4ec7d4
3 changed files with 56 additions and 54 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
backend/src/Modules/Conversations/Application/Messages/GetSharedMedia/GetSharedMediaQuery.cs
View File

@@ -102,7 +102,7 @@ internal sealed class GetSharedMediaQueryHandler : IQueryHandler<GetSharedMediaQ
if (filterType == "files")
{
return mType != "image" && mType != "video" && mType != "link";
return mType != "image" && mType != "video" && mType != "link" && !isGif;
}
if (filterType == "media")

96
backend/src/Modules/TelegramImport/Infrastructure/Parser/TelegramHtmlParser.cs
View File

@@ -43,15 +43,46 @@ public sealed class TelegramHtmlParser : ITelegramHtmlParser
var doc = await _parser.ParseDocumentAsync(htmlStream, ct);
var messages = new List<TelegramMessage>();
string? lastSenderName = null;
DateTime? lastCreatedAt = null;
var messageNodes = doc.QuerySelectorAll(".message");
foreach (var node in messageNodes)
{
// STOP ALL MERGING OF SEPARATE NODES.
// Telegram HTML nodes represent visual bubbles. If we merge them, we break the original hierarchy
// and often misplace captions or hide media.
bool isJoined = node.ClassList.Contains("joined");
var media = ExtractMedia(node, baseDirInZip);
var textNode = node.QuerySelector(".text");
var hasText = textNode != null && !string.IsNullOrWhiteSpace(textNode.TextContent);
var dateNode = node.QuerySelector(".date[title]") ?? node.QuerySelector("[title]");
var dateStr = dateNode?.GetAttribute("title") ?? "";
DateTime? currentAt = ParseDate(dateStr);
// REFINED ALBUM MERGING LOGIC:
// 1. Current is .joined
// 2. Current has media but NO text (important: if it has text, it's a new bubble group)
// 3. Sender is the same as previous message
// 4. Time difference is minimal (<= 2 seconds). Telegram albums usually share the exact same timestamp.
if (isJoined && media.Count > 0 && !hasText && messages.Count > 0)
{
var prev = messages[messages.Count - 1];
bool isSameSender = lastSenderName != null; // Since it's 'joined', it's the same visual group
bool isCloseInTime = lastCreatedAt.HasValue && currentAt.HasValue &&
Math.Abs((currentAt.Value - lastCreatedAt.Value).TotalSeconds) <= 2;
if (isSameSender && isCloseInTime)
{
var mergedMedia = (prev.Media ?? new List<TelegramMedia>()).Concat(media).ToList();
messages[messages.Count - 1] = prev with { Media = mergedMedia };
continue;
}
}
var msg = ParseSingleMessage(node, baseDirInZip, ref lastSenderName);
if (msg != null) messages.Add(msg);
if (msg != null)
{
messages.Add(msg);
lastCreatedAt = currentAt;
}
}
return messages;
@@ -82,11 +113,9 @@ public sealed class TelegramHtmlParser : ITelegramHtmlParser
if (idAttr.StartsWith("message")) long.TryParse(idAttr.Replace("message", ""), out numericId);
var fromNameNode = node.QuerySelector(".body > .from_name");
// Ignore forwarded from_name as message sender
if (fromNameNode != null && fromNameNode.Closest(".forwarded") != null) fromNameNode = null;
var senderName = fromNameNode != null ? CleanName(fromNameNode) : lastSenderName;
// If it's a new bubble group (no "joined"), update lastSenderName
if (senderName != null && !node.ClassList.Contains("joined")) lastSenderName = senderName;
var textNode = node.QuerySelector(".text");
@@ -94,30 +123,10 @@ public sealed class TelegramHtmlParser : ITelegramHtmlParser
var dateNode = node.QuerySelector(".date[title]") ?? node.QuerySelector("[title]");
var dateStr = dateNode?.GetAttribute("title") ?? "";
DateTime createdAt = DateTime.UtcNow;
if (!string.IsNullOrEmpty(dateStr))
{
var cleanDateStr = dateStr.Trim();
var formats = new[] { "dd.MM.yyyy HH:mm:ss 'UTC'zzz", "dd.MM.yyyy HH:mm:ss", "d.MM.yyyy HH:mm:ss", "dd.M.yyyy HH:mm:ss", "d.M.yyyy HH:mm:ss" };
try
{
if (DateTimeOffset.TryParseExact(cleanDateStr, formats, System.Globalization.CultureInfo.InvariantCulture, System.Globalization.DateTimeStyles.None, out var dto))
{
createdAt = dto.UtcDateTime;
}
else if (DateTimeOffset.TryParse(cleanDateStr.Replace("UTC", ""), System.Globalization.CultureInfo.InvariantCulture, System.Globalization.DateTimeStyles.None, out var dto2))
{
createdAt = dto2.UtcDateTime;
}
}
catch { }
}
DateTime createdAt = ParseDate(dateStr) ?? DateTime.UtcNow;
var mediaList = ExtractMedia(node, baseDir);
// Replies
string? replyToId = null;
var replyNode = node.QuerySelector(".reply_to a[href^=\"#go_to_message\"]");
if (replyNode != null)
@@ -126,12 +135,10 @@ public sealed class TelegramHtmlParser : ITelegramHtmlParser
replyToId = href?.Replace("#go_to_message", "message");
}
// Forwards
string? forwardedFrom = null;
var forwardNode = node.QuerySelector(".forwarded .from_name");
if (forwardNode != null) forwardedFrom = CleanName(forwardNode);
// Reactions
var reactions = new List<TelegramReaction>();
var reactionNodes = node.QuerySelectorAll(".reactions .reaction");
foreach (var r in reactionNodes)
@@ -153,11 +160,20 @@ public sealed class TelegramHtmlParser : ITelegramHtmlParser
}
}
private static DateTime? ParseDate(string? dateStr)
{
if (string.IsNullOrEmpty(dateStr)) return null;
var cleanDateStr = dateStr.Trim();
var formats = new[] { "dd.MM.yyyy HH:mm:ss 'UTC'zzz", "dd.MM.yyyy HH:mm:ss", "d.MM.yyyy HH:mm:ss", "dd.M.yyyy HH:mm:ss", "d.M.yyyy HH:mm:ss" };
if (DateTimeOffset.TryParseExact(cleanDateStr, formats, System.Globalization.CultureInfo.InvariantCulture, System.Globalization.DateTimeStyles.None, out var dto)) return dto.UtcDateTime;
if (DateTimeOffset.TryParse(cleanDateStr.Replace("UTC", ""), System.Globalization.CultureInfo.InvariantCulture, System.Globalization.DateTimeStyles.None, out var dto2)) return dto2.UtcDateTime;
return null;
}
private List<TelegramMedia> ExtractMedia(IElement node, string baseDir)
{
var mediaList = new List<TelegramMedia>();
// Look for links that represent media containers
var allLinks = node.QuerySelectorAll("a[href]").ToList();
foreach (var link in allLinks)
@@ -165,7 +181,6 @@ public sealed class TelegramHtmlParser : ITelegramHtmlParser
var href = NormalizeHref(link.GetAttribute("href"));
if (href == null || !href.Contains("/")) continue;
// Skip internal pages
if (href.EndsWith(".html", StringComparison.OrdinalIgnoreCase)) continue;
var fullPath = BuildPath(baseDir, href);
@@ -174,28 +189,15 @@ public sealed class TelegramHtmlParser : ITelegramHtmlParser
var fileName = Path.GetFileName(href);
var mimeType = GetMimeType(href);
// ANIMATION DETECTION:
// 1. Check classes (standard TG export)
bool hasGifClass = link.ClassList.Contains("animated_wrap") || link.ClassList.Contains("media_video");
// 2. Check full text of the link container for "Animation" string
// This is the most reliable way to find TG "GIFs" which are actually MP4s
bool hasAnimationText = link.TextContent.Contains("Animation", StringComparison.OrdinalIgnoreCase);
if (hasGifClass || hasAnimationText)
{
mimeType = "image/gif";
}
if (link.ClassList.Contains("media_voice_message") || href.Contains("voice_messages"))
{
mimeType = "audio/ogg";
}
if (hasGifClass || hasAnimationText) mimeType = "image/gif";
if (link.ClassList.Contains("media_voice_message") || href.Contains("voice_messages")) mimeType = "audio/ogg";
mediaList.Add(new TelegramMedia(fullPath, fileName, mimeType));
}
// Alternative check for voice messages in audio tags
foreach (var audioEl in node.QuerySelectorAll("audio[src]"))
{
var href = NormalizeHref(audioEl.GetAttribute("src"));

12
client-web/src/modules/chats/presentation/components/GroupSettings.tsx
View File

@@ -308,10 +308,10 @@ export default function GroupSettings({ chat, onClose, onGoToMessage }: GroupSet
const tabsConfig = [
{ key: 'members' as const, label: t('membersCount') || 'Участники', icon: Users, count: chat.members.length },
{ key: 'gifs' as const, label: t('gifs') || 'GIF', icon: Play, count: sortedGifs.length },
{ key: 'media' as const, label: t('mediaTab'), icon: ImageIcon, count: sortedMedia.length },
{ key: 'files' as const, label: t('filesTab'), icon: FileText, count: sortedFiles.flatMap(msg => msg.media || []).length },
{ key: 'links' as const, label: t('linksTab'), icon: LinkIcon, count: sortedLinks.flatMap(msg => msg.links || []).length },
{ key: 'gifs' as const, label: 'GIF', icon: Play, count: sortedGifs.length },
{ key: 'media' as const, label: t('mediaTab') || 'Медиа', icon: ImageIcon, count: sortedMedia.length },
{ key: 'files' as const, label: t('filesTab') || 'Файлы', icon: FileText, count: sortedFiles.flatMap(msg => msg.media || []).length },
{ key: 'links' as const, label: t('linksTab') || 'Ссылки', icon: LinkIcon, count: sortedLinks.flatMap(msg => msg.links || []).length },
];
const availableTabs = tabsConfig.filter(tab => tab.key === 'members' || !loadedTabs.has(tab.key) || tab.count > 0);
@@ -336,7 +336,7 @@ export default function GroupSettings({ chat, onClose, onGoToMessage }: GroupSet
animate={{ opacity: 1, x: 0, scale: 1 }}
exit={{ opacity: 0, x: 50, scale: 0.95 }}
transition={{ type: 'spring', damping: 25, stiffness: 300 }}
className="fixed right-3 top-3 bottom-3 w-[650px] max-w-[calc(100%-24px)] bg-surface-secondary/90 backdrop-blur-3xl shadow-2xl shadow-black/80 border border-white/5 rounded-[2rem] z-50 flex flex-col overflow-hidden ring-1 ring-white/10"
className="fixed right-3 top-3 bottom-3 w-[720px] max-w-[calc(100%-24px)] bg-surface-secondary/90 backdrop-blur-3xl shadow-2xl shadow-black/80 border border-white/5 rounded-[2rem] z-50 flex flex-col overflow-hidden ring-1 ring-white/10"
>
{/* Header */}
<div className="flex items-center justify-between p-4 border-b border-border/40">
@@ -496,7 +496,7 @@ export default function GroupSettings({ chat, onClose, onGoToMessage }: GroupSet
<button
key={tab.key}
onClick={() => setActiveTab(tab.key)}
className={`flex-1 flex flex-col items-center justify-center gap-1 py-2 text-[10px] font-bold uppercase tracking-widest transition-all ${
className={`flex-1 flex flex-col items-center justify-center gap-1.5 py-3 text-[10px] font-black uppercase tracking-wider transition-all border-r last:border-r-0 border-white/5 ${
activeTab === tab.key
? 'bg-white/5 text-knot-400'
: 'text-zinc-500 hover:text-zinc-300'