feat(x): парсинг X через api.fxtwitter.com — 30 живых хендлов как источники, адаптер x_user, полный текст твитов, метрики; старый syndication-путь (429) убран

omni/feeds.py

@@ -648,7 +648,15 @@ def _models_dev(src, limit):
     return out
 
 
+def _x_user(src, limit=12):
+    """Лента X-аккаунта через api.fxtwitter.com (без ключей)."""
+    from .providers import x as xp
+    handle = (src.extra or {}).get("handle") or src.url.rsplit("/", 1)[-1]
+    return xp.profile(handle, limit=max(4, min(limit, 25)))
+
+
 ADAPTERS = {
+    "x_user": _x_user,
     "rss": _rss,
     "generic_json": _generic_json,
     "or_models": _or_models,

omni/providers/x.py

@@ -77,30 +77,64 @@ def search_x(query, limit=15, handles=None):
                       extra={"handle": d["handle"]})
             items.append(it)
 
-    # путь 3: конкретные профили
-    for h in (handles or [])[:4]:
-        r = net.get(f"https://syndication.twitter.com/srv/timeline-profile/screen-name/{h}",
-                    headers=UA)
-        if r.status != 200:
-            continue
-        m = re.search(r'<script id="__NEXT_DATA__" type="application/json">(.*?)</script>', r.text, re.S)
-        if not m:
-            continue
-        try:
-            data = json.loads(m.group(1))
-        except Exception:
-            continue
-        entries = (((data.get("props") or {}).get("pageProps") or {}).get("timeline") or {}).get("entries") or []
-        for e in entries[:12]:
-            tw = ((e.get("content") or {}).get("tweet") or {})
-            txt = tw.get("full_text") or tw.get("text") or ""
-            if not txt:
-                continue
-            tid = tw.get("id_str") or ""
-            it = Item("x", "tweet", url=f"https://x.com/{h}/status/{tid}" if tid else f"https://x.com/{h}",
-                      title=txt[:120], text=txt, author=h, source="x.com/" + h,
-                      published=tw.get("created_at", ""),
-                      extra={"likes": tw.get("favorite_count"), "rt": tw.get("retweet_count")})
-            it.score = float(tw.get("favorite_count") or 0) ** 0.5
+    # путь 3: конкретные профили через fxtwitter (без ключей, отдаёт полный текст)
+    for h in (handles or [])[:6]:
+        for it in profile(h, limit=12):
             items.append(it)
+
+    # дедуп по ссылке
+    uniq, seen_u = [], set()
+    for it in items:
+        if it.url in seen_u:
+            continue
+        seen_u.add(it.url)
+        uniq.append(it)
+    return uniq[:max(limit, 12) if handles else len(uniq)]
+
+
+def profile(handle, limit=20, cursor=None):
+    """Лента профиля X через api.fxtwitter.com — без токенов и логина.
+
+    Возвращает список Item; при ошибке сети/404 — пустой список.
+    """
+    h = (handle or "").lstrip("@").strip()
+    if not h:
+        return []
+    url = f"https://api.fxtwitter.com/2/profile/{h}/statuses"
+    if cursor:
+        url += "?cursor=" + urllib.parse.quote(str(cursor))
+    try:
+        r = net.get(url, timeout=25, json_mode=True)
+    except Exception:
+        return []
+    if r.status != 200:
+        return []
+    data = r.json() or {}
+    out = []
+    for t in (data.get("results") or []):
+        if t.get("type") != "status":
+            continue
+        rt = t.get("raw_text")
+        txt = (rt.get("text") if isinstance(rt, dict) else rt) or t.get("text") or ""
+        if not isinstance(txt, str):
+            txt = str(txt)
+        if len(txt) < 15:
+            continue
+        tid = t.get("id") or ""
+        author = (t.get("author") or {}).get("screen_name") or h             if isinstance(t.get("author"), dict) else h
+        likes = t.get("likes") or 0
+        it = Item("x", "tweet",
+                  url=t.get("url") or f"https://x.com/{h}/status/{tid}",
+                  title=txt.strip().splitlines()[0][:140], text=txt,
+                  author=author, source="x.com/" + h,
+                  published=t.get("created_at", ""),
+                  extra={"handle": h, "lang": t.get("lang", ""), "likes": likes, "replies": t.get("replies"),
+                         "reposts": t.get("reposts"), "views": t.get("views"),
+                         "quote": (t.get("quote") or {}).get("url") if isinstance(t.get("quote"), dict) else None})
+        it.score = float(likes) ** 0.5 + float(t.get("replies") or 0) ** 0.4
+        out.append(it)
+        if len(out) >= limit:
+            break
+    return out
+
     return items[:limit * 2]

omni/sources.py

@@ -16,6 +16,7 @@ from .sources_ext3 import EXTRA3
 from .sources_abuse import ABUSE_LIVE
 from .sources_ext4 import EXTRA5
 from .sources_ext5 import EXTRA6
+from .sources_x import X_SOURCES
 
 
 @dataclass
@@ -199,7 +200,7 @@ def _dedup(rows):
     return out
 
 
-SOURCES = _dedup(SOURCES_BASE + [Source(**d) for d in EXTRA] + [Source(**d) for d in EXTRA2] + [Source(**d) for d in EXTRA3] + [Source(**d) for d in ABUSE_LIVE] + [Source(**d) for d in EXTRA4] + [Source(**d) for d in EXTRA5] + [Source(**d) for d in EXTRA6])
+SOURCES = _dedup(SOURCES_BASE + [Source(**d) for d in EXTRA] + [Source(**d) for d in EXTRA2] + [Source(**d) for d in EXTRA3] + [Source(**d) for d in ABUSE_LIVE] + [Source(**d) for d in EXTRA4] + [Source(**d) for d in EXTRA5] + [Source(**d) for d in EXTRA6] + [Source(**d) for d in X_SOURCES])
 
 BY_NAME = {s.name: s for s in SOURCES}
 CATEGORIES = sorted({s.category for s in SOURCES})

omni/sources_x.py

@@ -0,0 +1,53 @@
+"""X/Twitter источники — читаются без API-ключей через api.fxtwitter.com.
+
+Адаптер: feeds._x_user -> providers.x.profile(handle).
+Каждый хендл проверен живым прогоном (>=3 записи). Мёртвые (0 записей) не добавлены.
+"""
+
+_X = [
+    # (handle, category, теги)
+    ("vxunderground", "abuse", ("malware", "leak")),
+    ("malwrhunterteam", "abuse", ("malware", "stealer")),
+    ("DailyDarkWeb", "abuse", ("leak", "ransomware")),
+    ("DarkWebInformer", "abuse", ("leak", "darkweb")),
+    ("abuse_ch", "abuse", ("stealer", "malware")),
+    ("H4ckManac", "abuse", ("leak", "breach")),
+    ("falconfeedsio", "abuse", ("ransomware", "leak")),
+    ("nullcookies", "abuse", ("leak", "fraud")),
+    ("TheHackersNews", "security", ("news", "cve")),
+    ("BleepinComputer", "security", ("news", "cve")),
+    ("campuscodi", "security", ("news", "cve")),
+    ("ShadowServer", "security", ("scan", "cve")),
+    ("RecordedFuture", "security", ("intel", "cve")),
+    ("CISAgov", "security", ("cve", "advisory")),
+    ("msftsecintel", "security", ("cve", "advisory")),
+    ("Unit42_Intel", "security", ("intel", "cve")),
+    ("Mandiant", "security", ("intel", "cve")),
+    ("CrowdStrike", "security", ("intel", "cve")),
+    ("SentinelOne", "security", ("intel", "cve")),
+    ("Kaspersky", "security", ("intel", "cve")),
+    ("elasticlabs", "security", ("intel",)),
+    ("blackorbird", "security", ("apt", "intel")),
+    ("OpenAI", "models", ("model", "release")),
+    ("AnthropicAI", "models", ("model", "release")),
+    ("deepseek_ai", "models", ("model", "release")),
+    ("huggingface", "models", ("model", "dataset")),
+    ("AI_News_", "models", ("news", "model")),
+    ("karpathy", "community", ("ai", "guide")),
+    ("simonw", "community", ("ai", "tool")),
+    ("ylecun", "community", ("ai", "news")),
+]
+
+X_SOURCES = [
+    {
+        "name": "x_" + h.lower(),
+        "category": cat,
+        "kind": "tweet",
+        "url": "https://x.com/" + h,
+        "adapter": "x_user",
+        "tags": tags,
+        "limit": 15,
+        "extra": {"handle": h},
+    }
+    for h, cat, tags in _X
+]