video search: REAL slskd search for the Soulseek source (was mocked)

The Soulseek ⌕ now actually queries your slskd instance instead of fabricating
results. core/video/slskd_search.py (isolated): build_query(scope,…) → POST
/api/v0/searches (shared soulseek.* URL+key) → poll /responses (bounded ~8s,
early-exit at 25) → keep video files → group_video_files() folds them into one hit
per release folder with a peer count + the fastest available source. Same
{title,size_bytes,…} shape, so parse→evaluate→rank/cards are unchanged.

Endpoint routes source=='soulseek' to slskd (torrent/usenet stay mocked), surfaces
'slskd not configured' / network errors, carries username/peers/slots/filename
through for the cards + a future real Grab, and caps to 40. Pure helpers tested (4);
isolation guard + ruff clean.

api/video/downloads.py

@@ -108,6 +108,7 @@ def register_routes(bp):
         body = request.get_json(silent=True) or {}
         scope = str(body.get("scope") or "movie").lower()
         title = body.get("title") or ""
+        source = str(body.get("source") or "").lower()
 
         def _int(v):
             try:
@@ -117,9 +118,23 @@ def register_routes(bp):
 
         want_season, want_episode = _int(body.get("season")), _int(body.get("episode"))
         profile = load_profile(get_video_db())
-        raw = mock_search(scope, title, year=body.get("year"), season=want_season,
+
-                          episode=want_episode, season_end=_int(body.get("season_end")),
+        live = False
-                          source=body.get("source"))
+        if source == "soulseek":
+            # REAL Soulseek search (slskd). Torrent/Usenet stay mocked until those
+            # indexers are wired. Same {title, size_bytes, …} shape downstream.
+            from core.video.slskd_search import build_query, slskd_search
+            sres = slskd_search(build_query(scope, title, year=body.get("year"),
+                                            season=want_season, episode=want_episode))
+            if not sres.get("configured"):
+                return jsonify({"scope": scope, "results": [], "error": "slskd isn't configured — set its URL on Settings → Downloads."})
+            if sres.get("error"):
+                return jsonify({"scope": scope, "results": [], "error": "slskd: " + str(sres["error"])})
+            raw, live = sres["hits"], True
+        else:
+            raw = mock_search(scope, title, year=body.get("year"), season=want_season,
+                              episode=want_episode, season_end=_int(body.get("season_end")),
+                              source=source)
 
         results = []
         for hit in raw:
@@ -127,8 +142,12 @@ def register_routes(bp):
             size_gb = round((hit.get("size_bytes") or 0) / (1024 ** 3), 1)
             verdict = evaluate_release(parsed, profile, scope=scope, want_season=want_season,
                                        want_episode=want_episode, size_gb=size_gb)
+            avail = hit.get("seeders") if hit.get("seeders") is not None else (hit.get("peers") or 0)
             results.append({
-                "title": hit.get("title"), "size_gb": size_gb, "seeders": hit.get("seeders", 0),
+                "title": hit.get("title"), "size_gb": size_gb,
+                "seeders": hit.get("seeders"), "peers": hit.get("peers"),
+                "username": hit.get("username"), "slots": hit.get("slots"),
+                "filename": hit.get("filename"), "_avail": avail,
                 "quality_label": verdict["quality_label"], "accepted": verdict["accepted"],
                 "rejected": verdict["rejected"], "score": verdict["score"],
                 "resolution": parsed.get("resolution"), "source": parsed.get("source"),
@@ -136,9 +155,11 @@ def register_routes(bp):
                 "audio": parsed.get("audio"), "group": parsed.get("group"),
                 "repack": parsed.get("repack") or parsed.get("proper"),
             })
-        # accepted first, then best score, then most seeders.
+        # accepted first, then best score, then most availability (seeders/peers).
-        results.sort(key=lambda r: (r["accepted"], r["score"], r["seeders"]), reverse=True)
+        results.sort(key=lambda r: (r["accepted"], r["score"], r["_avail"]), reverse=True)
-        return jsonify({"scope": scope, "results": results})
+        for r in results:
+            r.pop("_avail", None)
+        return jsonify({"scope": scope, "results": results[:40], "live": live})
 
     @bp.route("/downloads/youtube-quality", methods=["GET"])
     def video_youtube_quality():

core/video/slskd_search.py

@@ -0,0 +1,146 @@
+"""Real Soulseek (slskd) search for the video Download view.
+
+Replaces the mock for the Soulseek source: it POSTs a search to the slskd instance
+(the shared ``soulseek.*`` config used by the music side too), polls for responses,
+keeps the video files, and GROUPS them by release folder so each card is one release
+(with a peer count) rather than one row per user. The parse → evaluate → rank pipeline
+downstream is unchanged — this just returns the same ``{title, size_bytes, …}`` shape.
+
+Isolated: imports only stdlib + requests + the shared ``config_manager`` (app config,
+not music code). The pure helpers (build_query / group_video_files) are unit-tested;
+the HTTP poll is thin I/O glue.
+"""
+
+from __future__ import annotations
+
+import time
+from typing import Any
+
+import requests
+
+# Container extensions we treat as the actual video (everything else — subs, nfo,
+# art, samples — is ignored for quality purposes).
+VIDEO_EXTS = frozenset((
+    "mkv", "mp4", "avi", "m4v", "mov", "wmv", "ts", "m2ts", "mpg", "mpeg",
+    "webm", "flv", "vob", "divx", "mk3d",
+))
+
+
+def build_query(scope: str, title: Any, *, year: Any = None, season: Any = None,
+                episode: Any = None) -> str:
+    """The text we hand slskd for a given scope (movie / episode / season / series)."""
+    t = str(title or "").strip()
+    scope = (scope or "movie").lower()
+    try:
+        s_i = int(season) if season is not None else None
+        e_i = int(episode) if episode is not None else None
+    except (TypeError, ValueError):
+        s_i = e_i = None
+    if scope == "episode" and s_i is not None and e_i is not None:
+        return "%s S%02dE%02d" % (t, s_i, e_i)
+    if scope == "season" and s_i is not None:
+        return "%s S%02d" % (t, s_i)
+    if scope == "movie" and year:
+        return ("%s %s" % (t, year)).strip()
+    return t   # series / fallback
+
+
+def _is_video(filename: str) -> bool:
+    fn = str(filename or "").replace("\\", "/")
+    base = fn.rsplit("/", 1)[-1]
+    if "sample" in base.lower():
+        return False
+    ext = base.rsplit(".", 1)[-1].lower() if "." in base else ""
+    return ext in VIDEO_EXTS
+
+
+def _release_name(filename: str) -> str:
+    """The release a file belongs to — its parent folder (where scene/p2p put the
+    quality), falling back to the bare filename (sans extension)."""
+    fn = str(filename or "").replace("\\", "/").strip("/")
+    parts = [p for p in fn.split("/") if p]
+    if len(parts) >= 2:
+        return parts[-2]
+    base = parts[-1] if parts else ""
+    return base.rsplit(".", 1)[0] if "." in base else base
+
+
+def group_video_files(responses: Any) -> list:
+    """Flatten slskd responses → one hit per release folder, with a peer count and the
+    best (fastest, most-available) source. Pure — drives the unit tests."""
+    groups: dict = {}
+    for resp in (responses if isinstance(responses, list) else []):
+        if not isinstance(resp, dict):
+            continue
+        user = resp.get("username")
+        speed = resp.get("uploadSpeed", 0) or 0
+        slots = resp.get("freeUploadSlots", 0) or 0
+        for f in (resp.get("files") or []):
+            fn = f.get("filename", "")
+            if not _is_video(fn):
+                continue
+            rel = _release_name(fn)
+            g = groups.get(rel)
+            if g is None:
+                g = groups[rel] = {"title": rel, "size_bytes": 0, "users": set(),
+                                   "best_speed": -1, "username": None, "slots": 0, "filename": fn}
+            g["size_bytes"] = max(g["size_bytes"], f.get("size", 0) or 0)
+            if user:
+                g["users"].add(user)
+            if speed > g["best_speed"]:
+                g["best_speed"] = speed
+                g["username"] = user
+                g["slots"] = slots
+                g["filename"] = fn
+    out = []
+    for g in groups.values():
+        out.append({"title": g["title"], "size_bytes": g["size_bytes"], "peers": len(g["users"]),
+                    "username": g["username"], "slots": g["slots"], "filename": g["filename"]})
+    out.sort(key=lambda h: (h["peers"], h["size_bytes"]), reverse=True)
+    return out
+
+
+def slskd_search(query: str, *, max_seconds: int = 8, slskd_timeout_ms: int = 4500) -> dict:
+    """POST a search to slskd, poll responses, return {configured, hits[, error]}.
+    Thin I/O glue around ``group_video_files``."""
+    from config.settings import config_manager
+    base = str(config_manager.get("soulseek.slskd_url", "") or "").rstrip("/")
+    if not base:
+        return {"configured": False, "hits": []}
+    key = config_manager.get("soulseek.api_key", "") or ""
+    headers = {"X-API-Key": key} if key else {}
+    try:
+        min_mbps = int(config_manager.get("soulseek.min_peer_upload_speed", 0) or 0)
+    except (TypeError, ValueError):
+        min_mbps = 0
+    payload = {"searchText": query, "timeout": slskd_timeout_ms, "filterResponses": True,
+               "minimumResponseFileCount": 1, "minimumPeerUploadSpeed": min_mbps * 125000}
+    try:
+        r = requests.post(base + "/api/v0/searches", json=payload, headers=headers, timeout=10)
+        r.raise_for_status()
+        data = r.json()
+    except Exception as e:   # noqa: BLE001 - surface any slskd/network failure to the UI
+        return {"configured": True, "hits": [], "error": str(e)}
+    sid = data.get("id") if isinstance(data, dict) else (
+        data[0].get("id") if isinstance(data, list) and data and isinstance(data[0], dict) else None)
+    if not sid:
+        return {"configured": True, "hits": [], "error": "slskd returned no search id"}
+
+    responses: list = []
+    deadline = time.monotonic() + max_seconds
+    while time.monotonic() < deadline:
+        try:
+            rr = requests.get(base + "/api/v0/searches/%s/responses" % sid, headers=headers, timeout=10)
+            if rr.ok:
+                body = rr.json()
+                if isinstance(body, list):
+                    responses = body
+        except Exception:   # noqa: BLE001, S110 - keep polling through transient errors
+            pass
+        if len(responses) >= 25:
+            break
+        time.sleep(1)
+    return {"configured": True, "hits": group_video_files(responses)}
+
+
+__all__ = ["VIDEO_EXTS", "build_query", "group_video_files", "slskd_search"]

tests/test_video_slskd_search.py

@@ -0,0 +1,51 @@
+"""Real slskd search — the pure, testable seams (query building + grouping slskd
+responses into per-release hits). The HTTP poll itself is thin I/O glue, not tested
+here. Isolated from music."""
+
+from __future__ import annotations
+
+from core.video.slskd_search import build_query, group_video_files
+
+
+def test_build_query_per_scope():
+    assert build_query("movie", "The Matrix", year=1999) == "The Matrix 1999"
+    assert build_query("movie", "The Matrix") == "The Matrix"
+    assert build_query("episode", "The Wire", season=2, episode=3) == "The Wire S02E03"
+    assert build_query("season", "The Wire", season=2) == "The Wire S02"
+    assert build_query("series", "The Wire") == "The Wire"
+
+
+def test_group_video_files_groups_by_release_folder():
+    responses = [
+        {"username": "alice", "uploadSpeed": 900, "freeUploadSlots": 1, "files": [
+            {"filename": r"@@x\The.Wire.S02.1080p.BluRay.x265-GRP\the.wire.s02e01.mkv", "size": 3_000_000_000},
+            {"filename": r"@@x\The.Wire.S02.1080p.BluRay.x265-GRP\the.wire.s02e02.mkv", "size": 3_200_000_000},
+            {"filename": r"@@x\The.Wire.S02.1080p.BluRay.x265-GRP\readme.nfo", "size": 1024},
+        ]},
+        {"username": "bob", "uploadSpeed": 1500, "freeUploadSlots": 2, "files": [
+            {"filename": "The.Wire.S02.1080p.BluRay.x265-GRP/the.wire.s02e01.mkv", "size": 3_000_000_000},
+        ]},
+    ]
+    hits = group_video_files(responses)
+    assert len(hits) == 1                      # both users → one release
+    h = hits[0]
+    assert h["title"] == "The.Wire.S02.1080p.BluRay.x265-GRP"
+    assert h["peers"] == 2                      # alice + bob
+    assert h["username"] == "bob"               # bob is faster → the chosen source
+    assert h["slots"] == 2
+    assert h["size_bytes"] == 3_200_000_000     # largest video file in the folder
+
+
+def test_group_skips_non_video_and_samples():
+    responses = [{"username": "u", "uploadSpeed": 1, "freeUploadSlots": 0, "files": [
+        {"filename": "Movie.2020.1080p/sample.mkv", "size": 50_000_000},   # sample dropped
+        {"filename": "Movie.2020.1080p/Movie.2020.1080p.srt", "size": 40000},  # subs dropped
+        {"filename": "Movie.2020.1080p/movie.mp4", "size": 8_000_000_000},
+    ]}]
+    hits = group_video_files(responses)
+    assert len(hits) == 1 and hits[0]["size_bytes"] == 8_000_000_000
+
+
+def test_group_handles_garbage():
+    assert group_video_files(None) == []
+    assert group_video_files([{"nope": 1}, "junk"]) == []