From 155c5712c2db2cbe56d7cf261a83192a546d7e55 Mon Sep 17 00:00:00 2001
From: Leo Vasanko <git@vasanko.com>
Date: Fri, 22 May 2026 16:44:08 +0000
Subject: [PATCH] Prefer ffmpeg-derived resolution and normalize to
 SD/HD/FHD/4K/8K

---
 mediahive/hivescan/indexer.py  |  4 +--
 mediahive/hivescan/parsing.py  |  3 +-
 mediahive/hivescan/showreel.py |  6 +++-
 mediahive/hivescan/utils.py    | 66 ++++++++++++++++++++++++++++++++--
 4 files changed, 72 insertions(+), 7 deletions(-)

diff --git a/mediahive/hivescan/indexer.py b/mediahive/hivescan/indexer.py
index 241c237..2963675 100644
--- a/mediahive/hivescan/indexer.py
+++ b/mediahive/hivescan/indexer.py
@@ -67,7 +67,7 @@ async def _build_torrent_info(
     return Torrent(
         title=item.title,
         playable_file=make_relative_path(playable_file, media_root),
-        resolution=item.resolution or (probe_info.resolution if probe_info else None),
+        resolution=(probe_info.resolution if probe_info else None) or item.resolution,
         quality=item.quality,
         codec=item.codec,
         audio=item.audio,
@@ -252,7 +252,7 @@ def _build_episodes_data(
             torrents[relpath] = Torrent(
                 title=f["torrent_title"],
                 playable_file=make_relative_path(f["path"], media_root),
-                resolution=f.get("resolution") or f.get("probed_resolution"),
+                resolution=f.get("probed_resolution") or f.get("resolution"),
                 quality=f.get("quality"),
                 codec=f.get("codec"),
                 audio=f.get("audio"),
diff --git a/mediahive/hivescan/parsing.py b/mediahive/hivescan/parsing.py
index 817f86a..952efba 100644
--- a/mediahive/hivescan/parsing.py
+++ b/mediahive/hivescan/parsing.py
@@ -8,6 +8,7 @@ import PTN
 from aiopathlib import AsyncPath
 
 from mediahive.hivescan.models import ContentHash, ContentType, ParsedContent
+from mediahive.hivescan.utils import normalize_resolution_label
 
 
 def determine_content_type(parsed: dict) -> ContentType:
@@ -36,7 +37,7 @@ async def parse_download(path: Path) -> ParsedContent:
         content_type=content_type,
         title=parsed.get("title", name),
         year=parsed.get("year"),
-        resolution=parsed.get("resolution"),
+        resolution=normalize_resolution_label(parsed.get("resolution")),
         quality=parsed.get("quality"),
         codec=parsed.get("codec"),
         audio=parsed.get("audio"),
diff --git a/mediahive/hivescan/showreel.py b/mediahive/hivescan/showreel.py
index f4ddae4..711978b 100644
--- a/mediahive/hivescan/showreel.py
+++ b/mediahive/hivescan/showreel.py
@@ -19,6 +19,8 @@ from typing import Optional
 
 from aiopathlib import AsyncPath
 
+from mediahive.hivescan.utils import classify_resolution_from_dimensions
+
 logger = logging.getLogger("hivescan.showreel")
 
 
@@ -388,7 +390,9 @@ async def probe_media_info(video_path: str) -> MediaProbeInfo:
             if dim_match:
                 info.width = int(dim_match.group(1))
                 info.height = int(dim_match.group(2))
-                info.resolution = f"{info.width}x{info.height}"
+                info.resolution = classify_resolution_from_dimensions(
+                    info.width, info.height
+                )
 
         info.is_hdr = (
             "smpte2084" in lower_text
diff --git a/mediahive/hivescan/utils.py b/mediahive/hivescan/utils.py
index a7cf6a2..397a4ce 100644
--- a/mediahive/hivescan/utils.py
+++ b/mediahive/hivescan/utils.py
@@ -2,11 +2,10 @@
 
 import time
 from pathlib import Path
-from typing import Optional, List
+from typing import List, Optional
 
 from aiopathlib import AsyncPath
 
-
 # Default output folder name (created at common root of scanned paths)
 DEFAULT_OUTPUT_FOLDER = ".mediahive"
 
@@ -15,15 +14,76 @@ _ATIME_FRESHNESS_THRESHOLD = 3600
 
 # Resolution priority for quality sorting (higher = better)
 RESOLUTION_PRIORITY = {
-    "2160p": 4,
+    "8K": 5,
     "4K": 4,
+    "FHD": 3,
+    "HD": 2,
+    "SD": 1,
+    # Backward compatibility for existing snapshot data
+    "4320p": 5,
+    "2160p": 4,
+    "UHD": 4,
     "1080p": 3,
     "1080i": 3,
     "720p": 2,
+    "576p": 1,
     "480p": 1,
 }
 
 
+def classify_resolution_from_dimensions(
+    width: int | None, height: int | None
+) -> str | None:
+    """Map raw frame dimensions to SD/HD/FHD/4K/8K buckets.
+
+    Uses the smallest standard frame bucket that can contain the source frame,
+    which keeps cropped cinematic encodes in their expected class.
+    """
+    if not width or not height or width <= 0 or height <= 0:
+        return None
+
+    long_edge = max(width, height)
+    short_edge = min(width, height)
+
+    buckets = [
+        (1024, 576, "SD"),
+        (1280, 720, "HD"),
+        (1920, 1080, "FHD"),
+        (4096, 2160, "4K"),
+        (8192, 4320, "8K"),
+    ]
+
+    for max_w, max_h, label in buckets:
+        if long_edge <= max_w and short_edge <= max_h:
+            return label
+
+    return "8K"
+
+
+def normalize_resolution_label(value: str | None) -> str | None:
+    """Normalize PTN/legacy resolution text into SD/HD/FHD/4K/8K labels."""
+    if not value:
+        return None
+
+    normalized = str(value).strip().upper()
+    mapping = {
+        "SD": "SD",
+        "HD": "HD",
+        "FHD": "FHD",
+        "4K": "4K",
+        "8K": "8K",
+        "4320P": "8K",
+        "2160P": "4K",
+        "UHD": "4K",
+        "1080P": "FHD",
+        "1080I": "FHD",
+        "720P": "HD",
+        "576P": "SD",
+        "480P": "SD",
+    }
+    return mapping.get(normalized)
+
+
 async def get_added_timestamp(path: Path) -> Optional[int]:
     """
     Get the timestamp when a torrent was added to the collection.