chore(release): v0.6.0

Reviewed-on: #12

.gitea/workflows/build-and-publish.yaml

@@ -45,7 +45,8 @@ jobs:
 
       # The suite runs inside the image, against the interpreter that ships,
       # rather than against whatever the runner happens to provide. A failing
-      # test fails the build. Layers are shared with the push build below.
+      # test fails the build. The runtime stage below is built from the same
+      # daemon afterwards, so its layers are already in cache.
       - name: Run the test suite inside the image
         run: docker build --target test -t music-curator:test .
 
@@ -124,9 +125,6 @@ jobs:
           echo "release=${release}" >> "$GITHUB_OUTPUT"
           echo "Computed bump=${bump}, release=${release}, base=${base}"
 
-      - name: Set up Buildx
-        uses: docker/setup-buildx-action@d7f5e7f509e45cec5c76c4d5afdd7de93d0b3df5 # v4
-
       - name: Log in to the Gitea container registry
         if: github.event_name != 'pull_request'
         uses: docker/login-action@650006c6eb7dba73a995cc03b0b2d7f5ca915bee # v4
@@ -135,21 +133,33 @@ jobs:
           username: ${{ github.repository_owner }}
           password: ${{ secrets.PACKAGES_TOKEN }}
 
-      - name: Build and push
+      # Plain `docker build` rather than buildx. buildx boots its own buildkit
-        uses: docker/build-push-action@f9f3042f7e2789586610d6e8b85c8f03e5195baf # v7
+      # in a container with a cache of its own, so it shared nothing with the
-        with:
+      # test build above and rebuilt the image from the base image up -- two
-          context: .
+      # full builds per run. It earns that cost when building for several
-          # Without this the last stage in the Dockerfile -- the test stage --
+      # platforms; this only ever targets the amd64 NAS, so it does not.
-          # would be what gets published.
+      #
-          target: runtime
+      # `--target runtime` is a strict prefix of the test stage, so every layer
-          # The NAS is the only host this runs on. Building arm64 as well would
+      # is already in the daemon's cache and this resolves in seconds.
-          # mean emulating it under QEMU for no consumer.
+      - name: Build the runtime image
-          platforms: linux/amd64
+        run: |
-          push: ${{ github.event_name != 'pull_request' }}
+          set -euo pipefail
-          tags: ${{ steps.version.outputs.tags }}
+          tags=()
-          labels: |
+          while IFS= read -r tag; do
-            org.opencontainers.image.source=${{ github.server_url }}/${{ github.repository }}
+            [ -n "$tag" ] && tags+=(-t "$tag")
-            org.opencontainers.image.revision=${{ github.sha }}
+          done <<< "${{ steps.version.outputs.tags }}"
+          docker build --target runtime \
+            --label "org.opencontainers.image.source=${GITHUB_SERVER_URL}/${GITHUB_REPOSITORY}" \
+            --label "org.opencontainers.image.revision=${GITHUB_SHA}" \
+            "${tags[@]}" .
+
+      - name: Push
+        if: github.event_name != 'pull_request'
+        run: |
+          set -euo pipefail
+          while IFS= read -r tag; do
+            [ -n "$tag" ] && docker push "$tag"
+          done <<< "${{ steps.version.outputs.tags }}"
 
       # Record the release: write the computed version into pyproject.toml, then
       # commit and tag it, so the packaging metadata always matches the release

README.md

@@ -7,9 +7,11 @@ which keeps an MP3 copy of a lossless library for an iPod. This one answers the
 question that mirror cannot: which of it is worth carrying, and which of it has
 not been played in years.
 
-**This is stage two.** It ingests the scrobble history, indexes the library from
+**This is stage three.** It ingests the scrobble history, indexes the library
-Lidarr, and matches one to the other. There are no playlists yet, and it writes
+from Lidarr, matches one to the other, and writes playlists into the mirror —
-nothing back — every Lidarr call is a `GET`. See "Where this is going" below.
+by listening history and by mood.
+Nothing is written back to Lidarr — every call there is a `GET`. See "Where this
+is going" below.
 
 ## What it does today
 
@@ -18,6 +20,7 @@ nothing back — every Lidarr call is a `GET`. See "Where this is going" below.
 - Indexes every artist, album and track Lidarr knows about, with file paths and
   the date each file landed.
 - Ties the two together and reports how well it managed.
+- Writes M3U playlists into the mirror, from the listening history.
 
 ## Matching
 
@@ -29,7 +32,14 @@ Two tiers, and no third.
 | `name` | Normalised artist and title | Everything the first tier could not carry |
 | `none` | — | Recorded as a miss, never guessed at |
 
-The normalisation is the load-bearing part, because the two sides disagree in
+The name tier does almost all of the work. A recording MBID is exact when it
+lands, but MusicBrainz holds a separate recording per release, and Last.fm and
+Lidarr rarely pick the same one: on a real library, two thirds of scrobbles
+carry a recording id and barely a twentieth of them join on it. The report
+counts how many carried an id and matched on name anyway, which is the measure
+of that disagreement.
+
+So the normalisation is the load-bearing part, because the two sides disagree in
 predictable ways. It folds case and accents, drops guest credits (`Yellowcard
 feat. Tay Jardine` against a tag of `Yellowcard`), strips a trailing
 version suffix (`(Remastered 2011)`, `- Live`), expands `&`, and removes a
@@ -42,6 +52,56 @@ It leans towards collapsing too much. A false match makes something look
 played; a missed match makes something look abandoned. Only one of those
 deletes music.
 
+### Indexing quirks
+
+Albums are fetched from the **unfiltered** `GET /api/v1/album` first: one
+request, and the only path that skips albums whose artist metadata is missing
+rather than dereferencing it.
+
+That is not enough on its own. Every album endpoint maps through a resource
+that picks the release with `SingleOrDefault(x => x.Monitored)`, which throws
+for an album with **two monitored releases** and takes the whole response with
+it:
+
+```
+HTTP 500: Sequence contains more than one element
+```
+
+When the bulk call dies that way, the indexer falls back to one request per
+artist. It cannot avoid the exception, but it confines it to whichever artist
+owns the offending album and names them in the log — which is the only
+practical way to find it in a large library. Open that artist in Lidarr and
+check the Releases tab of each album: exactly one release may be monitored.
+
+Losing an artist's albums does not cost their tracks, which come from a
+different endpoint with a different mapper, so matching is unaffected. A cull
+would not be, and the report says so.
+
+### Talking to Lidarr
+
+Indexing is two requests per artist, and more when the album fallback fires. On
+a large library that is thousands of requests in a few minutes. `urllib` opens a
+new TCP connection and performs a new DNS lookup for every one of them, which is
+enough to exhaust a container's resolver and produce `[Errno -3] Try again` on
+everything at once. The client therefore holds one connection open per host and
+resolves once.
+
+Transient failures — a dropped connection, a resolver hiccup, `429`, `502`,
+`503`, `504` — are retried with a backoff. An HTTP `500` is not: it is an
+unhandled exception inside Lidarr's own serialisation and will be raised again
+identically. That distinction also decides whether a failure is worth
+investigating; a library-wide outage is not probed artist by artist, because
+doing so multiplies the load that caused it.
+
+A local address is preferable to a public hostname here. It removes DNS, the
+reverse proxy and its timeouts from a path that needs none of them.
+
+Tracks and files have no unfiltered endpoint — Lidarr rejects a call with no
+filter — so they stay per artist. If one artist cannot be served, that artist is
+skipped and the run continues, but the count is recorded and the coverage report
+says so loudly. A missing artist makes their played music look cold, so an
+incomplete index must never be culled against.
+
 ### Reading the coverage report
 
 Matched against unmatched is the wrong comparison — most unmatched listening is
@@ -52,10 +112,169 @@ matcher. The line to watch is:
 unmatched by an artist the library holds: N pairs, M plays
 ```
 
-That is a track that was played, sitting beside a file it should have matched.
+That is a track that was played by an artist the library holds. It is then
-Those are the matcher's real misses, and every one is a candidate for being
+split three ways, because owning an artist is a weak proxy for owning a track
-wrongly called cold in stage four. The report lists the worst fifteen by play
+and a shared title is a weak proxy for a shared song:
-count so they can be eyeballed.
+
+- **the library's own title credits the scrobbled artist** — `Voodoo People
+  (Pendulum Remix)` against a play credited to Pendulum. Same song, filed
+  under the original artist. These are the genuine misses.
+- **the same title under an unrelated artist** — a collision, not a miss.
+  Across fifty thousand tracks these are constant: `Everyday` is Rusko and
+  also Def Leppard, `Kaleidoscope` is Delta Heavy and also Chappell Roan.
+  Matching on title alone would be far worse than missing them, which is why
+  there is no such tier.
+- **the title is nowhere in the library** — never bought.
+
+Only the first is worth chasing. Counting all three as matcher failures
+overstates the problem and would over-block the cull.
+
+## Playlists
+
+Written into `<mirror>/_playlists/` as extended M3U, rebuilt every pass. Six
+rules, capped at `--playlist-limit` tracks each:
+
+| Playlist | Rule |
+| -------------------- | --------------------------------------------------------- |
+| `heavy-rotation` | Most played over the last twelve months |
+| `all-time` | Most played ever |
+| `neglected` | Played heavily once, silent for twelve months |
+| `deep-cuts` | Never played, from albums whose other tracks you play constantly |
+| `unheard-favourites` | Never played, by the artists you play most |
+| `unheard` | Never played, anywhere in the library |
+
+Ninety days was the obvious window for "recent" and is the wrong one: on a real
+history it holds a few hundred plays spread thinly across a twenty-thousand
+track rotation, so nothing ranks meaningfully. Twelve months does.
+
+The two `unheard` playlists rotate **weekly**, not per pass. A pass runs every
+few hours, and a playlist that reorders itself each time is one that has to be
+re-imported each time — the Music app imports a snapshot of a file, it does not
+track it.
+
+### Moods
+
+A second set of playlists selects by **Last.fm's crowd tags** rather than by
+listening history. MusicBrainz genres arrive free with the Lidarr index and are
+no use for this: they are sparse and formal, and will not tell you a record is
+screamo or synthwave. People typing tags will.
+
+Tags are fetched once per artist — one `artist.getTopTags` call each — and
+refreshed every ninety days. `--tag-limit` spreads the first sweep over several
+passes.
+
+Artists are looked up **by name**, not by MusicBrainz id, despite Lidarr having
+an id for every one of them. Last.fm's mbid index is stale and partial: it
+answers "the artist you supplied could not be found" for Devo, Escape the Fate,
+Blasterjaxx and a few hundred others whose pages plainly exist and carry exactly
+the tags wanted. Its name index is the one its own site runs on. The id is kept
+only as a fallback, for a name Lidarr spells differently.
+
+An artist neither key resolves is recorded as fetched with no tags, so the next
+pass does not spend a request on it again. A genuine failure — a rate limit, a
+bad key — is *not* recorded, so that one is retried.
+
+The built-in moods are built from the **tag distribution of this library**,
+measured, rather than from a general taxonomy:
+
+| Mood | Selected on |
+| --------------- | -------------------------------------------------------- |
+| `drum-and-bass` | drum and bass and its six spellings, liquid funk, neurofunk, jungle, techstep, hospital records |
+| `bass` | dubstep, brostep, grime, trip-hop, big beat |
+| `dance` | house and its variants, trance, techno, electro, rave |
+| `pop-punk` | pop punk, punk, emo, emocore, easycore, power pop |
+| `screamo` | screamo, post-hardcore, metalcore, melodic hardcore, trancecore |
+| `heavy-metal` | heavy metal, thrash, speed, power, death, prog, NWOBHM |
+| `hair-metal` | hair metal, glam metal, glam rock, arena rock, AOR — 1975-1994 |
+| `nu-metal` | nu metal, alternative metal, rapcore, industrial |
+| `classic-rock` | classic rock, prog, psychedelic, blues rock, 70s, 60s |
+| `80s-synths` | 80s, new wave, synth pop, electropop, post-punk — 1975-1992, rock excluded |
+| `indie` | indie, indie rock, indie pop, britpop, singer-songwriter |
+
+Measuring first mattered. `synthwave`, `edm`, `big room` and `hardstyle` are
+plausible tags that carry **nothing at all** here, while `techstep`, `easycore`
+and `hospital records` carry real weight. Guessing produces the first list.
+
+Three kinds of tag are never used, and there is a test enforcing it:
+
+- **Nationality** — `american` alone spans 241 artists. A passport is not a
+  sound.
+- **`rock` and `electronic`** — 340 and 275 artists, most of the library. A
+  mood that matches everything is not a mood.
+- **Artist names** — Last.fm's most popular tag for an artist is frequently
+  their own name. `green day`, `paramore` and `queen` are single-artist
+  playlists waiting to happen.
+
+### Exclusions
+
+A mood may also list `exclude`. An excluded tag drops the artist outright rather
+than docking their score, and it exists because `80s-synths` cannot be written
+any other way.
+
+`80s` is the eleventh most-played tag here, and it sits on Def Leppard and Bon
+Jovi exactly as heavily as on Eurythmics. Weighting cannot separate them,
+because the tag it would weight is the one they share. What does separate them
+is that the stadium rock also carries `hard rock` and `hair metal`, and the
+synth acts do not.
+
+Checked against live Last.fm pages, since the tag census only sees artists
+already in the library:
+
+| Artist | Tags |
+| --- | --- |
+| Eurythmics | `80s`, `new wave`, `pop`, `female vocalists`, `synth pop` |
+| Frankie Goes to Hollywood | `80s`, `new wave`, `pop`, `british`, `dance` |
+| Depeche Mode | `electronic`, `synthpop`, `new wave`, `80s`, `synth pop` |
+| Duran Duran | `new wave`, `80s`, `pop`, `synth pop`, `rock` |
+
+Four of Eurythmics' five tags are ones no mood may use. Three of the four
+artists spell it **`synth pop`** with a space; only one spells it `synthpop`.
+Guessing one spelling would have missed most of the canon.
+
+An artist qualifies when their tag weights inside a mood sum to at least 30 out
+of Last.fm's 0-100 scale. One low-weight tag is not a genre, it is somebody's
+stray opinion.
+
+`years` filters on the album's release date, which is what separates eighties
+synth records from everything else a synthpop tag drags in.
+
+`--vibes` replaces the whole set with a JSON file of the same shape, so a new
+mood does not need a new release:
+
+```json
+[{ "name": "shoegaze", "tags": ["shoegaze", "dream pop"], "min_score": 40 }]
+```
+
+Names are validated when the file is read, not when the file is written. A bad
+one would otherwise surface as a playlist created somewhere unintended.
+
+### Paths
+
+Lidarr knows where the lossless source is; the playlists have to point at the
+MP3s music-mirror made from it. The mapping strips a library root from Lidarr's
+track paths and re-roots them under the mirror, with the suffix changed.
+
+`--library-root` is derived from the common parent of the indexed artist folders
+when unset, so it agrees with Lidarr by construction rather than by being kept
+in step by hand. Override it if that guess is wrong.
+
+Entries are written **relative to the playlist file**, so one playlist works
+from the NAS, from a Mac over SMB, and from Linux, without rewriting.
+
+Each playlist is given the **owner and group of the mirror** it is written
+into. The image runs as root by default so that a bind mount of any ownership
+stays writable, and the cost of that is output owned by root — which the account
+serving the share cannot read, group bit or no group bit, because the group is
+also root. Copying the mirror's own ownership avoids having to be told what it
+should be, and does nothing when the two already agree.
+
+A track is only listed once its mirror file has been confirmed to exist. Lidarr
+holding the FLAC says nothing about whether the MP3 has been encoded yet. If a
+large number are missing, the run says so — that is what a wrong `--library-root`
+or `--mirror` looks like, since the paths then map to nothing at all.
+
+`_playlists/` survives music-mirror's prune: it only deletes `*.mp3`, and its
+empty-directory sweep skips a directory holding M3Us.
 
 ## How the ingest works
 
@@ -111,6 +330,11 @@ music-curator --report-only   # report on the store, fetch nothing
 | `--backfill-limit` | `MUSIC_CURATOR_BACKFILL_LIMIT` | `0` | Cap backfill requests per pass; 0 for no cap |
 | `--lidarr-url` | `MUSIC_CURATOR_LIDARR_URL` | unset | Lidarr base URL, e.g. `http://lidarr:8686` |
 | `--lidarr-api-key` | `MUSIC_CURATOR_LIDARR_API_KEY` | unset | Lidarr API key |
+| `--mirror` | `MUSIC_CURATOR_MIRROR` | unset | Root of the MP3 mirror; playlists go here |
+| `--library-root` | `MUSIC_CURATOR_LIBRARY_ROOT` | derived | Prefix to strip from Lidarr's paths |
+| `--playlist-limit` | `MUSIC_CURATOR_PLAYLIST_LIMIT` | `100` | Most tracks in any one playlist |
+| `--vibes` | `MUSIC_CURATOR_VIBES` | built-in | JSON file of mood definitions |
+| `--tag-limit` | `MUSIC_CURATOR_TAG_LIMIT` | `0` | Cap artist tag lookups per pass |
 | `--skip-index` | — | off | Match against the index already held |
 | `--report-only` | — | off | Report without fetching |
 
@@ -159,7 +383,8 @@ nix shell nixpkgs#python3Packages.pytest -c pytest
 | ------------------------------------------------ | ------------ |
 | Last.fm ingest and store | done |
 | Lidarr index and the scrobble-to-track matcher | done |
-| M3U playlists written into the mirror | next |
+| M3U playlists from the listening history | done |
+| Genre and mood playlists from Last.fm tags | done |
 | Cold-music report, unmonitoring what is not played | last |
 
 The cull will unmonitor cold albums in Lidarr and tag their artists. It will

compose.yaml

@@ -30,5 +30,11 @@ services:
       # Cap the backfill at this many requests per pass. Unlimited by default,
       # which finishes a long history in one go.
       # MUSIC_CURATOR_BACKFILL_LIMIT: "0"
+      # The MP3 mirror music-mirror maintains. Playlists are written into
+      # _playlists/ inside it; leave unset to skip them.
+      MUSIC_CURATOR_MIRROR: /mirror
+      # Most tracks in any one playlist.
+      # MUSIC_CURATOR_PLAYLIST_LIMIT: "100"
     volumes:
       - /mnt/tank/apps/music-curator:/data
+      - /mnt/tank/media/music-mp3:/mirror

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "music-curator"
-version = "0.2.0"
+version = "0.6.0"
 description = "Ingest a Last.fm listening history and curate a music library from it"
 readme = "README.md"
 requires-python = ">=3.11"

tests/conftest.py

@@ -1,6 +1,10 @@
+import http.server
+import io
 import json
 import os
 import sys
+import threading
+import urllib.error
 import urllib.parse
 
 import pytest
@@ -52,7 +56,9 @@ class FakeLastfm:
     is prepended to the first page with no `date`.
     """
 
-    def __init__(self, tracks=(), loved=(), nowplaying=None, outcomes=()):
+    def __init__(self, tracks=(), loved=(), nowplaying=None, outcomes=(), tags=None):
+        # Keyed by mbid or by artist name, whichever the caller asked with.
+        self.tags = tags or {}
         self.tracks = sorted(tracks, key=lambda track: int(track["date"]["uts"]), reverse=True)
         self.loved = list(loved)
         self.nowplaying = nowplaying
@@ -78,6 +84,15 @@ class FakeLastfm:
             return json.dumps(self._recent(query))
         if method == "user.getlovedtracks":
             return json.dumps(self._loved(query))
+        if method == "artist.gettoptags":
+            key = query.get("mbid") or query.get("artist", "")
+            if key not in self.tags:
+                # What the real service says for a key it cannot resolve, which
+                # for mbids is a great many artists whose pages plainly exist.
+                return json.dumps(
+                    {"error": 6, "message": "The artist you supplied could not be found"}
+                )
+            return json.dumps({"toptags": {"tag": self.tags[key]}})
         raise AssertionError(f"unexpected method {method}")
 
     def _recent(self, query):
@@ -128,7 +143,10 @@ class FakeLidarr:
     exercise the same stitching the real thing does.
     """
 
-    def __init__(self, artists=()):
+    def __init__(self, artists=(), fail=()):
+        # (path, artistId) pairs the fake refuses to serve, standing in for the
+        # 500s Lidarr returns on data it cannot hydrate.
+        self.fail = set(fail)
         self.artists, self.albums, self.tracks, self.files = [], [], [], []
         for artist_index, entry in enumerate(artists, start=1):
             artist_id = artist_index
@@ -190,10 +208,39 @@ class FakeLidarr:
         }
         self.calls.append((path, query))
 
+        artist_id = int(query.get("artistId", 0))
+        if (path, artist_id) in self.fail:
+            raise urllib.error.HTTPError(
+                url, 500, "Internal Server Error", {}, io.BytesIO(b'{"message": "boom"}')
+            )
+
+        album_ids = query.get("albumIds")
+        if path == "album" and album_ids:
+            album_id = int(album_ids)
+            if ("albumid", album_id) in self.fail:
+                raise urllib.error.HTTPError(
+                    url,
+                    500,
+                    "Internal Server Error",
+                    {},
+                    io.BytesIO(b'{"message": "Sequence contains more than one element"}'),
+                )
+            return json.dumps([row for row in self.albums if row["id"] == album_id])
+
         if path == "artist":
             return json.dumps(self.artists)
-        artist_id = int(query.get("artistId", 0))
         source = {"album": self.albums, "track": self.tracks, "trackfile": self.files}[path]
+        if path == "album" and not artist_id:
+            # Lidarr's unfiltered album endpoint returns the lot.
+            return json.dumps(source)
+        if not artist_id:
+            raise urllib.error.HTTPError(
+                url,
+                400,
+                "Bad Request",
+                {},
+                io.BytesIO(b'{"message": "artistId must be provided"}'),
+            )
         return json.dumps([row for row in source if row["artistId"] == artist_id])
 
 
@@ -207,3 +254,53 @@ def now_playing():
         "mbid": "",
         "@attr": {"nowplaying": "true"},
     }
+
+
+class _CountingServer(http.server.ThreadingHTTPServer):
+    """Counts accepted connections, which is what connection reuse is about."""
+
+    daemon_threads = True
+
+    def __init__(self, *args, **kwargs):
+        self.connections = 0
+        super().__init__(*args, **kwargs)
+
+    def process_request(self, request, client_address):
+        self.connections += 1
+        super().process_request(request, client_address)
+
+
+class _Handler(http.server.BaseHTTPRequestHandler):
+    # Without HTTP/1.1 the server closes after every response and no client
+    # could reuse anything, which would make the test prove nothing.
+    protocol_version = "HTTP/1.1"
+
+    def do_GET(self):
+        if self.path.startswith("/boom"):
+            body, status = b'{"message": "boom"}', 500
+        else:
+            body = json.dumps(
+                {"path": self.path, "key": self.headers.get("X-Api-Key")}
+            ).encode()
+            status = 200
+        self.send_response(status)
+        self.send_header("Content-Type", "application/json")
+        self.send_header("Content-Length", str(len(body)))
+        self.end_headers()
+        self.wfile.write(body)
+
+    def log_message(self, *args):
+        pass
+
+
+@pytest.fixture
+def http_server():
+    """A real local HTTP server, for the one component that talks sockets."""
+    server = _CountingServer(("127.0.0.1", 0), _Handler)
+    thread = threading.Thread(target=server.serve_forever, daemon=True)
+    thread.start()
+    try:
+        yield server, f"http://127.0.0.1:{server.server_port}"
+    finally:
+        server.shutdown()
+        server.server_close()

tests/test_music_curator.py

@@ -1,4 +1,8 @@
+import json
+import os
+import stat
 import urllib.error
+from pathlib import Path
 
 import pytest
 from conftest import FakeLastfm, FakeLidarr, make_loved, make_tracks
@@ -54,9 +58,23 @@ LIBRARY = [
     {
         "name": "The Prodigy",
         "albums": [
-            {"title": "The Fat of the Land", "tracks": [{"title": "Breathe (Remastered)"}]}
+            {
+                "title": "The Fat of the Land",
+                "tracks": [
+                    {"title": "Breathe (Remastered)"},
+                    # Credits its remixer in the title, which is the only signal
+                    # separating a real attribution miss from a title collision.
+                    {"title": "Voodoo People (Pendulum Remix)"},
+                ],
+            }
         ],
     },
+    # Held by the library in its own right, which is what puts its scrobbles
+    # inside the "artist the library holds" filter at all.
+    {
+        "name": "Pendulum",
+        "albums": [{"title": "Immersion", "tracks": [{"title": "Watercolour"}]}],
+    },
 ]
 
 
@@ -392,7 +410,7 @@ def test_keys_carry_the_play_count_and_the_span(tmp_path):
 def test_re_indexing_drops_what_lidarr_no_longer_has(tmp_path):
     """The index is Lidarr's mirror, not an accumulation of everything ever seen."""
     store = indexed(tmp_path, [scrobble_of("AC/DC", "Hells Bells")])
-    assert store.scalar("SELECT COUNT(*) FROM lidarr_artist") == 3
+    assert store.scalar("SELECT COUNT(*) FROM lidarr_artist") == 4
 
     music_curator.index_library(
         music_curator.Lidarr("http://lidarr", "key", transport=FakeLidarr(LIBRARY[:1])), store
@@ -500,3 +518,827 @@ def test_matching_is_redone_when_new_scrobbles_arrive(tmp_path):
     music_curator.run_once(client_for(api), store, "lyra", NOW, 0)
 
     assert verdict(store, "Yellowcard", "Transmission Home")[0] == "name"
+
+
+def test_albums_come_from_the_unfiltered_endpoint(tmp_path):
+    """One request, and the only path that skips albums it cannot hydrate."""
+    api = FakeLidarr(LIBRARY)
+    store = store_at(tmp_path)
+
+    music_curator.index_library(music_curator.Lidarr("http://lidarr", "key", transport=api), store)
+
+    album_calls = [query for path, query in api.calls if path == "album"]
+    assert album_calls == [{}]
+    assert store.scalar("SELECT COUNT(*) FROM lidarr_album") == 4
+
+
+def test_a_bad_album_falls_back_to_asking_per_artist(tmp_path):
+    """An album with two monitored releases throws in the resource mapper, so
+    the bulk call dies wholesale. Per artist, only its owner is lost."""
+    # The bulk call fails because artist 2 owns the offending album.
+    api = FakeLidarr(LIBRARY, fail=[("album", 0), ("album", 2)])
+    store = store_at(tmp_path)
+
+    music_curator.index_library(music_curator.Lidarr("http://lidarr", "key", transport=api), store)
+
+    assert store.scalar("SELECT COUNT(*) FROM lidarr_album WHERE artist_id = 2") == 0
+    # The other two artists keep their albums.
+    assert store.scalar("SELECT COUNT(*) FROM lidarr_album") == 3
+    assert store.get_state("index_albums_skipped") == "1"
+
+
+def test_a_bad_album_does_not_cost_that_artist_their_tracks(tmp_path):
+    """Tracks come from a different endpoint with a different mapper, so
+    matching survives an album Lidarr cannot serialise."""
+    api = FakeLidarr(LIBRARY, fail=[("album", 0), ("album", 2)])
+    store = store_at(tmp_path)
+
+    music_curator.index_library(music_curator.Lidarr("http://lidarr", "key", transport=api), store)
+
+    assert store.scalar("SELECT COUNT(*) FROM lidarr_track WHERE artist_id = 2") == 1
+    assert store.get_state("index_skipped") == "0"
+
+
+def test_an_artist_lidarr_cannot_serve_does_not_kill_the_index(tmp_path):
+    # Artist 2 is AC/DC in LIBRARY; its track lookup fails.
+    api = FakeLidarr(LIBRARY, fail=[("track", 2)])
+    store = store_at(tmp_path)
+
+    music_curator.index_library(music_curator.Lidarr("http://lidarr", "key", transport=api), store)
+
+    assert store.scalar("SELECT COUNT(*) FROM lidarr_artist") == 4
+    assert store.scalar("SELECT COUNT(*) FROM lidarr_track WHERE artist_id = 2") == 0
+    # The other two artists are indexed in full.
+    assert store.scalar("SELECT COUNT(*) FROM lidarr_track") == 5
+    assert store.get_state("index_skipped") == "1"
+
+
+def test_a_skipped_artist_is_recorded_so_the_report_can_disown_the_numbers(tmp_path):
+    """An incomplete index makes played music look cold. It has to be loud."""
+    api = FakeLidarr(LIBRARY, fail=[("trackfile", 1)])
+    store = store_at(tmp_path)
+
+    music_curator.index_library(music_curator.Lidarr("http://lidarr", "key", transport=api), store)
+    music_curator.match_library(store)
+
+    assert store.get_state("index_skipped") == "1"
+
+
+def test_a_clean_index_records_no_skips(tmp_path):
+    store = indexed(tmp_path, [])
+
+    assert store.get_state("index_skipped") == "0"
+
+
+def test_a_lidarr_error_carries_the_url_and_what_the_server_said():
+    """A bare status code sends you looking at the wrong thing entirely."""
+    api = FakeLidarr(LIBRARY, fail=[("album", 0)])
+    client = music_curator.Lidarr("http://lidarr:8686", "key", transport=api)
+
+    with pytest.raises(music_curator.LidarrError) as raised:
+        client.get("album")
+
+    message = str(raised.value)
+    assert "http://lidarr:8686/api/v1/album" in message
+    assert "HTTP 500" in message
+    assert "boom" in message
+
+
+def test_the_offending_album_is_named_not_just_its_artist(tmp_path, caplog):
+    """An artist's whole discography is too much to click through by hand."""
+    # AC/DC is artist 2; its only album is id 201, holding "Hells Bells".
+    api = FakeLidarr(LIBRARY, fail=[("album", 0), ("album", 2), ("albumid", 201)])
+    store = store_at(tmp_path)
+
+    with caplog.at_level("WARNING"):
+        music_curator.index_library(
+            music_curator.Lidarr("http://lidarr", "key", transport=api), store
+        )
+
+    assert "album id 201" in caplog.text
+    assert "Hells Bells" in caplog.text
+
+
+def test_a_transient_failure_is_not_probed_album_by_album(tmp_path, caplog):
+    """When the resolver is the problem every artist fails, and probing each of
+    them multiplies the load that caused it."""
+
+    def unresolvable(url, timeout=None, headers=None):
+        if "artistId" in url:
+            raise urllib.error.URLError("[Errno -3] Try again")
+        return json.dumps(FakeLidarr(LIBRARY).artists) if url.endswith("artist") else "[]"
+
+    store = store_at(tmp_path)
+    client = music_curator.Lidarr("http://lidarr", "key", transport=unresolvable, backoff=0)
+
+    with caplog.at_level("WARNING"):
+        music_curator.index_library(client, store)
+
+    assert "Try again" in caplog.text
+    assert "cannot serialise" not in caplog.text
+
+
+def test_a_transient_failure_is_retried():
+    attempts = []
+
+    def flaky(url, timeout=None, headers=None):
+        attempts.append(url)
+        if len(attempts) < 3:
+            raise urllib.error.URLError("[Errno -3] Try again")
+        return "[]"
+
+    client = music_curator.Lidarr("http://lidarr", "key", transport=flaky, backoff=0)
+
+    assert client.get("artist") == []
+    assert len(attempts) == 3
+
+
+def test_a_lidarr_500_is_not_retried():
+    """It is an exception inside Lidarr's serialisation, not a busy server."""
+    api = FakeLidarr(LIBRARY, fail=[("album", 0)])
+    client = music_curator.Lidarr("http://lidarr", "key", transport=api, backoff=0)
+
+    with pytest.raises(music_curator.LidarrError) as raised:
+        client.get("album")
+
+    assert raised.value.transient is False
+    assert len(api.calls) == 1
+
+
+def test_keep_alive_uses_one_connection_for_many_requests(http_server):
+    """The point of the whole class: one DNS lookup and one socket, not N."""
+    server, base = http_server
+    transport = music_curator.KeepAlive()
+
+    try:
+        for index in range(5):
+            body = transport(f"{base}/api/v1/artist?n={index}", headers={"X-Api-Key": "key"})
+            assert json.loads(body)["key"] == "key"
+    finally:
+        transport.close()
+
+    assert server.connections == 1
+
+
+def test_keep_alive_maps_an_error_status_onto_httperror(http_server):
+    _, base = http_server
+    transport = music_curator.KeepAlive()
+
+    try:
+        with pytest.raises(urllib.error.HTTPError) as raised:
+            transport(f"{base}/boom", headers={"X-Api-Key": "key"})
+        assert raised.value.code == 500
+        assert music_curator.error_detail(raised.value) == "boom"
+    finally:
+        transport.close()
+
+
+def test_lidarr_talks_to_a_real_server_through_keep_alive(http_server):
+    server, base = http_server
+    client = music_curator.Lidarr(base, "secret")
+
+    try:
+        assert client.get("artist", {"x": 1})["key"] == "secret"
+        assert client.get("album")["path"] == "/api/v1/album"
+    finally:
+        client.transport.close()
+
+    assert server.connections == 1
+
+
+# Titles taken verbatim from a real coverage report's unmatched list. Each one
+# was a genuine miss before the normaliser handled it.
+@pytest.mark.parametrize(
+    ("scrobbled", "tagged"),
+    [
+        ("Self vs Self (feat. In Flames)", "Self vs Self"),
+        ("Grime Battle of Hastings (feat. The Town Crier)", "Grime Battle of Hastings"),
+        ("Gold Dust - Shy FX Re-Edit", "Gold Dust"),
+        ("Back To Your Roots - Friction & K-Tee Remix", "Back To Your Roots"),
+        ("Constellations - Forza Horizon 3 VIP", "Constellations"),
+        ("Everyday (Netsky Remix)", "Everyday"),
+        ("Voodoo People [Pendulum Remix] [Live At Brixton Academy]", "Voodoo People"),
+    ],
+)
+def test_real_unmatched_titles_now_agree_with_their_tags(scrobbled, tagged):
+    assert music_curator.normalise(scrobbled) == music_curator.normalise(tagged)
+
+
+@pytest.mark.parametrize(
+    "title",
+    [
+        "Dancing with Myself",
+        "(Don't Fear) The Reaper",
+        "Live and Let Die",
+        "Radio Ga Ga",
+        "Editors",
+        "Mixed Emotions",
+        "Vipassana",
+    ],
+)
+def test_the_version_words_do_not_eat_ordinary_titles(title):
+    """Every one of these contains a version word and must survive intact."""
+    assert music_curator.normalise(title) == music_curator.normalise(title.lower())
+    assert len(music_curator.normalise(title).split()) == len(title.split())
+
+
+def test_a_bracketed_guest_credit_matches_the_bare_tag(tmp_path):
+    """Whitespace-then-feat misses the bracketed form, which is most of them."""
+    store = indexed(tmp_path, [scrobble_of("Yellowcard", "Here I Am Alive (feat. Someone)")])
+
+    method, track_id = verdict(store, "Yellowcard", "Here I Am Alive (feat. Someone)")
+    assert method == "name"
+    assert track_id is not None
+
+
+def test_misses_are_split_by_whether_the_library_holds_the_title(tmp_path):
+    """Owning an artist is a weak proxy for owning a track. Counting both as
+    matcher failures overstates the problem and would over-block the cull."""
+    store = indexed(
+        tmp_path,
+        [
+            # The library holds "Hells Bells", but under AC/DC, not Yellowcard:
+            # an attribution disagreement, and a real miss.
+            scrobble_of("Yellowcard", "Hells Bells"),
+            # Yellowcard is in the library; this track is not, under any artist.
+            scrobble_of("Yellowcard", "A Single She Never Bought"),
+        ],
+    )
+
+    def count(extra):
+        return store.connection.execute(
+            "SELECT COUNT(*) FROM scrobble_key k WHERE k.track_id IS NULL"
+            "   AND EXISTS (SELECT 1 FROM lidarr_artist a WHERE a.norm_name = k.norm_artist)"
+            f"  {extra}"
+        ).fetchone()[0]
+
+    assert count("") == 2
+    assert count("AND EXISTS (SELECT 1 FROM lidarr_track t WHERE t.norm_title = k.norm_track)") == 1
+
+
+def test_the_report_survives_the_attribution_split(tmp_path):
+    store = indexed(tmp_path, [scrobble_of("Yellowcard", "Hells Bells")])
+
+    music_curator.report(store, NOW)
+
+
+def attribution_pairs(store):
+    """Unmatched pairs where the library's own title credits the scrobbled artist."""
+    return [
+        row["artist"]
+        for row in store.connection.execute(
+            "SELECT k.artist FROM scrobble_key k WHERE k.track_id IS NULL"
+            "   AND EXISTS (SELECT 1 FROM lidarr_artist a WHERE a.norm_name = k.norm_artist)"
+            "   AND EXISTS (SELECT 1 FROM lidarr_track t"
+            "                WHERE t.norm_title = k.norm_track"
+            "                  AND instr(lower(t.title), lower(k.artist)) > 0)"
+        )
+    ]
+
+
+def test_a_shared_title_is_not_an_attribution_miss(tmp_path):
+    """Across fifty thousand tracks, titles collide constantly: "Everyday" is
+    Rusko and also Def Leppard. Matching those would be worse than missing."""
+    store = indexed(tmp_path, [scrobble_of("Yellowcard", "Hells Bells")])
+
+    # The library holds "Hells Bells", by AC/DC, and its title says nothing
+    # about Yellowcard. A collision, not a miss.
+    assert verdict(store, "Yellowcard", "Hells Bells") == ("none", None)
+    assert attribution_pairs(store) == []
+
+
+def test_a_remix_credited_in_the_library_title_is_an_attribution_miss(tmp_path):
+    """The library has "Voodoo People (Pendulum Remix)" under The Prodigy; the
+    scrobble credits Pendulum. Same song, different filing."""
+    store = indexed(tmp_path, [scrobble_of("Pendulum", "Voodoo People")])
+
+    assert attribution_pairs(store) == ["Pendulum"]
+
+
+def test_the_title_index_is_used_for_the_report_lookup(tmp_path):
+    """Without it the report scans every track for every unmatched key: forty-
+    seven seconds on a real library."""
+    store = indexed(tmp_path, [])
+
+    plan = "\n".join(
+        row[-1]
+        for row in store.connection.execute(
+            "EXPLAIN QUERY PLAN SELECT 1 FROM scrobble_key k WHERE k.track_id IS NULL"
+            " AND EXISTS (SELECT 1 FROM lidarr_track t WHERE t.norm_title = k.norm_track)"
+        )
+    )
+    assert "lidarr_track_title" in plan, plan
+
+
+def playlist_library(tmp_path):
+    """A library on disk, with mirror MP3s beside the Lidarr source paths."""
+    source = tmp_path / "music"
+    mirror = tmp_path / "mirror"
+    library = [
+        {
+            "name": "Played Band",
+            "albums": [
+                {
+                    "title": "Known",
+                    "tracks": [{"title": "Hit"}, {"title": "Album Track"}],
+                }
+            ],
+        },
+        {
+            "name": "Silent Band",
+            "albums": [{"title": "Unknown", "tracks": [{"title": "Never Heard"}]}],
+        },
+    ]
+    api = FakeLidarr(library)
+    # FakeLidarr invents /music/<artist>/<album>/<title>.flac; put the mirror
+    # MP3s at the paths music-mirror would have produced from those.
+    for handle in api.files:
+        relative = Path(handle["path"]).relative_to("/music")
+        handle["path"] = str(source / relative)
+        mp3 = (mirror / relative).with_suffix(".mp3")
+        mp3.parent.mkdir(parents=True, exist_ok=True)
+        mp3.write_bytes(b"not really an mp3")
+    for artist in api.artists:
+        artist["path"] = str(source / Path(artist["path"]).name)
+    return api, source, mirror
+
+
+def test_playlists_are_written_into_the_mirror(tmp_path):
+    api, source, mirror = playlist_library(tmp_path)
+    store = store_at(tmp_path)
+    ingest(store, FakeLastfm([scrobble_of("Played Band", "Hit")] * 1))
+    music_curator.index_library(
+        music_curator.Lidarr("http://lidarr", "key", transport=api), store
+    )
+    music_curator.match_library(store)
+
+    music_curator.build_playlists(store, mirror, str(source), 100, NOW)
+
+    written = sorted(p.name for p in (mirror / "_playlists").glob("*.m3u"))
+    assert written == [
+        "all-time.m3u",
+        "deep-cuts.m3u",
+        "heavy-rotation.m3u",
+        "neglected.m3u",
+        "unheard-favourites.m3u",
+        "unheard.m3u",
+    ]
+
+    played = (mirror / "_playlists" / "all-time.m3u").read_text().splitlines()
+    assert played[0] == "#EXTM3U"
+    assert played[1].startswith("#EXTINF:")
+    assert "Played Band - Hit" in played[1]
+    # Relative to the playlist file, so the same file works from any mount.
+    assert played[2] == "../Played Band/Known/Hit.mp3"
+    assert (mirror / "_playlists" / "all-time.m3u").parent.joinpath(played[2]).resolve().is_file()
+
+
+def test_an_unplayed_track_lands_in_the_unheard_playlist(tmp_path):
+    api, source, mirror = playlist_library(tmp_path)
+    store = store_at(tmp_path)
+    ingest(store, FakeLastfm([scrobble_of("Played Band", "Hit")]))
+    music_curator.index_library(
+        music_curator.Lidarr("http://lidarr", "key", transport=api), store
+    )
+    music_curator.match_library(store)
+
+    music_curator.build_playlists(store, mirror, str(source), 100, NOW)
+
+    unheard = (mirror / "_playlists" / "unheard.m3u").read_text()
+    assert "Never Heard" in unheard
+    assert "Album Track" in unheard
+    # The one thing that was played must not be in it.
+    assert "Played Band - Hit\n" not in unheard
+
+
+def test_a_track_with_no_mirror_file_is_left_out(tmp_path):
+    """The index knows what Lidarr holds; whether music-mirror has encoded the
+    MP3 yet is a different question, and is checked rather than assumed."""
+    api, source, mirror = playlist_library(tmp_path)
+    for mp3 in mirror.rglob("*.mp3"):
+        mp3.unlink()
+    store = store_at(tmp_path)
+    ingest(store, FakeLastfm([scrobble_of("Played Band", "Hit")]))
+    music_curator.index_library(
+        music_curator.Lidarr("http://lidarr", "key", transport=api), store
+    )
+    music_curator.match_library(store)
+
+    total = music_curator.build_playlists(store, mirror, str(source), 100, NOW)
+
+    assert total == 0
+    assert (mirror / "_playlists" / "all-time.m3u").read_text() == "#EXTM3U\n"
+
+
+def test_the_library_root_is_derived_from_the_artist_folders(tmp_path):
+    api, source, mirror = playlist_library(tmp_path)
+    store = store_at(tmp_path)
+    music_curator.index_library(
+        music_curator.Lidarr("http://lidarr", "key", transport=api), store
+    )
+
+    assert music_curator.library_root_of(store) == str(source)
+
+
+def test_the_playlist_limit_is_honoured(tmp_path):
+    api, source, mirror = playlist_library(tmp_path)
+    store = store_at(tmp_path)
+    music_curator.index_library(
+        music_curator.Lidarr("http://lidarr", "key", transport=api), store
+    )
+    music_curator.match_library(store)
+
+    music_curator.build_playlists(store, mirror, str(source), 1, NOW)
+
+    unheard = (mirror / "_playlists" / "unheard.m3u").read_text().splitlines()
+    assert len([line for line in unheard if line.startswith("#EXTINF")]) == 1
+
+
+def test_the_rotation_moves_weekly_not_every_pass(tmp_path):
+    """A playlist that reorders on every pass is one that has to be re-imported
+    on every pass; the Music app imports a snapshot, it does not track a file."""
+    api, source, mirror = playlist_library(tmp_path)
+    store = store_at(tmp_path)
+    music_curator.index_library(
+        music_curator.Lidarr("http://lidarr", "key", transport=api), store
+    )
+    music_curator.match_library(store)
+
+    def unheard_at(when):
+        music_curator.build_playlists(store, mirror, str(source), 100, when)
+        return (mirror / "_playlists" / "unheard.m3u").read_text()
+
+    same_week = unheard_at(NOW), unheard_at(NOW + 3600)
+    assert same_week[0] == same_week[1]
+
+
+def test_playlists_are_group_readable(tmp_path):
+    api, source, mirror = playlist_library(tmp_path)
+    store = store_at(tmp_path)
+    music_curator.index_library(
+        music_curator.Lidarr("http://lidarr", "key", transport=api), store
+    )
+    music_curator.match_library(store)
+
+    music_curator.build_playlists(store, mirror, str(source), 100, NOW)
+
+    for playlist in (mirror / "_playlists").glob("*.m3u"):
+        assert playlist.stat().st_mode & stat.S_IRGRP, playlist
+
+
+def test_tag_weights_fall_back_to_rank_when_no_count_is_sent():
+    """The documented sample carries only a name and a URL; a live response also
+    carries a count. Neither may be relied on alone."""
+    with_count = music_curator.parse_tags(
+        {"toptags": {"tag": [{"name": "Screamo", "count": 100}, {"name": "emo", "count": 40}]}}
+    )
+    assert with_count == [("screamo", 100), ("emo", 40)]
+
+    without = music_curator.parse_tags(
+        {"toptags": {"tag": [{"name": "screamo"}, {"name": "emo"}]}}
+    )
+    assert [tag for tag, _ in without] == ["screamo", "emo"]
+    assert without[0][1] > without[1][1]
+
+
+def test_a_lone_tag_is_not_a_list():
+    assert music_curator.parse_tags({"toptags": {"tag": {"name": "dnb", "count": 90}}}) == [
+        ("dnb", 90)
+    ]
+
+
+def test_an_artist_lastfm_cannot_answer_for_is_not_asked_again(tmp_path):
+    """Recording the fetch even when it returns nothing is what stops a pass
+    spending a request per unknown artist, forever."""
+    api, source, mirror = playlist_library(tmp_path)
+    store = store_at(tmp_path)
+    music_curator.index_library(
+        music_curator.Lidarr("http://lidarr", "key", transport=api), store
+    )
+    lastfm = FakeLastfm(tags={})
+
+    assert music_curator.sync_tags(client_for(lastfm), store, NOW) == 2
+    before = len(lastfm.calls)
+
+    assert music_curator.sync_tags(client_for(lastfm), store, NOW) == 0
+    assert len(lastfm.calls) == before
+
+
+def test_tags_are_refetched_once_they_go_stale(tmp_path):
+    api, source, mirror = playlist_library(tmp_path)
+    store = store_at(tmp_path)
+    music_curator.index_library(
+        music_curator.Lidarr("http://lidarr", "key", transport=api), store
+    )
+    lastfm = FakeLastfm(tags={})
+    music_curator.sync_tags(client_for(lastfm), store, NOW)
+
+    later = NOW + music_curator.TAG_REFRESH_SECONDS + 1
+    assert music_curator.sync_tags(client_for(lastfm), store, later) == 2
+
+
+def tagged_store(tmp_path, tags):
+    api, source, mirror = playlist_library(tmp_path)
+    store = store_at(tmp_path)
+    music_curator.index_library(
+        music_curator.Lidarr("http://lidarr", "key", transport=api), store
+    )
+    music_curator.sync_tags(client_for(FakeLastfm(tags=tags)), store, NOW)
+    return store, source, mirror
+
+
+def test_a_vibe_selects_by_tag(tmp_path):
+    store, source, mirror = tagged_store(
+        tmp_path,
+        {
+            "Played Band": [{"name": "screamo", "count": 100}],
+            "Silent Band": [{"name": "classic rock", "count": 100}],
+        },
+    )
+    vibes = [{"name": "screamo", "tags": ["screamo"]}]
+
+    music_curator.build_vibe_playlists(store, vibes, mirror, str(source), 100, NOW)
+
+    written = (mirror / "_playlists" / "screamo.m3u").read_text()
+    assert "Played Band" in written
+    assert "Silent Band" not in written
+
+
+def test_a_weakly_tagged_artist_is_below_the_threshold(tmp_path):
+    """A single low-weight tag is not a genre, it is somebody's stray opinion."""
+    store, source, mirror = tagged_store(
+        tmp_path, {"Played Band": [{"name": "screamo", "count": 3}]}
+    )
+    vibes = [{"name": "screamo", "tags": ["screamo"]}]
+
+    music_curator.build_vibe_playlists(store, vibes, mirror, str(source), 100, NOW)
+
+    assert (mirror / "_playlists" / "screamo.m3u").read_text() == "#EXTM3U\n"
+
+
+def test_a_vibe_can_be_restricted_by_release_year(tmp_path):
+    """What separates eighties synth records from everything else a synthpop
+    tag drags in."""
+    store, source, mirror = tagged_store(
+        tmp_path, {"Played Band": [{"name": "synthpop", "count": 100}]}
+    )
+    inside = [{"name": "eighties", "tags": ["synthpop"], "years": [1975, 1992]}]
+    outside = [{"name": "nineties", "tags": ["synthpop"], "years": [1993, 1999]}]
+
+    # FakeLidarr dates every album 2019, so neither window should catch it.
+    music_curator.build_vibe_playlists(store, inside, mirror, str(source), 100, NOW)
+    music_curator.build_vibe_playlists(store, outside, mirror, str(source), 100, NOW)
+    assert (mirror / "_playlists" / "eighties.m3u").read_text() == "#EXTM3U\n"
+    assert (mirror / "_playlists" / "nineties.m3u").read_text() == "#EXTM3U\n"
+
+    modern = [{"name": "modern", "tags": ["synthpop"], "years": [2000, 2030]}]
+    music_curator.build_vibe_playlists(store, modern, mirror, str(source), 100, NOW)
+    assert "Played Band" in (mirror / "_playlists" / "modern.m3u").read_text()
+
+
+def test_the_built_in_vibes_are_all_usable_filenames():
+    for vibe in music_curator.DEFAULT_VIBES:
+        assert music_curator.SAFE_VIBE_NAME.fullmatch(vibe["name"]), vibe["name"]
+        assert vibe["tags"]
+
+
+def test_a_vibes_file_replaces_the_built_in_set(tmp_path):
+    path = tmp_path / "vibes.json"
+    path.write_text(json.dumps([{"name": "mine", "tags": ["shoegaze"]}]))
+
+    assert music_curator.load_vibes(str(path)) == ({"name": "mine", "tags": ["shoegaze"]},)
+    assert music_curator.load_vibes(None) is music_curator.DEFAULT_VIBES
+
+
+@pytest.mark.parametrize(
+    "content",
+    [
+        '{"not": "a list"}',
+        '[{"tags": ["x"]}]',
+        '[{"name": "../escape", "tags": ["x"]}]',
+        '[{"name": "ok"}]',
+        "not json at all",
+    ],
+)
+def test_a_bad_vibes_file_is_refused_up_front(tmp_path, content):
+    """A bad name would otherwise surface as a file written somewhere
+    unintended, which is a poor way to learn about a typo."""
+    path = tmp_path / "vibes.json"
+    path.write_text(content)
+
+    with pytest.raises(ValueError):
+        music_curator.load_vibes(str(path))
+
+
+def test_tags_are_looked_up_by_name_not_by_mbid(tmp_path):
+    """Last.fm's mbid index is stale: it cannot find Devo or Escape the Fate by
+    one, though their pages plainly exist. Its name index can."""
+    api, source, mirror = playlist_library(tmp_path)
+    store = store_at(tmp_path)
+    music_curator.index_library(
+        music_curator.Lidarr("http://lidarr", "key", transport=api), store
+    )
+    lastfm = FakeLastfm(tags={"Played Band": [{"name": "screamo", "count": 90}]})
+
+    music_curator.sync_tags(client_for(lastfm), store, NOW)
+
+    asked = [call for call in lastfm.calls if call.get("method") == "artist.getTopTags"]
+    # The name is always tried first; the mbid only appears as a fallback for
+    # the artist that the name could not resolve.
+    assert "artist" in asked[0]
+    assert [call for call in asked if call.get("artist") == "Played Band"]
+    assert not [call for call in asked if call.get("mbid") == "artist-mbid-1"]
+    assert store.scalar("SELECT COUNT(*) FROM artist_tag WHERE tag = 'screamo'") == 1
+
+
+def test_the_mbid_is_tried_when_the_name_is_not_found(tmp_path):
+    """Kept only for a name Lidarr spells differently to Last.fm."""
+    api, source, mirror = playlist_library(tmp_path)
+    store = store_at(tmp_path)
+    music_curator.index_library(
+        music_curator.Lidarr("http://lidarr", "key", transport=api), store
+    )
+    lastfm = FakeLastfm(tags={"artist-mbid-1": [{"name": "dnb", "count": 80}]})
+
+    music_curator.sync_tags(client_for(lastfm), store, NOW)
+
+    assert store.scalar("SELECT COUNT(*) FROM artist_tag WHERE tag = 'dnb'") == 1
+    asked = [call for call in lastfm.calls if call.get("method") == "artist.getTopTags"]
+    assert any("mbid" in call for call in asked)
+
+
+def test_an_artist_neither_key_resolves_is_recorded_and_not_retried(tmp_path):
+    """Otherwise every pass spends a request on it again, for ever."""
+    api, source, mirror = playlist_library(tmp_path)
+    store = store_at(tmp_path)
+    music_curator.index_library(
+        music_curator.Lidarr("http://lidarr", "key", transport=api), store
+    )
+    lastfm = FakeLastfm(tags={})
+
+    assert music_curator.sync_tags(client_for(lastfm), store, NOW) == 2
+    spent = len(lastfm.calls)
+
+    assert music_curator.sync_tags(client_for(lastfm), store, NOW) == 0
+    assert len(lastfm.calls) == spent
+
+
+def test_a_real_failure_is_not_recorded_so_the_next_pass_retries(tmp_path):
+    """A rate limit is not the same as an artist not existing."""
+    api, source, mirror = playlist_library(tmp_path)
+    store = store_at(tmp_path)
+    music_curator.index_library(
+        music_curator.Lidarr("http://lidarr", "key", transport=api), store
+    )
+    lastfm = FakeLastfm(
+        tags={"Played Band": [], "Silent Band": []},
+        outcomes=[{"error": 10, "message": "Invalid API key"}],
+    )
+
+    music_curator.sync_tags(client_for(lastfm), store, NOW)
+
+    # One artist failed hard and must still be pending.
+    assert store.scalar("SELECT COUNT(*) FROM artist_tag_fetched") == 1
+
+
+# Tags that carry real weight in a real library, against tags that describe a
+# passport, span most of the collection, or are somebody's artist name.
+USELESS_TAGS = {
+    "american", "british", "australian", "canadian", "swedish", "dutch", "german",
+    "scottish", "english", "uk", "usa", "canada",
+    "rock", "electronic", "pop", "alternative", "metal ", "all", "heavy",
+    "female vocalists", "male vocalists", "female vocalist",
+    "my top songs", "cover", "covers", "not emo",
+    "green day", "paramore", "queen", "bon jovi", "shinedown", "aerosmith",
+    "journey", "fleetwood mac",
+}
+
+
+def test_no_mood_selects_on_a_useless_tag():
+    """Nationality is not a sound; `rock` and `electronic` span most of the
+    library; and Last.fm's top tag for an artist is often their own name."""
+    for vibe in music_curator.DEFAULT_VIBES:
+        overlap = {tag.casefold() for tag in vibe["tags"]} & USELESS_TAGS
+        assert not overlap, f"{vibe['name']} selects on {overlap}"
+
+
+def test_mood_names_are_unique():
+    names = [vibe["name"] for vibe in music_curator.DEFAULT_VIBES]
+    assert len(names) == len(set(names))
+
+
+ROCK_TAGS = {
+    "classic rock", "hard rock", "blues rock", "southern rock", "arena rock",
+    "glam rock", "hair metal", "glam metal", "heavy metal", "metal", "art rock",
+    "psychedelic rock", "progressive rock", "rock and roll", "rock n roll",
+}
+
+
+def test_a_decade_tag_in_a_non_rock_mood_must_exclude_the_rock():
+    """`80s` sits on Def Leppard and Bon Jovi as heavily as on Eurythmics, so a
+    mood that reaches for a decade without wanting rock has to say so. A mood
+    that does want it -- classic-rock reaching for 70s -- is exempt."""
+    for vibe in music_curator.DEFAULT_VIBES:
+        tags = {tag.casefold() for tag in vibe["tags"]}
+        decades = {"60s", "70s", "80s", "90s"} & tags
+        if not decades or tags & ROCK_TAGS:
+            continue
+        excluded = {tag.casefold() for tag in vibe.get("exclude", [])}
+        missing = {"hard rock", "hair metal"} - excluded
+        assert not missing, f"{vibe['name']} selects on {decades} without excluding {missing}"
+
+
+def test_the_eighties_mood_covers_the_canon():
+    """Depeche Mode, Duran Duran, Eurythmics and Frankie Goes to Hollywood --
+    checked against their live Last.fm tags. Three of the four carry "synth pop"
+    with a space; only one carries it without."""
+    synths = next(v for v in music_curator.DEFAULT_VIBES if v["name"] == "80s-synths")
+    tags = {t.casefold() for t in synths["tags"]}
+    for artist_tags in (
+        {"80s", "new wave", "pop", "female vocalists", "synth pop"},   # Eurythmics
+        {"80s", "new wave", "pop", "british", "dance"},                # Frankie
+        {"electronic", "synthpop", "new wave", "80s", "synth pop"},    # Depeche Mode
+        {"new wave", "80s", "pop", "synth pop", "rock"},               # Duran Duran
+    ):
+        assert tags & artist_tags, artist_tags
+    assert synths["years"] == [1975, 1992]
+
+
+def test_an_excluded_tag_drops_the_artist(tmp_path):
+    """Weighting cannot separate the eighties synth acts from the eighties
+    stadium rock, because the tag they would be weighted on is the one they
+    share."""
+    store, source, mirror = tagged_store(
+        tmp_path,
+        {
+            "Played Band": [{"name": "80s", "count": 100}, {"name": "hard rock", "count": 90}],
+            "Silent Band": [{"name": "80s", "count": 100}, {"name": "synth pop", "count": 90}],
+        },
+    )
+    vibes = [{"name": "eighties", "tags": ["80s", "synth pop"], "exclude": ["hard rock"]}]
+
+    music_curator.build_vibe_playlists(store, vibes, mirror, str(source), 100, NOW)
+
+    written = (mirror / "_playlists" / "eighties.m3u").read_text()
+    assert "Silent Band" in written
+    assert "Played Band" not in written
+
+
+def test_a_vibe_cannot_both_select_and_exclude_a_tag(tmp_path):
+    path = tmp_path / "vibes.json"
+    path.write_text(json.dumps([{"name": "x", "tags": ["80s"], "exclude": ["80s"]}]))
+
+    with pytest.raises(ValueError, match="selects on and excludes"):
+        music_curator.load_vibes(str(path))
+
+
+def test_matching_ownership_is_a_no_op_when_it_already_agrees(tmp_path):
+    target = tmp_path / "file"
+    target.write_text("x")
+
+    assert music_curator.match_ownership(target, tmp_path) is False
+
+
+def test_ownership_failure_is_tolerated(tmp_path):
+    """Not permitted unless running as root -- which is exactly the case where
+    the ownership is already whatever the caller runs as."""
+    target = tmp_path / "file"
+    target.write_text("x")
+
+    # uid 0 from a non-root test process: refused, and must not raise.
+    assert music_curator.set_ownership(target, 0, 0) is False
+
+
+def test_a_playlist_is_chowned_to_match_the_mirror(tmp_path, monkeypatch):
+    """The image runs as root, so its output is root-owned, and a root-owned
+    playlist in an apps-owned mirror is unreadable to whatever serves it."""
+    api, source, mirror = playlist_library(tmp_path)
+    store = store_at(tmp_path)
+    music_curator.index_library(
+        music_curator.Lidarr("http://lidarr", "key", transport=api), store
+    )
+    music_curator.match_library(store)
+
+    attempted = []
+    real_stat = music_curator.Path.stat
+
+    def pretend_mirror_is_owned_by_568(self, *args, **kwargs):
+        info = real_stat(self, *args, **kwargs)
+        if self == mirror:
+            return os.stat_result(
+                (info.st_mode, info.st_ino, info.st_dev, info.st_nlink, 568, 568,
+                 info.st_size, int(info.st_atime), int(info.st_mtime), int(info.st_ctime))
+            )
+        return info
+
+    monkeypatch.setattr(music_curator.Path, "stat", pretend_mirror_is_owned_by_568)
+    monkeypatch.setattr(
+        music_curator.os, "chown", lambda p, u, g: attempted.append((str(p), u, g))
+    )
+
+    music_curator.build_playlists(store, mirror, str(source), 100, NOW)
+
+    assert attempted, "no ownership was applied"
+    assert all(tuple(owner) == (568, 568) for _, *owner in attempted)
+    # The temporary file, before the rename, never the finished playlist.
+    assert all(path.endswith(".part") or path.endswith("_playlists") for path, *_ in attempted)