Files
Varaverk/Arrs_Stack/sonarr_classification_scan.sh
T
Gmer4Lfe 79059c20fa Add --move mode to sonarr_classification_scan.sh, matching Radarr
Acts on forward misplacements and reverse-kids-leak (adult content in
the kids root — not a judgment call, same reasoning as the Radarr
side), moving to the new SONARR_GENERAL_ROOT config var. Reverse-anime
-leak stays report-only — deliberate style placements like Castlevania
are common and valid there. episodeFileCount is Sonarr's hasFile
equivalent: series with 0 files get relocated + an immediate
SeriesSearch instead of a file move. One series at a time, verified
after each — matches the race condition found doing this by hand for
the Fate/Zero and Fate/Stay Night moves earlier in the session.
2026-07-17 13:02:22 -04:00

416 lines
21 KiB
Bash
Executable File

#!/bin/bash
# ==============================================================================================
# ============================ Sonarr Content Classification Scan ==============================
# ==============================================================================================
#
# PURPOSE
# ─────────────────────────────────────────────────────────────────────────────
# Same problem as radarr_classification_scan.sh, TV side: Overseerr lets any user request a
# show into the wrong root folder (kids shows added to the general TV share, anime added to
# Kids_Tv_Shows, etc.). This script reads Sonarr's tracked series list and classifies every
# series as anime / kids-only / regular using metadata signals alone (genre, certification,
# network, original language) — then reports where a series' computed classification
# disagrees with the root folder it's actually sitting in, in both directions:
#
# FORWARD — a series classified as anime/kids is sitting outside its dedicated root
# REVERSE — a series sitting inside the kids/anime root doesn't match that classification
#
# Report-only by default. Every rule below was validated against this library's real data
# before being adopted — see the companion comment block in master.conf above the curated
# lists. Pass --move to actually act (see MOVE MODE below) — nothing writes to Sonarr unless
# that flag is given.
#
# ==============================================================================================
# CLASSIFICATION RULES — DIFFERENT FIELD MODEL THAN RADARR, NOT A COPY-PASTE
# ==============================================================================================
#
# TV metadata (TheTVDB, via Sonarr) shapes these signals differently than movie metadata
# (TMDb, via Radarr) — every difference below was confirmed live, not assumed:
# - Sonarr has an explicit "Anime" genre tag; Radarr does not.
# - Sonarr uses a single "network" field (TheTVDB's broadcaster), not "studio".
# - Certification is on the US TV Parental Guidelines scale (TV-Y/TV-Y7/TV-G/TV-PG/
# TV-14/TV-MA), not the MPAA scale — the tiers do not mean the same thing at the same
# position (TV-G is "general audience", not "for children", unlike movie G).
#
# is_anime:
# genre "Anime" (corroborated by Japanese language OR a Japan network — the bare tag alone
# produced a real false positive: "Craig of the Creek", an all-American Cartoon Network
# show, carries an "Anime" genre tag on TheTVDB for no discernible reason)
# OR (genre Animation AND originalLanguage Japanese)
# OR network in SONARR_ANIME_NETWORKS
# Always wins over kids when both could apply — explicit priority, not a tiebreak.
#
# is_kids ("kids will end up watching this alone" — NOT "family show night"):
# not is_anime AND (
# genre "Children" (NOT "Family" — see below)
# OR certification in (TV-Y, TV-Y7) (NOT TV-G — see below)
# OR network in SONARR_KIDS_NETWORKS
# )
# "Family" genre and "TV-G" certification were both tested standalone and rejected —
# both catch general-audience live-action content the whole household watches together
# (I Love Lucy, The Brady Bunch, Full House, Homestead Rescue), not kids-only content.
# Blanket "Animation" genre was also tested and rejected — it's dominated on TV by adult
# animated sitcoms (Rick and Morty, BoJack Horseman, Family Guy, South Park), unlike the
# movie side where it's a usable (gated) signal.
#
# No junk-detection tier here (unlike Radarr) — TheTVDB's ratings/imdbId data is far
# sparser than TMDb's even for completely legitimate shows (confirmed live: "The Pussycat
# Dolls Present: The Search for the Next Doll", a real 2007 MTV show, has ratings.votes=0
# and imdbId=null) — the vote-count heuristic that works for Radarr would flag real content
# for removal here, so it's deliberately not reused. --remove-junk from the Radarr script has
# no Sonarr equivalent for the same reason.
#
# ==============================================================================================
# MOVE MODE (--move)
# ==============================================================================================
#
# Acts on FORWARD misplacements (classified anime/kids, sitting in the wrong root) and on
# REVERSE-KIDS leaks (adult certification sitting in the kids root — moved back to
# SONARR_GENERAL_ROOT). Does NOT act on REVERSE-ANIME leaks — those are genuine judgment
# calls, since deliberate style placements (Castlevania-type Western/Chinese animation
# grouped with anime by choice) legitimately live in the anime root without matching the
# anime signal.
#
# episodeFileCount is Sonarr's equivalent of Radarr's hasFile — a series can have 0 files
# (fully monitored, nothing downloaded) even while correctly classified. Those get their
# rootFolderPath/path corrected and an immediate SeriesSearch triggered rather than a file
# move (mirrors radarr_classification_scan.sh's handling of hasFile=false movies).
#
# One series at a time, verified after each — a rapid-fire batch of Sonarr moves is exactly
# what raced Sonarr's own file-move worker earlier this session doing this by hand (2 series
# reported "success" while their files sat at an intermediate path). A short sleep plus a
# real re-fetch-and-check after every single move catches that here before it can compound.
#
# ==============================================================================================
# DESIGN PRINCIPLES
# ==============================================================================================
#
# Curated Lists, Not Bare Genre/Cert Matching — see master.conf comments for exclusions.
# Cache-First — arr_get_tracked_data() same as sonarr_cleanup.sh, single call regardless
# of library size. Refreshed after --move writes so no other script reads stale data.
#
# ==============================================================================================
# CONFIGURATION
# ==============================================================================================
#
# host*.conf
# SONARR_URL / SONARR_API_KEY / SONARR_TV_ROOT — existing, aliased by detect_hosts()
# SONARR_GENERAL_ROOT / SONARR_KIDS_ROOT / SONARR_ANIME_ROOT — rootFolderPath literals as
# reported by the API (e.g. "/tv", "/kids tv", "/ext-anime-shows") — leave blank on a
# host with no dedicated root for that category; the corresponding checks are skipped,
# not treated as an error.
#
# master.conf
# SONARR_ANIME_NETWORKS / SONARR_KIDS_NETWORKS — curated network allowlists
# SONARR_VERSION_MAJOR — expected API major version (reused from sonarr_cleanup.sh)
#
# ==============================================================================================
# RUNTIME MODES
# ==============================================================================================
#
# sonarr_classification_scan.sh — normal run, prints report
# sonarr_classification_scan.sh --log — verbose (per-series list)
# sonarr_classification_scan.sh --status — show config and exit
# sonarr_classification_scan.sh --move — act on forward misplacements + reverse-kids-leak
#
# ==============================================================================================
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
source "$SCRIPT_DIR/../load_config.sh"
# --move is a script-local flag, not one parse_args recognizes — check the raw args before
# they get filtered into PARSED_ARGS.
MOVE_MODE=false
for _arg in "$@"; do
[[ "$_arg" == "--move" ]] && MOVE_MODE=true
done
unset _arg
parse_args "$@"
# ==============================================================================================
# ━━━ Setup ━━━
# ==============================================================================================
if [[ "$EUID" -ne 0 ]]; then
error "Must be run as root"
exit 1
fi
if ! command -v jq >/dev/null 2>&1; then
error "jq not found — required for JSON parsing"
exit 1
fi
detect_hosts
if [[ -z "${SONARR_URL:-}" ]] || [[ -z "${SONARR_API_KEY:-}" ]]; then
info "Sonarr not configured on $MY_ID ($LOCAL_SERVER_NAME) — skipping"
exit 0
fi
require_var SONARR_URL
require_var SONARR_API_KEY
if [[ "$SHOW_STATUS" == true ]]; then
echo ""
echo "━━━━━ $ICON_SUMMARY STATUS ━━━━━"
echo "$ICON_HOST Identity: $MY_ID ($LOCAL_SERVER_NAME)"
echo "$ICON_GEAR Sonarr URL: $SONARR_URL"
echo "$ICON_GEAR TV root: $SONARR_TV_ROOT"
echo "$ICON_GEAR General root: ${SONARR_GENERAL_ROOT:-<not configured>}"
echo "$ICON_GEAR Kids root: ${SONARR_KIDS_ROOT:-<not configured>}"
echo "$ICON_GEAR Anime root: ${SONARR_ANIME_ROOT:-<not configured>}"
echo "$ICON_GEAR Anime networks: ${#SONARR_ANIME_NETWORKS[@]} curated"
echo "$ICON_GEAR Kids networks: ${#SONARR_KIDS_NETWORKS[@]} curated"
echo "$ICON_GEAR Move mode: $MOVE_MODE"
echo "━━━━━━━━━━━━━━━━━━━━━━━"
exit 0
fi
echo ""
echo "━━━ $ICON_SYNC Fetching Sonarr Library ━━━"
if ! check_api "$SONARR_URL" "Sonarr" 10; then
exit 1
fi
check_arr_version "$SONARR_URL" "$SONARR_API_KEY" "v3" "$SONARR_VERSION_MAJOR" "Sonarr" || exit 1
SERIES_RESPONSE=$(arr_get_tracked_data "sonarr" "$SONARR_URL" "$SONARR_API_KEY" "v3") || {
error "Failed to fetch series from Sonarr"
exit 1
}
SERIES_COUNT=$(echo "$SERIES_RESPONSE" | jq -r 'length' 2>/dev/null)
if [[ -z "$SERIES_COUNT" ]] || [[ "$SERIES_COUNT" -eq 0 ]]; then
error "API returned 0 series — aborting"
exit 1
fi
info "$SERIES_COUNT series loaded"
# ==============================================================================================
# ━━━ Classify ━━━
# ==============================================================================================
echo ""
echo "━━━ $ICON_CLEAN Classifying ━━━"
ANIME_NETWORKS_JSON=$(printf '%s\n' "${SONARR_ANIME_NETWORKS[@]}" | jq -R . | jq -s .)
KIDS_NETWORKS_JSON=$(printf '%s\n' "${SONARR_KIDS_NETWORKS[@]}" | jq -R . | jq -s .)
# Not just the US TV-MA/TV-14 tiers — non-US certification scales use different labels for the
# same "clearly adult" tier (confirmed live: "Tomb Raider: The Legend of Lara Croft" is "16+").
ADULT_CERT_JSON='["TV-MA","TV-14","MA15+","16","16+","18","15","14"]'
RESULTS=$(echo "$SERIES_RESPONSE" | jq \
--argjson animeNetworks "$ANIME_NETWORKS_JSON" \
--argjson kidsNetworks "$KIDS_NETWORKS_JSON" \
--argjson adultCert "$ADULT_CERT_JSON" \
--arg animeRoot "${SONARR_ANIME_ROOT:-}" \
--arg kidsRoot "${SONARR_KIDS_ROOT:-}" '
def is_anime:
(any(.genres[]?; . == "Anime")
and (.originalLanguage.name == "Japanese" or (.network as $n | $animeNetworks | index($n) != null)))
or (any(.genres[]?; . == "Animation") and .originalLanguage.name == "Japanese")
or (.network as $n | $animeNetworks | index($n) != null);
def is_kids:
(is_anime | not) and (
any(.genres[]?; . == "Children")
or (.certification as $c | ["TV-Y","TV-Y7"] | index($c) != null)
or (.network as $n | $kidsNetworks | index($n) != null)
);
map(
{
title, id, network, certification, rootFolderPath,
episodeFileCount: (.statistics.episodeFileCount // 0),
is_anime: is_anime,
is_kids: is_kids
} |
. + {
forward_anime_miss: (.is_anime and $animeRoot != "" and .rootFolderPath != $animeRoot),
forward_kids_miss: (.is_kids and $kidsRoot != "" and .rootFolderPath != $kidsRoot),
reverse_anime_leak: ((.is_anime | not) and $animeRoot != "" and .rootFolderPath == $animeRoot),
reverse_kids_leak: ((.is_anime | not) and (.is_kids | not) and $kidsRoot != "" and .rootFolderPath == $kidsRoot
and (.certification as $c | $adultCert | index($c) != null))
}
)
')
FORWARD_ANIME_COUNT=$(echo "$RESULTS" | jq '[.[] | select(.forward_anime_miss)] | length')
FORWARD_KIDS_COUNT=$(echo "$RESULTS" | jq '[.[] | select(.forward_kids_miss)] | length')
REVERSE_ANIME_COUNT=$(echo "$RESULTS" | jq '[.[] | select(.reverse_anime_leak)] | length')
REVERSE_KIDS_COUNT=$(echo "$RESULTS" | jq '[.[] | select(.reverse_kids_leak)] | length')
if [[ "$ENABLE_LOGGING" == true ]]; then
echo "$RESULTS" | jq -r '.[] | select(.forward_anime_miss or .forward_kids_miss or .reverse_anime_leak or .reverse_kids_leak) |
" [\(if .forward_anime_miss then "FORWARD-ANIME" elif .forward_kids_miss then "FORWARD-KIDS" elif .reverse_anime_leak then "REVERSE-ANIME" elif .reverse_kids_leak then "REVERSE-KIDS" else "?" end)] \(.title) (root: \(.rootFolderPath), network: \(.network // "n/a"), cert: \(.certification // "n/a"))"'
fi
# ==============================================================================================
# ━━━ Summary ━━━
# ==============================================================================================
echo ""
echo "━━━━━ $ICON_SUMMARY SONARR CLASSIFICATION SUMMARY ━━━━━"
echo "$ICON_HOST Identity: $MY_ID ($LOCAL_SERVER_NAME)"
echo "$ICON_SYNC Series scanned: $SERIES_COUNT"
echo "$ICON_TRASH Forward — anime miss: $FORWARD_ANIME_COUNT (classified anime, outside ${SONARR_ANIME_ROOT:-<unconfigured>})"
echo "$ICON_TRASH Forward — kids miss: $FORWARD_KIDS_COUNT (classified kids, outside ${SONARR_KIDS_ROOT:-<unconfigured>})"
echo "$ICON_WARN Reverse — anime leak: $REVERSE_ANIME_COUNT (in ${SONARR_ANIME_ROOT:-<unconfigured>}, no anime signal — review, may be deliberate style placement)"
echo "$ICON_WARN Reverse — kids leak: $REVERSE_KIDS_COUNT (in ${SONARR_KIDS_ROOT:-<unconfigured>}, adult-rated content)"
echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
[[ "$ENABLE_LOGGING" != true ]] && echo " (run with --log for the per-title list)"
# ==============================================================================================
# ━━━ Move Mode ━━━
# ==============================================================================================
# See MOVE MODE in the header for scope (forward + reverse-kids-leak, not reverse-anime-leak).
if [[ "$MOVE_MODE" == true ]]; then
echo ""
echo "━━━ $ICON_SYNC Move Mode ━━━"
acquire_lock "wait"
trap "_release_all_locks" EXIT
build_arr_path_map "SONARR"
MOVE_TARGETS=$(echo "$RESULTS" | jq -c '[.[] | select(.forward_anime_miss or .forward_kids_miss or .reverse_kids_leak)]')
MOVE_COUNT=$(echo "$MOVE_TARGETS" | jq 'length')
if [[ "$MOVE_COUNT" -eq 0 ]]; then
info "Nothing to move"
exit 0
fi
warn "About to process $MOVE_COUNT series — one at a time, verifying after each"
MOVED=0
RELOCATED_SEARCH=0
FAILED=0
while IFS= read -r item; do
id=$(echo "$item" | jq -r '.id')
title=$(echo "$item" | jq -r '.title')
is_anime_flag=$(echo "$item" | jq -r '.is_anime')
is_forward_kids=$(echo "$item" | jq -r '.forward_kids_miss')
had_files_count=$(echo "$item" | jq -r '.episodeFileCount')
if [[ "$is_anime_flag" == "true" ]]; then
target_root="$SONARR_ANIME_ROOT"
elif [[ "$is_forward_kids" == "true" ]]; then
target_root="$SONARR_KIDS_ROOT"
else
target_root="$SONARR_GENERAL_ROOT"
fi
if [[ -z "$target_root" ]]; then
error " ✗ $title — target root not configured (SONARR_GENERAL_ROOT blank), skipping"
(( FAILED++ ))
continue
fi
# RESULTS only carries the reduced report fields — Sonarr's PUT expects the complete
# resource representation, so fetch a fresh full series record to modify and send back.
full_series=$(arr_api "$SONARR_URL" "$SONARR_API_KEY" "v3" "series/$id" "Sonarr")
if [[ -z "$full_series" ]]; then
error " ✗ $title — could not fetch full series record, skipping"
(( FAILED++ ))
continue
fi
old_path=$(echo "$full_series" | jq -r '.path')
folder_name="${old_path##*/}"
# A literal "/" in the folder name would build a broken nested directory instead of
# moving to one clean folder — this is exactly the self-inflicted bug hit doing the
# Fate/Zero and Fate/Stay Night moves by hand earlier this session.
if [[ "$folder_name" == *"/"* ]]; then
error " ✗ $title — folder name contains '/', skipping (needs manual handling)"
(( FAILED++ ))
continue
fi
new_path="${target_root}/${folder_name}"
if [[ "$had_files_count" -gt 0 ]]; then
info " → $title: $old_path$new_path (moving $had_files_count episode file(s))"
move_qs="?moveFiles=true"
else
info " → $title: $old_path$new_path (no files — relocating + search)"
move_qs=""
fi
updated_series=$(echo "$full_series" | jq --arg root "$target_root" --arg path "$new_path" \
'.rootFolderPath = $root | .path = $path')
http_code=$(curl -sf -o /dev/null -w "%{http_code}" -X PUT \
--max-time 30 \
-H "X-Api-Key: $SONARR_API_KEY" \
-H "Content-Type: application/json" \
-d "$updated_series" \
"${SONARR_URL}/api/v3/series/${id}${move_qs}" 2>/dev/null)
if [[ "$http_code" != "200" && "$http_code" != "202" ]]; then
error " ✗ $title — API returned HTTP $http_code — stopping (review before re-running)"
(( FAILED++ ))
break
fi
sleep 3
# Never trust the PUT response alone — re-fetch and confirm the change actually landed.
# This exact check is what caught the earlier race condition doing this by hand: two
# series reported "success" while episodeFileCount had silently dropped to 0.
verify_series=$(arr_api "$SONARR_URL" "$SONARR_API_KEY" "v3" "series/$id" "Sonarr")
verify_root=$(echo "$verify_series" | jq -r '.rootFolderPath')
verify_filecount=$(echo "$verify_series" | jq -r '.statistics.episodeFileCount // 0')
if [[ "$verify_root" != "$target_root" ]]; then
error " ✗ $title — verification failed (root: $verify_root) — stopping"
(( FAILED++ ))
break
fi
if [[ "$had_files_count" -gt 0 ]]; then
if [[ "$verify_filecount" -eq "$had_files_count" ]]; then
echo " $ICON_SUCCESS $title — moved and verified ($verify_filecount files)"
(( MOVED++ ))
else
error " ✗ $title — verification failed (root updated but episode count $verify_filecount != expected $had_files_count) — stopping"
(( FAILED++ ))
break
fi
else
search_code=$(curl -sf -o /dev/null -w "%{http_code}" -X POST \
--max-time 30 \
-H "X-Api-Key: $SONARR_API_KEY" \
-H "Content-Type: application/json" \
-d "{\"name\":\"SeriesSearch\",\"seriesId\":${id}}" \
"${SONARR_URL}/api/v3/command" 2>/dev/null)
if [[ "$search_code" == "200" || "$search_code" == "201" ]]; then
echo " $ICON_SUCCESS $title — relocated, search triggered"
else
warn " $title — relocated but search trigger returned HTTP $search_code (will pick up on next scheduled search)"
fi
(( RELOCATED_SEARCH++ ))
fi
done < <(echo "$MOVE_TARGETS" | jq -c '.[]')
# arr_get_tracked_data() is cache-first — every write above changed rootFolderPath, so the
# shared cache is now stale until the next scheduled arr_cache_prefill run. Refresh it now
# rather than leave that window open for every other script reading this cache.
if [[ "$(( MOVED + RELOCATED_SEARCH ))" -gt 0 ]]; then
info "Refreshing shared tracked-data cache..."
fresh_series=$(arr_api "$SONARR_URL" "$SONARR_API_KEY" "v3" "series" "Sonarr")
[[ -n "$fresh_series" ]] && arr_cache_write "sonarr" "$fresh_series"
fi
echo ""
echo "━━━━━ $ICON_SUMMARY MOVE SUMMARY ━━━━━"
echo "$ICON_SUCCESS Moved (files relocated): $MOVED"
echo "$ICON_SUCCESS Relocated + search triggered: $RELOCATED_SEARCH"
echo "$ICON_ERROR Failed: $FAILED"
echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
fi
exit 0