fix: Media Flow collector - use contains() instead of regex

OpenWrt jq is compiled without ONIGURUMA regex library, so test()
function doesn't work. Replace all regex patterns with contains()
for streaming service detection.

- Use ascii_downcase + contains() for pattern matching
- Define is_streaming, get_category, get_quality as jq functions
- Detects: YouTube, Netflix, Spotify, WhatsApp, Discord, Zoom, etc.
- Bump version to 0.6.2

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
This commit is contained in:
CyberMind-FR 2026-01-10 14:04:47 +01:00
parent 6de24bd4a2
commit 03552c55e9
2 changed files with 94 additions and 107 deletions

View File

@ -4,7 +4,7 @@
include $(TOPDIR)/rules.mk
PKG_NAME:=luci-app-media-flow
PKG_VERSION:=0.6.1
PKG_VERSION:=0.6.2
PKG_RELEASE:=1
PKG_ARCH:=all
PKG_LICENSE:=Apache-2.0

View File

@ -5,6 +5,7 @@
# Uses nDPId's local DPI detection (no cloud subscription required)
#
# Copyright (C) 2025 CyberMind.fr
# NOTE: Uses contains() instead of test() for jq without ONIGURUMA regex
HISTORY_FILE="/tmp/media-flow-history.json"
NDPID_FLOWS="/tmp/ndpid-flows.json"
@ -13,13 +14,6 @@ MEDIA_CACHE="/tmp/media-flow-ndpid-cache.json"
MAX_ENTRIES=1000
LOCK_FILE="/tmp/media-flow-ndpid-collector.lock"
# Streaming services patterns for nDPId applications
# These match nDPId's application protocol names
STREAMING_VIDEO="YouTube|Netflix|Disney|AmazonVideo|PrimeVideo|Twitch|HboMax|Hulu|Vimeo|Peacock|Paramount|Crunchyroll|DailyMotion|Vevo|Plex|AppleTV"
STREAMING_AUDIO="Spotify|AppleMusic|Deezer|SoundCloud|Tidal|Pandora|AmazonMusic|YouTubeMusic|iHeartRadio|Audible"
STREAMING_VISIO="Zoom|Teams|GoogleMeet|Discord|Skype|Webex|FaceTime|WhatsApp|Signal|Telegram|Slack|GoToMeeting"
STREAMING_ALL="${STREAMING_VIDEO}|${STREAMING_AUDIO}|${STREAMING_VISIO}"
# Check if already running
if [ -f "$LOCK_FILE" ]; then
pid=$(cat "$LOCK_FILE" 2>/dev/null)
@ -37,8 +31,7 @@ enabled=$(uci -q get media_flow.global.enabled 2>/dev/null || echo "1")
# Check if nDPId data is available
if [ ! -f "$NDPID_FLOWS" ]; then
# Fall back to checking if ndpid is running
if ! pgrep -x ndpid > /dev/null 2>&1; then
if ! pgrep ndpid > /dev/null 2>&1; then
exit 0
fi
fi
@ -46,71 +39,85 @@ fi
# Initialize history file
[ ! -f "$HISTORY_FILE" ] && echo '[]' > "$HISTORY_FILE"
# Function to categorize streaming service
categorize_service() {
local app="$1"
if echo "$app" | grep -qiE "$STREAMING_VIDEO"; then
echo "video"
elif echo "$app" | grep -qiE "$STREAMING_AUDIO"; then
echo "audio"
elif echo "$app" | grep -qiE "$STREAMING_VISIO"; then
echo "visio"
else
echo "other"
fi
}
# Function to estimate quality from bandwidth
estimate_quality() {
local kbps="$1"
local category="$2"
if [ "$category" = "audio" ]; then
# Audio quality tiers (kbps)
if [ "$kbps" -lt 96 ]; then
echo "Low"
elif [ "$kbps" -lt 192 ]; then
echo "Normal"
elif [ "$kbps" -lt 320 ]; then
echo "High"
else
echo "Lossless"
fi
elif [ "$category" = "visio" ]; then
# Video call quality tiers (kbps)
if [ "$kbps" -lt 500 ]; then
echo "Audio Only"
elif [ "$kbps" -lt 1500 ]; then
echo "SD"
elif [ "$kbps" -lt 3000 ]; then
echo "HD"
else
echo "FHD"
fi
else
# Video streaming quality tiers (kbps)
if [ "$kbps" -lt 1000 ]; then
echo "SD"
elif [ "$kbps" -lt 3000 ]; then
echo "HD"
elif [ "$kbps" -lt 8000 ]; then
echo "FHD"
else
echo "4K"
fi
fi
}
# Process nDPId flows
# Process nDPId flows using contains() instead of test() regex
# This works with jq compiled without ONIGURUMA
if [ -f "$NDPID_FLOWS" ] && command -v jq >/dev/null 2>&1; then
timestamp=$(date -Iseconds)
# Extract streaming flows from nDPId data
new_entries=$(jq -c --arg ts "$timestamp" --arg pattern "$STREAMING_ALL" '
# Extract streaming flows - using contains() for pattern matching
# Matches: YouTube, Netflix, Disney, Twitch, Spotify, WhatsApp, Zoom, Teams, etc.
new_entries=$(jq -c --arg ts "$timestamp" '
# Helper function to check if app is a streaming service
def is_streaming:
(. | ascii_downcase) as $app |
($app | contains("youtube")) or
($app | contains("netflix")) or
($app | contains("disney")) or
($app | contains("amazon")) or
($app | contains("prime")) or
($app | contains("twitch")) or
($app | contains("hbo")) or
($app | contains("hulu")) or
($app | contains("vimeo")) or
($app | contains("peacock")) or
($app | contains("paramount")) or
($app | contains("plex")) or
($app | contains("appletv")) or
($app | contains("spotify")) or
($app | contains("applemusic")) or
($app | contains("deezer")) or
($app | contains("soundcloud")) or
($app | contains("tidal")) or
($app | contains("pandora")) or
($app | contains("audible")) or
($app | contains("zoom")) or
($app | contains("teams")) or
($app | contains("meet")) or
($app | contains("discord")) or
($app | contains("skype")) or
($app | contains("webex")) or
($app | contains("facetime")) or
($app | contains("whatsapp")) or
($app | contains("signal")) or
($app | contains("telegram")) or
($app | contains("slack"));
# Helper function to get category
def get_category:
(. | ascii_downcase) as $app |
if ($app | contains("youtube")) or ($app | contains("netflix")) or ($app | contains("disney")) or
($app | contains("amazon")) or ($app | contains("twitch")) or ($app | contains("hbo")) or
($app | contains("hulu")) or ($app | contains("vimeo")) or ($app | contains("plex")) or
($app | contains("appletv")) or ($app | contains("paramount")) or ($app | contains("peacock"))
then "video"
elif ($app | contains("spotify")) or ($app | contains("applemusic")) or ($app | contains("deezer")) or
($app | contains("soundcloud")) or ($app | contains("tidal")) or ($app | contains("pandora")) or
($app | contains("audible"))
then "audio"
elif ($app | contains("zoom")) or ($app | contains("teams")) or ($app | contains("meet")) or
($app | contains("discord")) or ($app | contains("skype")) or ($app | contains("webex")) or
($app | contains("facetime")) or ($app | contains("whatsapp")) or ($app | contains("signal")) or
($app | contains("telegram")) or ($app | contains("slack"))
then "visio"
else "other"
end;
# Helper function to estimate quality from bandwidth (kbps)
def get_quality(cat):
if cat == "audio" then
if . < 96 then "Low" elif . < 192 then "Normal" elif . < 320 then "High" else "Lossless" end
elif cat == "visio" then
if . < 500 then "Audio" elif . < 1500 then "SD" elif . < 3000 then "HD" else "FHD" end
else
if . < 1000 then "SD" elif . < 3000 then "HD" elif . < 8000 then "FHD" else "4K" end
end;
[.[] |
select(.app != null and .app != "" and .app != "Unknown") |
select(.app | test($pattern; "i")) |
select(.state == "active" or .bytes_rx > 10000) |
select(.app | is_streaming) |
select(.state == "active" or .bytes_rx > 10000 or .bytes_tx > 10000) |
(.app | get_category) as $cat |
(((.bytes_rx // 0) + (.bytes_tx // 0)) * 8 / 1000 | floor) as $bw |
{
timestamp: $ts,
app: .app,
@ -124,63 +131,43 @@ if [ -f "$NDPID_FLOWS" ] && command -v jq >/dev/null 2>&1; then
confidence: (.confidence // "Unknown"),
ndpi_category: (.category // "Unknown"),
flow_id: (.id // 0),
state: (.state // "active")
state: (.state // "active"),
duration: 1,
bandwidth: $bw,
category: $cat,
quality: ($bw | get_quality($cat))
}
] |
# Calculate bandwidth and quality for each entry
map(. + {
duration: 1,
bandwidth: ((.bytes_rx + .bytes_tx) * 8 / 1000 | floor),
category: (
if (.app | test("YouTube|Netflix|Disney|Amazon|Twitch|Hbo|Hulu|Vimeo|Peacock|Paramount|Plex|AppleTV"; "i")) then "video"
elif (.app | test("Spotify|Apple.*Music|Deezer|SoundCloud|Tidal|Pandora|iHeart|Audible"; "i")) then "audio"
elif (.app | test("Zoom|Teams|Meet|Discord|Skype|Webex|Face.*Time|WhatsApp|Signal|Telegram|Slack"; "i")) then "visio"
else "other"
end
)
}) |
# Add quality estimation
map(. + {
quality: (
if .category == "audio" then
(if .bandwidth < 96 then "Low" elif .bandwidth < 192 then "Normal" elif .bandwidth < 320 then "High" else "Lossless" end)
elif .category == "visio" then
(if .bandwidth < 500 then "Audio Only" elif .bandwidth < 1500 then "SD" elif .bandwidth < 3000 then "HD" else "FHD" end)
else
(if .bandwidth < 1000 then "SD" elif .bandwidth < 3000 then "HD" elif .bandwidth < 8000 then "FHD" else "4K" end)
end
)
}) |
# Only include flows with significant traffic
[.[] | select(.bytes_rx > 10000 or .packets > 100)]
[.[] | select(.bytes_rx > 5000 or .bytes_tx > 5000 or .packets > 50)]
' "$NDPID_FLOWS" 2>/dev/null)
# Save current state to cache for frontend
if [ -n "$new_entries" ] && [ "$new_entries" != "[]" ] && [ "$new_entries" != "null" ]; then
echo "$new_entries" > "$MEDIA_CACHE"
# Merge with history (avoid duplicates by flow_id within same minute)
# Merge with history (avoid duplicates)
jq -c --argjson new "$new_entries" '
# Add new entries
. + ($new | map(del(.flow_id, .state))) |
# Remove duplicates (same client+app within 60 seconds)
unique_by(.client + .app + (.timestamp | split("T")[0])) |
# Keep only last MAX_ENTRIES
.[-'"$MAX_ENTRIES"':]
' "$HISTORY_FILE" > "${HISTORY_FILE}.tmp" 2>/dev/null && mv "${HISTORY_FILE}.tmp" "$HISTORY_FILE"
else
# No active streams, save empty cache
echo '[]' > "$MEDIA_CACHE"
fi
fi
# Also process nDPId apps file for aggregated stats
# Also process nDPId apps file for aggregated stats (without regex)
if [ -f "$NDPID_APPS" ] && command -v jq >/dev/null 2>&1; then
# Extract streaming apps from aggregated data
jq -c --arg pattern "$STREAMING_ALL" '
[.[] | select(.name | test($pattern; "i"))] |
sort_by(-.bytes) |
.[0:20]
jq -c '
def is_streaming:
(.name | ascii_downcase) as $app |
($app | contains("youtube")) or ($app | contains("netflix")) or
($app | contains("spotify")) or ($app | contains("whatsapp")) or
($app | contains("discord")) or ($app | contains("zoom")) or
($app | contains("teams")) or ($app | contains("twitch")) or
($app | contains("disney")) or ($app | contains("amazon"));
[.[] | select(is_streaming)] | sort_by(-.bytes) | .[0:20]
' "$NDPID_APPS" > "/tmp/media-flow-apps.json" 2>/dev/null
fi