Add channel checker, scheduled health checks, and a status dashboard

Builds on the idea in #1145 (a checker that maps failures back to a channel
and list, and tells a geo-blocked channel from a dead one) with a few
additions:

check_channels.py
- Standard library only, same as #1145's version.
- Adds a `disputed` state: an optional --confirm-dead pass gives ffprobe a
  second opinion on anything that looks dead over HTTP, before it gets
  reported as dead. This is one-directional (can only pull a verdict out
  of `dead`, never push one into it) because ffprobe itself is not
  reliable enough to trust in the other direction - a known-good DASH
  channel needed longer than any sane per-channel budget to open while
  testing this, and a header check alone had already been fooled by an
  isolated media fragment sitting at a URL that looked like a live channel
  (a mistake made and caught while triaging #1149/#1151 - see the
  docstring for the details).
- Restructured to share one worker pool across every list in a run
  instead of a fresh pool per list, which stopped small lists from paying
  the same wall-clock floor as large ones; a full run across all lists
  dropped from not finishing in 15 minutes to about 10.
- --json writes one record per channel per run, for building a history.

generate_dashboard.py
- Turns that history into a single self-contained docs/index.html: current
  state breakdown, an alive-share trend across every run kept, a per-list
  breakdown, and a searchable/filterable table of everything that is not
  currently alive.

Two new scheduled workflows
- check_channels_fast.yml: every 6 hours, HTTP checks only.
- check_channels_deep.yml: every 2 days, with --confirm-dead (needs ffmpeg).
Both append to .github/checker-history/history.jsonl (pruned to 90 days),
then build and deploy the dashboard to GitHub Pages.

.github/workflows/check_channels_deep.yml

@@ -0,0 +1,91 @@
+name: Channel check (deep)
+
+on:
+  schedule:
+    - cron: '0 6 */2 * *'  # every 2 days at 06:00 UTC
+  workflow_dispatch:
+
+permissions:
+  contents: write
+  pages: write
+  id-token: write
+
+concurrency:
+  group: pages
+  cancel-in-progress: false
+
+jobs:
+  check-deep:
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@v6
+        with:
+          fetch-depth: 0
+
+      - name: Install ffmpeg (for the ffprobe second opinion on dead channels)
+        run: sudo apt-get update && sudo apt-get install -y ffmpeg
+
+      - name: Run the deep check (HTTP probes + ffprobe confirmation on dead channels)
+        run: |
+          python3 check_channels.py --attempts 3 --timeout 12 --pause 5 --workers 40 --confirm-dead --json run.jsonl | tee check_summary.txt
+
+      - name: Append to history and prune entries older than 90 days
+        run: |
+          mkdir -p .github/checker-history
+          history=".github/checker-history/history.jsonl"
+          touch "$history"
+          cat run.jsonl >> "$history"
+          cutoff=$(date -u -d '90 days ago' +%Y-%m-%dT%H:%M:%SZ)
+          python3 - "$history" "$cutoff" <<'PYEOF'
+          import json, sys
+          path, cutoff = sys.argv[1], sys.argv[2]
+          kept = []
+          with open(path, encoding="utf-8") as f:
+              for line in f:
+                  line = line.strip()
+                  if line and json.loads(line)["checked_at"] >= cutoff:
+                      kept.append(line)
+          with open(path, "w", encoding="utf-8") as f:
+              f.write("\n".join(kept) + ("\n" if kept else ""))
+          PYEOF
+
+      - name: Commit history
+        run: |
+          git config user.name "ChannelCheckerBot" || true
+          git config user.email "channelcheckerbot@users.noreply.github.com" || true
+          git add .github/checker-history/history.jsonl
+          git diff --staged --quiet || git commit --quiet -m "Channel check (deep) - $(date -u +%Y-%m-%dT%H:%M:%SZ)"
+          git pull --rebase origin master || true
+          git diff --quiet HEAD @{u} || git push origin HEAD
+
+      - name: Build the status dashboard
+        run: python3 generate_dashboard.py --history .github/checker-history/history.jsonl --out docs/index.html
+
+      - name: Upload dashboard for Pages
+        uses: actions/upload-pages-artifact@v3
+        with:
+          path: docs
+
+      - name: Summarize
+        if: always()
+        run: |
+          {
+            echo "## Deep channel check - $(date -u +%Y-%m-%dT%H:%M:%SZ)"
+            echo ""
+            echo "Channels that looked \`dead\` over HTTP were given a second opinion by \`ffprobe\`. A channel still reported \`dead\` here survived both checks and is safe to act on; one reported \`disputed\` looked dead over HTTP but ffprobe found a real stream behind it - needs a human look, not an automatic edit."
+            echo ""
+            echo '```'
+            cat check_summary.txt
+            echo '```'
+          } >> "$GITHUB_STEP_SUMMARY"
+
+  deploy-pages:
+    needs: check-deep
+    runs-on: ubuntu-latest
+    environment:
+      name: github-pages
+      url: ${{ steps.deployment.outputs.page_url }}
+    steps:
+      - id: deployment
+        uses: actions/deploy-pages@v4

.github/workflows/check_channels_fast.yml

@@ -0,0 +1,88 @@
+name: Channel check (fast)
+
+on:
+  schedule:
+    - cron: '0 */6 * * *'  # every 6 hours
+  workflow_dispatch:
+
+permissions:
+  contents: write
+  pages: write
+  id-token: write
+
+concurrency:
+  group: pages
+  cancel-in-progress: false
+
+jobs:
+  check-fast:
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@v6
+        with:
+          fetch-depth: 0
+
+      - name: Run the fast check (HTTP probes only, no ffprobe)
+        run: |
+          python3 check_channels.py --attempts 3 --timeout 12 --pause 5 --workers 40 --json run.jsonl | tee check_summary.txt
+
+      - name: Append to history and prune entries older than 90 days
+        run: |
+          mkdir -p .github/checker-history
+          history=".github/checker-history/history.jsonl"
+          touch "$history"
+          cat run.jsonl >> "$history"
+          cutoff=$(date -u -d '90 days ago' +%Y-%m-%dT%H:%M:%SZ)
+          python3 - "$history" "$cutoff" <<'PYEOF'
+          import json, sys
+          path, cutoff = sys.argv[1], sys.argv[2]
+          kept = []
+          with open(path, encoding="utf-8") as f:
+              for line in f:
+                  line = line.strip()
+                  if line and json.loads(line)["checked_at"] >= cutoff:
+                      kept.append(line)
+          with open(path, "w", encoding="utf-8") as f:
+              f.write("\n".join(kept) + ("\n" if kept else ""))
+          PYEOF
+
+      - name: Commit history
+        run: |
+          git config user.name "ChannelCheckerBot" || true
+          git config user.email "channelcheckerbot@users.noreply.github.com" || true
+          git add .github/checker-history/history.jsonl
+          git diff --staged --quiet || git commit --quiet -m "Channel check (fast) - $(date -u +%Y-%m-%dT%H:%M:%SZ)"
+          git pull --rebase origin master || true
+          git diff --quiet HEAD @{u} || git push origin HEAD
+
+      - name: Build the status dashboard
+        run: python3 generate_dashboard.py --history .github/checker-history/history.jsonl --out docs/index.html
+
+      - name: Upload dashboard for Pages
+        uses: actions/upload-pages-artifact@v3
+        with:
+          path: docs
+
+      - name: Summarize
+        if: always()
+        run: |
+          {
+            echo "## Fast channel check - $(date -u +%Y-%m-%dT%H:%M:%SZ)"
+            echo ""
+            echo "Every channel probed 3x over HTTP; no ffprobe confirmation (see the *Channel check (deep)* workflow for that). \`blocked\` means geo-restricted, not broken - see \`check_channels.py\` for what each state means."
+            echo ""
+            echo '```'
+            cat check_summary.txt
+            echo '```'
+          } >> "$GITHUB_STEP_SUMMARY"
+
+  deploy-pages:
+    needs: check-fast
+    runs-on: ubuntu-latest
+    environment:
+      name: github-pages
+      url: ${{ steps.deployment.outputs.page_url }}
+    steps:
+      - id: deployment
+        uses: actions/deploy-pages@v4

.gitignore

@@ -1,2 +1,4 @@
 __pycache__/
 *.pyc
+# generated by generate_dashboard.py; deployed straight to Pages, never committed
+docs/

check_channels.py

@@ -0,0 +1,326 @@
+#!/usr/bin/python3
+"""Report which channels in `lists/` are no longer reachable.
+
+The playlist checker records the URLs it could not open, but a bare URL does not say
+which channel or which list it came from, and one failed attempt does not mean a
+channel is gone. Streams time out, rate limit, move between CDNs, and some are only
+served inside their own country.
+
+This script maps every URL back to its channel and list, probes each one several
+times, and separates the cases that look alike from the outside:
+
+  dead         every attempt said "not found", so the stream is really gone
+  disputed     looked dead over HTTP, but ffprobe found a real audio/video stream
+               behind it - needs a human, not a bot, to decide
+  blocked      the server answered but refused us, which is what a channel served
+               only in its own country looks like from anywhere else
+  unreachable  every attempt timed out or failed to connect
+  flaky        answered at least once, so it is up but unreliable from here
+  alive        answered every time
+
+Only `dead` is safe to act on without a second opinion. `blocked` in particular must
+not be treated as a fault: the lists mark geo-blocked channels deliberately.
+
+Both playlist formats used by the lists are understood, HLS (.m3u8) and MPEG-DASH
+(.mpd), so a DASH stream is not mistaken for a broken one. Neither check is airtight
+on its own: a manifest that starts with `#EXTM3U`/`<MPD` is not proof the media
+behind it plays, and a payload ffprobe can decode is not proof it is a live channel
+rather than a stray media fragment sitting at a familiar-looking URL - one such
+fragment is what first prompted the --confirm-dead option below. Treat `dead` as
+"nothing here answered like a stream, twice, two different ways" rather than proof.
+
+With --confirm-dead, every channel that looks `dead` over HTTP gets a second,
+independent opinion from `ffprobe` (part of ffmpeg) before being reported as dead:
+ffprobe actually tries to decode audio/video from the URL, which catches streams a
+header check alone cannot judge either way. This second opinion is one-directional -
+it can only pull a channel *out* of `dead` into `disputed`, never push a channel that
+looks fine over HTTP into a worse state - because a slow or unusual server can make
+ffprobe time out on a channel that plays fine elsewhere (this happened while writing
+this script: a known-good DASH channel needed longer than any reasonable per-channel
+budget for ffprobe to open), and a timeout there must not be read as confirmation of
+anything. It is off by default because it materially changes the runtime: ffprobe
+does real network I/O per candidate, on top of the HTTP probes already spent finding
+that candidate, and only makes sense where that time is available (a weekly run),
+not where it is not (a PR check blocking on a handful of changed links).
+
+Usage:
+    ./check_channels.py                          # every list
+    ./check_channels.py greece italy             # only those lists
+    ./check_channels.py --attempts 5 greece      # more attempts per channel
+    ./check_channels.py --confirm-dead greece    # + ffprobe second opinion on dead ones
+"""
+
+import argparse
+import json
+import os
+import re
+import shutil
+import ssl
+import subprocess
+import sys
+import time
+import urllib.error
+import urllib.request
+from collections import namedtuple
+from concurrent.futures import ThreadPoolExecutor
+from datetime import datetime, timezone
+
+LISTS_DIR = os.path.join(os.path.dirname(os.path.abspath(__file__)), "lists")
+
+# one bundle for the knobs every probe/list/run function otherwise had to repeat
+ProbeSettings = namedtuple(
+    "ProbeSettings", ["attempts", "timeout", "pause", "workers", "confirm_dead"],
+)
+
+# `| 1 | Channel name | [>](url) | ...` is the row format used by every list
+ROW = re.compile(r"^\|[^|]*\|([^|]+)\|[^|]*\[>\]\((https?://[^)\s]+)\)")
+
+USER_AGENT = (
+    "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 "
+    "(KHTML, like Gecko) Chrome/131.0 Safari/537.36"
+)
+
+OK = "ok"
+GONE = "gone"
+REFUSED = "refused"
+UNREACHABLE = "unreachable"
+
+ALIVE = "alive"
+DEAD = "dead"
+DISPUTED = "disputed"
+BLOCKED = "blocked"
+UNREACHED = "unreachable"
+FLAKY = "flaky"
+
+# a channel served only in its own country answers, then refuses us
+REFUSING_CODES = (401, 402, 403, 451)
+GONE_CODES = (404, 410)
+
+# ffprobe gets much longer than an HTTP probe: it has to open a connection, read
+# enough of the stream to find a decodable frame, and do that over whatever the
+# channel's own CDN feels like doing today, not just get a response header back
+FFPROBE_TIMEOUT = 25
+
+
+def read_channels(name):
+    """Return the `(channel, url)` pairs of the list called `name`."""
+    path = os.path.join(LISTS_DIR, name + ".md")
+    channels = []
+    with open(path, encoding="utf-8") as handle:
+        for line in handle:
+            match = ROW.match(line.strip())
+            if match:
+                channels.append((match.group(1).strip(), match.group(2)))
+    return channels
+
+
+def looks_like_a_playlist(head):
+    """Return True if `head` is the start of an HLS or a DASH playlist."""
+    text = head.lstrip()
+    if text.startswith("#EXTM3U"):
+        return True
+    # DASH manifests are XML, and may carry a declaration, a comment or neither
+    return "<MPD" in text[:2000]
+
+
+def probe(url, timeout):
+    """Open `url` once and report what the server did."""
+    request = urllib.request.Request(url, headers={"User-Agent": USER_AGENT})
+    # a stream whose certificate does not verify is still a working stream
+    context = ssl.create_default_context()
+    context.check_hostname = False
+    context.verify_mode = ssl.CERT_NONE
+    try:
+        with urllib.request.urlopen(request, timeout=timeout, context=context) as response:
+            head = response.read(3000).decode("utf-8", "ignore")
+    except urllib.error.HTTPError as error:
+        if error.code in GONE_CODES:
+            return GONE
+        if error.code in REFUSING_CODES:
+            return REFUSED
+        return UNREACHABLE
+    except (urllib.error.URLError, OSError, ValueError):
+        return UNREACHABLE
+    return OK if looks_like_a_playlist(head) else GONE
+
+
+def ffprobe_finds_a_stream(url):
+    """Ask ffprobe whether it can decode real audio/video from `url`.
+
+    Returns True/False for a definite answer, or None if ffprobe itself could not
+    be asked (missing binary, timed out, or errored) - None must never be treated
+    as "no stream found", only as "no second opinion available this time".
+    """
+    try:
+        result = subprocess.run(
+            [
+                "ffprobe", "-v", "error",
+                "-user_agent", USER_AGENT,
+                "-show_entries", "stream=codec_type",
+                "-of", "csv=p=0",
+                url,
+            ],
+            capture_output=True, text=True, timeout=FFPROBE_TIMEOUT, check=False,
+        )
+    except subprocess.TimeoutExpired:
+        return None
+    except OSError:
+        return None
+    if result.returncode != 0:
+        return False
+    return any(kind in result.stdout for kind in ("video", "audio"))
+
+
+def classify(outcomes):
+    """Turn the outcomes of the attempts on one channel into a single state."""
+    if all(outcome == OK for outcome in outcomes):
+        return ALIVE
+    if any(outcome == OK for outcome in outcomes):
+        return FLAKY
+    if any(outcome == REFUSED for outcome in outcomes):
+        return BLOCKED
+    if all(outcome == GONE for outcome in outcomes):
+        return DEAD
+    return UNREACHED
+
+
+def check(channel, settings):
+    """Probe one channel `settings.attempts` times and return its state.
+
+    If the channel looks dead and `settings.confirm_dead` is set, give it one more
+    chance through ffprobe before reporting it as dead - see the module docstring
+    for why this can only pull a verdict out of `dead`, never push one into it.
+    """
+    name, url = channel
+    outcomes = []
+    for attempt in range(settings.attempts):
+        outcomes.append(probe(url, settings.timeout))
+        if attempt + 1 < settings.attempts:
+            time.sleep(settings.pause)
+    state = classify(outcomes)
+    if state == DEAD and settings.confirm_dead:
+        if ffprobe_finds_a_stream(url):
+            state = DISPUTED
+    return name, url, state, outcomes
+
+
+def check_all(names, settings):
+    """Probe every channel of every named list and return `{list_name: [results]}`.
+
+    All channels of all lists share one pool of `settings.workers`, so a run across
+    many lists is not slower per list than a run of one - a list with two channels
+    does not pay the same wall-clock floor as a list with two hundred just because
+    it was handed its own pool that then sits mostly idle.
+    """
+    channels_by_list = {
+        name: [c for c in read_channels(name) if not c[1].startswith("https://www.youtube.com")]
+        for name in names
+    }
+
+    def run(item):
+        list_name, channel = item
+        return list_name, check(channel, settings)
+
+    jobs = [(name, channel) for name, channels in channels_by_list.items() for channel in channels]
+    results_by_list = {name: [] for name in names}
+    with ThreadPoolExecutor(max_workers=settings.workers) as pool:
+        for list_name, result in pool.map(run, jobs):
+            results_by_list[list_name].append(result)
+    return results_by_list
+
+
+def report(name, results, attempts):
+    """Print the results of one list and return how many channels are dead."""
+    states = {}
+    for _, _, state, _ in results:
+        states[state] = states.get(state, 0) + 1
+    summary = ", ".join(f"{states[s]} {s}" for s in sorted(states))
+    print(f"{name}: {len(results)} checked, {summary}")
+    for channel, url, state, outcomes in results:
+        if state == ALIVE:
+            continue
+        detail = "/".join(outcomes) if state != FLAKY else f"{outcomes.count(OK)}/{attempts} ok"
+        print(f"  {state:12} {channel} [{detail}] -> {url}")
+    return states.get(DEAD, 0)
+
+
+def as_records(list_name, results, checked_at, confirm_dead):
+    """Turn one list's results into flat dicts, one per channel, for --json output.
+
+    `confirm_dead` is recorded on every row so a later reader can tell a `dead`
+    verdict that already survived an ffprobe second opinion from one that has not
+    been asked yet - the two are not the same strength of evidence.
+    """
+    return [
+        {
+            "checked_at": checked_at,
+            "list": list_name,
+            "channel": channel,
+            "url": url,
+            "state": state,
+            "outcomes": outcomes,
+            "confirm_dead": confirm_dead,
+        }
+        for channel, url, state, outcomes in results
+    ]
+
+
+def parse_args():
+    """Parse the command line into an `argparse.Namespace`."""
+    parser = argparse.ArgumentParser(
+        description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter,
+    )
+    parser.add_argument("lists", nargs="*", help="lists to check, without the .md suffix")
+    parser.add_argument("--attempts", type=int, default=3, help="probes per channel")
+    parser.add_argument("--timeout", type=int, default=12, help="seconds per probe")
+    parser.add_argument("--pause", type=int, default=5, help="seconds between probes")
+    parser.add_argument("--workers", type=int, default=8, help="channels probed at once")
+    parser.add_argument(
+        "--confirm-dead", action="store_true",
+        help="give ffprobe a second opinion on channels that look dead (slower; needs ffmpeg)",
+    )
+    parser.add_argument(
+        "--json", metavar="PATH",
+        help="also write every channel's result as one JSON record per line to PATH, for "
+             "building a history across runs (state, outcomes, timestamp, list, channel, url)",
+    )
+    return parser.parse_args()
+
+
+def write_json_records(handle, names, results_by_list, checked_at, confirm_dead):
+    """Write every list's results to `handle` as one JSON record per channel per line."""
+    for name in names:
+        for record in as_records(name, results_by_list[name], checked_at, confirm_dead):
+            handle.write(json.dumps(record, ensure_ascii=False) + "\n")
+
+
+def main():
+    """Check the lists named on the command line, or every list."""
+    args = parse_args()
+
+    if args.confirm_dead and shutil.which("ffprobe") is None:
+        print(
+            "--confirm-dead needs ffprobe (part of ffmpeg) on PATH; proceeding without it.",
+            file=sys.stderr,
+        )
+        args.confirm_dead = False
+
+    all_lists = (f[:-3] for f in os.listdir(LISTS_DIR) if f.endswith(".md"))
+    names = args.lists or sorted(all_lists)
+    checked_at = datetime.now(timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")
+    settings = ProbeSettings(
+        args.attempts, args.timeout, args.pause, args.workers, args.confirm_dead,
+    )
+
+    results_by_list = check_all(names, settings)
+
+    dead_total = sum(report(name, results_by_list[name], args.attempts) for name in names)
+    if args.json:
+        with open(args.json, "w", encoding="utf-8") as handle:
+            write_json_records(handle, names, results_by_list, checked_at, args.confirm_dead)
+
+    return 1 if dead_total else 0
+
+
+if __name__ == "__main__":
+    sys.exit(main())

generate_dashboard.py

@@ -0,0 +1,503 @@
+#!/usr/bin/python3
+"""Turn `.github/checker-history/history.jsonl` into a static status dashboard.
+
+Reads the JSON-lines history that `check_channels.py --json` appends to (one row
+per channel per run), and writes a single self-contained `docs/index.html`: no
+build step, no client-side fetch of the raw history - everything needed to render
+is computed here and baked into the page, so it opens instantly whether it is
+opened from disk or served by GitHub Pages.
+
+The page shows:
+  - today's state breakdown (how many channels are alive/dead/blocked/... right now)
+  - a trend of the alive share across every run kept in the history
+  - one row per list (country) with its current breakdown
+  - every channel that is not currently `alive`, searchable and filterable by state
+
+"Currently" means: for each (list, channel), the most recent row in the history -
+older rows for the same channel are trend data, not part of today's snapshot.
+
+Usage:
+    ./generate_dashboard.py                          # reads the default history path
+    ./generate_dashboard.py --history path/to.jsonl --out path/to.html
+"""
+
+import argparse
+import json
+import os
+from collections import defaultdict
+from datetime import datetime, timezone
+from html import escape
+
+BASE_DIR = os.path.dirname(os.path.abspath(__file__))
+DEFAULT_HISTORY = os.path.join(BASE_DIR, ".github", "checker-history", "history.jsonl")
+DEFAULT_OUT = os.path.join(BASE_DIR, "docs", "index.html")
+
+# rendering order and the CSS custom property carrying each state's color
+STATE_ORDER = ["alive", "flaky", "blocked", "unreachable", "disputed", "dead"]
+STATE_VAR = {s: f"var(--s-{s})" for s in STATE_ORDER}
+STATE_LABEL = {
+    "alive": "answers every time",
+    "flaky": "answers, but not every time",
+    "blocked": "refused - likely geo-restricted",
+    "unreachable": "times out or refuses to connect",
+    "disputed": "looked dead, ffprobe found a stream - needs a human",
+    "dead": "gone on every attempt, ffprobe agrees where asked",
+}
+
+
+def read_history(path):
+    """Return every row of the history file, oldest first, skipping malformed lines."""
+    rows = []
+    if not os.path.exists(path):
+        return rows
+    with open(path, encoding="utf-8") as handle:
+        for line in handle:
+            line = line.strip()
+            if not line:
+                continue
+            try:
+                rows.append(json.loads(line))
+            except json.JSONDecodeError:
+                continue
+    rows.sort(key=lambda r: r["checked_at"])
+    return rows
+
+
+def latest_snapshot(rows):
+    """Return the most recent row for each (list, channel, url) tuple."""
+    latest = {}
+    for row in rows:
+        key = (row["list"], row["channel"], row["url"])
+        latest[key] = row
+    return list(latest.values())
+
+
+def run_trend(rows):
+    """Group rows by `checked_at` (one run) and return the alive share of each run."""
+    by_run = defaultdict(lambda: defaultdict(int))
+    for row in rows:
+        by_run[row["checked_at"]][row["state"]] += 1
+    points = []
+    for checked_at in sorted(by_run):
+        counts = by_run[checked_at]
+        total = sum(counts.values())
+        alive_share = counts.get("alive", 0) / total if total else 0
+        points.append((checked_at, alive_share, total))
+    return points
+
+
+def _scope_grid_and_labels(pad_l, pad_r, pad_t, width, plot_h):
+    """Return the horizontal gridlines and their %-axis labels for `svg_trend`."""
+    grid = "".join(
+        f'<line x1="{pad_l}" y1="{pad_t + plot_h * f:.1f}" '
+        f'x2="{width - pad_r}" y2="{pad_t + plot_h * f:.1f}" class="scope-grid"/>'
+        for f in (0, 0.25, 0.5, 0.75, 1)
+    )
+    labels = "".join(
+        f'<text x="{pad_l - 8}" y="{pad_t + plot_h * f + 4:.1f}" class="scope-axis" '
+        f'text-anchor="end">{int((1 - f) * 100)}%</text>'
+        for f in (0, 0.5, 1)
+    )
+    return grid, labels
+
+
+def _scope_geometry(points, pad_l, pad_t, plot_w, plot_h):
+    """Return `(line_path, area_path, last_x, last_y)` for the trend's data points."""
+    xs = [pad_l + i * plot_w / (len(points) - 1) for i in range(len(points))]
+    ys = [pad_t + (1 - share) * plot_h for _, share, _ in points]
+    steps = (f"{'M' if i == 0 else 'L'}{x:.1f},{y:.1f}" for i, (x, y) in enumerate(zip(xs, ys)))
+    path = " ".join(steps)
+    area = path + f" L{xs[-1]:.1f},{pad_t + plot_h:.1f} L{xs[0]:.1f},{pad_t + plot_h:.1f} Z"
+    return path, area, xs[-1], ys[-1]
+
+
+def _scope_endpoint_labels(points):
+    """Return `(first_date, last_date, last_alive_pct)` for the trend's endpoints."""
+    return points[0][0][:10], points[-1][0][:10], f"{points[-1][1] * 100:.1f}%"
+
+
+def svg_trend(points, width=860, height=180):
+    """Render `points` as an oscilloscope-style trace of the alive share over time.
+
+    Kept as one function despite the local-variable count: it is a single visual
+    composite (grid, axis labels, line, fill, endpoint dot) whose pieces only make
+    sense read together, and splitting it further would trade a readable layout
+    calculation for indirection between fragments that all belong to one <svg>.
+    """
+    # pylint: disable=too-many-locals
+    pad_l, pad_r, pad_t = 34, 16, 16
+    plot_w, plot_h = width - pad_l - pad_r, height - pad_t - 26
+    grid, labels = _scope_grid_and_labels(pad_l, pad_r, pad_t, width, plot_h)
+
+    if len(points) < 2:
+        return (
+            f'<svg viewBox="0 0 {width} {height}" class="scope" role="img" '
+            f'aria-label="Not enough runs yet for a trend">{grid}{labels}'
+            f'<text x="{width / 2}" y="{height / 2}" text-anchor="middle" class="scope-empty">'
+            "collecting data - check back after a few runs</text></svg>"
+        )
+
+    path, area, last_x, last_y = _scope_geometry(points, pad_l, pad_t, plot_w, plot_h)
+    first_label, last_label, last_pct = _scope_endpoint_labels(points)
+    return (
+        f'<svg viewBox="0 0 {width} {height}" class="scope" role="img" aria-label='
+        f'"Alive share over time, {first_label} to {last_label}, currently {last_pct}">'
+        f"{grid}{labels}"
+        f'<path d="{area}" class="scope-fill"/><path d="{path}" class="scope-line"/>'
+        f'<circle cx="{last_x:.1f}" cy="{last_y:.1f}" r="3.5" class="scope-dot"/>'
+        f'<text x="{pad_l}" y="{height - 6}" class="scope-axis">{escape(first_label)}</text>'
+        f'<text x="{width - pad_r}" y="{height - 6}" class="scope-axis" text-anchor="end">'
+        f"{escape(last_label)}</text></svg>"
+    )
+
+
+def svg_signal_bar(counts, width=280, height=14):
+    """Render a `state -> count` mapping as one horizontal signal-strength bar."""
+    total = sum(counts.values()) or 1
+    x = 0
+    segments = []
+    for state in STATE_ORDER:
+        count = counts.get(state, 0)
+        if not count:
+            continue
+        w = count / total * width
+        segments.append(
+            f'<rect x="{x:.1f}" y="0" width="{w:.1f}" height="{height}" fill="{STATE_VAR[state]}">'
+            f"<title>{state}: {count} ({count / total * 100:.1f}%)</title></rect>"
+        )
+        x += w
+    return (
+        f'<svg viewBox="0 0 {width} {height}" class="signal-bar" role="img" '
+        f'aria-label="State breakdown">{"".join(segments)}</svg>'
+    )
+
+
+def legend_html():
+    """Return the small state-color legend shown above the by-list table."""
+    chips = "".join(
+        f'<span class="chip"><span class="dot" style="background:{STATE_VAR[s]}"></span>{s}'
+        f'<span class="chip-note">{STATE_LABEL[s]}</span></span>'
+        for s in STATE_ORDER
+    )
+    return f'<div class="legend">{chips}</div>'
+
+
+# pylint: disable=line-too-long
+EMPTY_PAGE = """<!doctype html>
+<html lang="en"><head><meta charset="utf-8"><title>Channel Signal</title>
+<style>body{{font-family:system-ui;padding:3rem;background:#0f1417;color:#e7edf0}}</style></head>
+<body><h1>Channel Signal</h1><p>No history yet at {path} - the checker workflows populate this after their first run.</p></body></html>"""
+# pylint: enable=line-too-long
+
+
+def build_cards(overall_counts):
+    """Return the row of big per-state count cards at the top of the page."""
+    return "".join(
+        f'<div class="card"><div class="card-num" style="color:{STATE_VAR[s]}">'
+        f'{overall_counts.get(s, 0)}</div><div class="card-label">{s}</div></div>'
+        for s in STATE_ORDER
+    )
+
+
+def build_list_rows(by_list):
+    """Return one `<tr>` per list (country), each with its own signal bar."""
+    rows = []
+    for name in sorted(by_list):
+        entries = by_list[name]
+        counts = defaultdict(int)
+        for entry in entries:
+            counts[entry["state"]] += 1
+        alive_pct = counts.get("alive", 0) / len(entries) * 100
+        rows.append(
+            f"<tr><td class='list-name'>{escape(name)}</td><td class='num'>{len(entries)}</td>"
+            f'<td class="num" style="color:{STATE_VAR["alive"]}">{alive_pct:.0f}%</td>'
+            f"<td>{svg_signal_bar(counts)}</td></tr>"
+        )
+    return "".join(rows)
+
+
+def build_problem_rows(current):
+    """Return one `<tr>` per channel that is not currently `alive`."""
+    ordered = sorted(current, key=lambda r: (r["state"] != "dead", r["list"], r["channel"]))
+    rows = []
+    for row in ordered:
+        if row["state"] == "alive":
+            continue
+        rows.append(
+            f'<tr data-state="{escape(row["state"])}">'
+            f'<td>{escape(row["list"])}</td>'
+            f'<td>{escape(row["channel"])}</td>'
+            f'<td><span class="pill" style="background:{STATE_VAR[row["state"]]}">'
+            f'{escape(row["state"])}</span></td>'
+            f'<td class="url"><a href="{escape(row["url"])}">{escape(row["url"])}</a></td>'
+            f'<td class="dim">{escape(row["checked_at"][:10])}</td>'
+            "</tr>"
+        )
+    return "".join(rows)
+
+
+def render(history_path, generated_at):
+    """Build the full dashboard page for the history found at `history_path`."""
+    rows = read_history(history_path)
+    if not rows:
+        return EMPTY_PAGE.format(path=escape(history_path))
+
+    current = latest_snapshot(rows)
+    overall_counts = defaultdict(int)
+    for row in current:
+        overall_counts[row["state"]] += 1
+    total_channels = len(current)
+
+    by_list = defaultdict(list)
+    for row in current:
+        by_list[row["list"]].append(row)
+
+    trend = run_trend(rows)
+    last_checked = rows[-1]["checked_at"]
+    alive_pct_overall = (
+        overall_counts.get("alive", 0) / total_channels * 100 if total_channels else 0
+    )
+
+    cards = build_cards(overall_counts)
+    list_rows = build_list_rows(by_list)
+    problem_rows = build_problem_rows(current)
+    state_options = "".join(
+        f'<option value="{s}">{s}</option>' for s in STATE_ORDER if s != "alive"
+    )
+
+    # pylint: disable=line-too-long
+    return f"""<!doctype html>
+<html lang="en">
+<head>
+<meta charset="utf-8">
+<meta name="viewport" content="width=device-width, initial-scale=1">
+<title>Channel Signal</title>
+<link rel="stylesheet" href="https://fonts.googleapis.com/css2?family=Archivo:wght@600;700&family=IBM+Plex+Sans:wght@400;500;600&family=IBM+Plex+Mono:wght@400;500&display=swap">
+<style>
+:root {{
+  --bg: #f5f3ef;
+  --surface: #ffffff;
+  --surface-2: #edeae4;
+  --text: #191d1f;
+  --text-dim: #5c6a70;
+  --border: #dfdad2;
+  --accent: #0f766e;
+  --accent-soft: #d6f3ef;
+  --s-alive: #157a3d;
+  --s-flaky: #b3620a;
+  --s-blocked: #1d5fd6;
+  --s-unreachable: #737f89;
+  --s-disputed: #a324ab;
+  --s-dead: #c22222;
+  --shadow: 0 1px 2px rgba(30,25,15,.06), 0 6px 20px -8px rgba(30,25,15,.12);
+}}
+@media (prefers-color-scheme: dark) {{
+  :root:not([data-theme="light"]) {{
+    --bg: #10181a;
+    --surface: #16211f;
+    --surface-2: #1b2725;
+    --text: #eaf2ee;
+    --text-dim: #8fa39d;
+    --border: #253634;
+    --accent: #2dd4bf;
+    --accent-soft: #123934;
+    --s-alive: #4ade80;
+    --s-flaky: #fbbf24;
+    --s-blocked: #60a5fa;
+    --s-unreachable: #93a4ab;
+    --s-disputed: #e879f9;
+    --s-dead: #f87171;
+    --shadow: 0 1px 2px rgba(0,0,0,.3), 0 8px 24px -8px rgba(0,0,0,.5);
+  }}
+}}
+:root[data-theme="dark"] {{
+  --bg: #10181a;
+  --surface: #16211f;
+  --surface-2: #1b2725;
+  --text: #eaf2ee;
+  --text-dim: #8fa39d;
+  --border: #253634;
+  --accent: #2dd4bf;
+  --accent-soft: #123934;
+  --s-alive: #4ade80;
+  --s-flaky: #fbbf24;
+  --s-blocked: #60a5fa;
+  --s-unreachable: #93a4ab;
+  --s-disputed: #e879f9;
+  --s-dead: #f87171;
+  --shadow: 0 1px 2px rgba(0,0,0,.3), 0 8px 24px -8px rgba(0,0,0,.5);
+}}
+
+* {{ box-sizing: border-box; }}
+body {{
+  margin: 0; background: var(--bg); color: var(--text);
+  font-family: "IBM Plex Sans", system-ui, sans-serif;
+  padding: 2.5rem 1.25rem 5rem;
+}}
+main {{ max-width: 980px; margin: 0 auto; }}
+
+.masthead {{ display: flex; justify-content: space-between; align-items: baseline; flex-wrap: wrap; gap: .5rem 1.5rem; margin-bottom: .35rem; }}
+h1 {{
+  font-family: "Archivo", system-ui, sans-serif; font-weight: 700; font-size: 1.55rem;
+  letter-spacing: -.01em; margin: 0; text-wrap: balance;
+}}
+h1 .on-air {{ color: var(--accent); font-variant-numeric: tabular-nums; }}
+.subline {{ color: var(--text-dim); font-size: .82rem; margin: 0 0 1.75rem; font-family: "IBM Plex Mono", monospace; }}
+
+.cards {{ display: grid; grid-template-columns: repeat(6, 1fr); gap: .6rem; margin-bottom: 1.75rem; }}
+.card {{
+  background: var(--surface); border: 1px solid var(--border); border-radius: 8px;
+  padding: .85rem .5rem; text-align: center; box-shadow: var(--shadow);
+}}
+.card-num {{ font-family: "IBM Plex Mono", monospace; font-size: 1.7rem; font-weight: 500; font-variant-numeric: tabular-nums; line-height: 1; }}
+.card-label {{ font-size: .72rem; color: var(--text-dim); margin-top: .35rem; text-transform: uppercase; letter-spacing: .06em; }}
+
+section {{
+  background: var(--surface); border: 1px solid var(--border); border-radius: 10px;
+  padding: 1.4rem 1.6rem; margin-bottom: 1.5rem; box-shadow: var(--shadow);
+}}
+section h2 {{
+  font-family: "Archivo", system-ui, sans-serif; font-weight: 700; font-size: 1rem;
+  margin: 0 0 .9rem; letter-spacing: -.005em;
+}}
+.section-note {{ color: var(--text-dim); font-size: .82rem; margin: -.5rem 0 1rem; }}
+.section-note code {{ font-family: "IBM Plex Mono", monospace; font-size: .8em; background: var(--surface-2); padding: .05rem .3rem; border-radius: 4px; }}
+
+.scope {{ width: 100%; height: auto; display: block; }}
+.scope-grid {{ stroke: var(--border); stroke-width: 1; }}
+.scope-axis {{ font: 10.5px "IBM Plex Mono", monospace; fill: var(--text-dim); }}
+.scope-empty {{ font: 12px "IBM Plex Mono", monospace; fill: var(--text-dim); }}
+.scope-fill {{ fill: var(--accent-soft); }}
+.scope-line {{ fill: none; stroke: var(--accent); stroke-width: 2; }}
+.scope-dot {{ fill: var(--accent); }}
+
+.legend {{ display: flex; gap: 1.1rem; flex-wrap: wrap; margin-bottom: 1rem; font-size: .78rem; }}
+.chip {{ display: inline-flex; align-items: baseline; gap: .4rem; font-family: "IBM Plex Mono", monospace; color: var(--text); }}
+.chip-note {{ color: var(--text-dim); font-family: "IBM Plex Sans", sans-serif; }}
+.dot {{ width: 8px; height: 8px; border-radius: 50%; display: inline-block; align-self: center; }}
+
+table {{ width: 100%; border-collapse: collapse; font-size: .87rem; }}
+th {{
+  text-align: left; padding: .4rem .55rem; border-bottom: 2px solid var(--border);
+  font-size: .72rem; text-transform: uppercase; letter-spacing: .05em; color: var(--text-dim); font-weight: 600;
+}}
+td {{ padding: .45rem .55rem; border-bottom: 1px solid var(--border); vertical-align: middle; }}
+tr:last-child td {{ border-bottom: none; }}
+td.num {{ text-align: right; font-family: "IBM Plex Mono", monospace; font-variant-numeric: tabular-nums; }}
+td.list-name {{ font-weight: 500; }}
+td.url {{ max-width: 320px; overflow: hidden; text-overflow: ellipsis; white-space: nowrap; font-family: "IBM Plex Mono", monospace; font-size: .8rem; }}
+td.url a {{ color: var(--text-dim); text-decoration: none; }}
+td.url a:hover {{ color: var(--accent); text-decoration: underline; }}
+td.dim {{ color: var(--text-dim); font-family: "IBM Plex Mono", monospace; font-size: .8rem; }}
+.signal-bar {{ width: 100%; max-width: 280px; height: 14px; display: block; border-radius: 3px; overflow: hidden; }}
+
+.pill {{
+  color: #08130f; font-family: "IBM Plex Mono", monospace; font-weight: 500;
+  border-radius: 4px; padding: .12rem .5rem; font-size: .76rem; display: inline-block;
+}}
+
+.controls {{ display: flex; gap: .6rem; margin-bottom: .9rem; flex-wrap: wrap; }}
+input, select {{
+  padding: .45rem .6rem; border-radius: 6px; border: 1px solid var(--border);
+  background: var(--surface-2); color: var(--text); font-size: .85rem; font-family: inherit;
+}}
+input:focus, select:focus {{ outline: 2px solid var(--accent); outline-offset: 1px; }}
+#search {{ flex: 1; min-width: 180px; }}
+.table-wrap {{ overflow-x: auto; }}
+
+footer {{ text-align: center; color: var(--text-dim); font-size: .78rem; margin-top: 2.5rem; font-family: "IBM Plex Mono", monospace; }}
+footer a {{ color: var(--accent); }}
+
+@media (max-width: 720px) {{
+  .cards {{ grid-template-columns: repeat(3, 1fr); }}
+}}
+</style>
+</head>
+<body>
+<main>
+  <div class="masthead">
+    <h1>Channel Signal <span class="on-air">&mdash; {alive_pct_overall:.0f}% on air</span></h1>
+  </div>
+  <p class="subline">{total_channels} channels &middot; {len(by_list)} lists &middot; last swept {escape(last_checked)} &middot; page built {escape(generated_at)}</p>
+
+  <div class="cards">{cards}</div>
+
+  <section>
+    <h2>Alive share, over time</h2>
+    {svg_trend(trend)}
+  </section>
+
+  <section>
+    <h2>By list</h2>
+    {legend_html()}
+    <div class="table-wrap">
+    <table>
+      <thead><tr><th>List</th><th>Channels</th><th>Alive</th><th>Signal</th></tr></thead>
+      <tbody>{"".join(list_rows)}</tbody>
+    </table>
+    </div>
+  </section>
+
+  <section>
+    <h2>Needs attention</h2>
+    <p class="section-note"><code>blocked</code> usually means geo-restricted on purpose (the Ⓖ marker in the lists), not broken.
+       <code>disputed</code> looked dead over HTTP but ffprobe found a real stream behind it - worth a human look before touching it.</p>
+    <div class="controls">
+      <input type="search" id="search" placeholder="Search list or channel&hellip;">
+      <select id="state-filter">
+        <option value="">All states</option>
+        {state_options}
+      </select>
+    </div>
+    <div class="table-wrap">
+    <table id="problems">
+      <thead><tr><th>List</th><th>Channel</th><th>State</th><th>URL</th><th>Last checked</th></tr></thead>
+      <tbody>{"".join(problem_rows)}</tbody>
+    </table>
+    </div>
+  </section>
+
+  <footer>generated by <a href="https://github.com/Free-TV/IPTV/blob/master/check_channels.py">check_channels.py</a></footer>
+</main>
+<script>
+  const search = document.getElementById('search');
+  const stateFilter = document.getElementById('state-filter');
+  const rows = [...document.querySelectorAll('#problems tbody tr')];
+  function applyFilters() {{
+    const q = search.value.trim().toLowerCase();
+    const state = stateFilter.value;
+    for (const row of rows) {{
+      const matchesText = !q || row.textContent.toLowerCase().includes(q);
+      const matchesState = !state || row.dataset.state === state;
+      row.style.display = (matchesText && matchesState) ? '' : 'none';
+    }}
+  }}
+  search.addEventListener('input', applyFilters);
+  stateFilter.addEventListener('change', applyFilters);
+</script>
+</body>
+</html>
+"""
+    # pylint: enable=line-too-long
+
+
+def main():
+    """Generate the dashboard from the command-line-given (or default) history path."""
+    parser = argparse.ArgumentParser(
+        description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter,
+    )
+    parser.add_argument("--history", default=DEFAULT_HISTORY, help="path to history.jsonl")
+    parser.add_argument("--out", default=DEFAULT_OUT, help="path to write the dashboard HTML")
+    args = parser.parse_args()
+
+    generated_at = datetime.now(timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")
+    html = render(args.history, generated_at)
+
+    os.makedirs(os.path.dirname(args.out), exist_ok=True)
+    with open(args.out, "w", encoding="utf-8") as handle:
+        handle.write(html)
+    print(f"Wrote {args.out}")
+
+
+if __name__ == "__main__":
+    main()