Update tvj_epg.py

tvj_epg.py

@@ -1,6 +1,8 @@
 #!/usr/bin/env python3
 
 import os
+import re
+import html
 import requests
 import xml.etree.ElementTree as ET
 from datetime import datetime, timezone
@@ -8,15 +10,26 @@ from zoneinfo import ZoneInfo
 
 # ---------------- CONFIG ----------------
 
-EPG_URL = "https://www.1spotmedia.com/index.php/api/epg/get_epg_timeline_by_id"
+EPG_URL = "https://insight-api-shared.univtec.com/interface/epg"
+
+# Stable TVJ object ID returned by the new Univtec API
 CHANNEL_ID = "66e312890478fd00235244cb"
 
 TVG_ID = "TVJ.jm@SD"
 CHANNEL_NAME = "Television Jamaica"
 CHANNEL_ICON = "https://www.televisionjamaica.com/Portals/0/tvj_logo.png"
 
-# Podman/container output path (mount this directory)
-OUTPUT_FILE = "/app/output/tvj.xml"
+# Podman/container output path
+OUTPUT_FILE = os.environ.get(
+    "OUTPUT_FILE",
+    "/app/output/tvj.xml",
+)
+
+REQUEST_HEADERS = {
+    "accept": "application/json",
+    "platform": "web",
+    "x-tenant-id": "rjr",
+}
 
 # ----------------------------------------
 
@@ -34,58 +47,104 @@ def fmt_log_time(dt_utc: datetime) -> str:
     return dt_utc.astimezone(LOG_TZ).strftime("%Y-%m-%d %H:%M:%S %Z")
 
 
-def epoch_ms_to_xmltv(ts_ms: int) -> str:
-    """Convert epoch milliseconds to XMLTV timestamp (UTC)."""
-    dt = datetime.fromtimestamp(ts_ms / 1000, tz=timezone.utc)
-    return dt.strftime("%Y%m%d%H%M%S +0000")
+def parse_iso8601(value: str) -> datetime:
+    """Parse an ISO-8601 timestamp returned by Univtec."""
+    if not value:
+        raise ValueError("Empty timestamp")
+
+    # Python accepts +00:00 more consistently than trailing Z.
+    if value.endswith("Z"):
+        value = value[:-1] + "+00:00"
+
+    dt = datetime.fromisoformat(value)
+
+    if dt.tzinfo is None:
+        dt = dt.replace(tzinfo=timezone.utc)
+
+    return dt.astimezone(timezone.utc)
 
 
-def clean_text(s):
-    if s is None:
+def datetime_to_xmltv(dt: datetime) -> str:
+    """Convert datetime to XMLTV UTC timestamp."""
+    return dt.astimezone(timezone.utc).strftime("%Y%m%d%H%M%S +0000")
+
+
+def clean_text(value):
+    if value is None:
         return None
-    s = str(s).strip()
-    return s if s else None
+
+    value = str(value).strip()
+    return value if value else None
 
 
-def get_program_obj(item: dict) -> dict:
-    prog = item.get("program")
-    return prog if isinstance(prog, dict) else {}
+def strip_html(value):
+    """Convert provider HTML descriptions to normal text."""
+    value = clean_text(value)
+    if not value:
+        return None
+
+    value = re.sub(r"<br\s*/?>", "\n", value, flags=re.IGNORECASE)
+    value = re.sub(r"</p\s*>", "\n", value, flags=re.IGNORECASE)
+    value = re.sub(r"<[^>]+>", "", value)
+    value = html.unescape(value)
+
+    lines = [line.strip() for line in value.splitlines()]
+    value = " ".join(line for line in lines if line)
+
+    return clean_text(value)
+
+
+def get_custom(item: dict) -> dict:
+    custom = item.get("custom")
+    return custom if isinstance(custom, dict) else {}
 
 
 def get_description(item: dict) -> str:
     """
-    Extract description from nested program object.
-    Treat 'Dummy description.' as missing.
+    Prefer the normal description.
+    Fall back to custom.longDescription with HTML removed.
     """
-    prog = get_program_obj(item)
+    custom = get_custom(item)
 
     desc = (
-        prog.get("longDescription")
-        or prog.get("description")
-        or prog.get("shortDescription")
+        clean_text(item.get("description"))
+        or strip_html(custom.get("longDescription"))
     )
-    desc = clean_text(desc)
-
-    if desc and desc.lower() == "dummy description.":
-        desc = None
 
     return desc or "No description available."
 
 
-def get_category(item: dict):
-    """Use displayGenre if present."""
-    return clean_text(get_program_obj(item).get("displayGenre"))
+def add_categories(programme_el, item: dict):
+    """Add each provider genre as an XMLTV category."""
+    custom = get_custom(item)
+    genres = custom.get("genre")
+
+    if isinstance(genres, str):
+        genres = [genres]
+
+    if not isinstance(genres, list):
+        return
+
+    for genre in genres:
+        genre = clean_text(genre)
+        if genre:
+            ET.SubElement(
+                programme_el,
+                "category",
+                lang="en"
+            ).text = genre
 
 
 def add_episode_numbers(programme_el, item: dict):
     """
-    Add episode number in xmltv_ns format if possible.
-    xmltv_ns: season is 0-based, episode is 0-based.
-    """
-    prog = get_program_obj(item)
+    Add episode number using xmltv_ns when available.
 
-    season = prog.get("tvSeasonNumber")
-    episode = prog.get("tvSeasonEpisodeNumber") or prog.get("seriesEpisodeNumber")
+    xmltv_ns is zero-based.
+    """
+    custom = get_custom(item)
+
+    season = custom.get("seasonNumber")
+    episode = custom.get("episodeNumber")
 
     try:
         season_i = int(season) if season is not None else None
@@ -100,11 +159,24 @@ def add_episode_numbers(programme_el, item: dict):
     if season_i is None and episode_i is None:
         return
 
-    # xmltv_ns is 0-based; assume provider values are 1-based if >= 1
-    s0 = (season_i - 1) if (season_i is not None and season_i >= 1) else 0
-    e0 = (episode_i - 1) if (episode_i is not None and episode_i >= 1) else 0
+    s0 = (
+        season_i - 1
+        if season_i is not None and season_i >= 1
+        else 0
+    )
+
+    e0 = (
+        episode_i - 1
+        if episode_i is not None and episode_i >= 1
+        else 0
+    )
+
+    ep = ET.SubElement(
+        programme_el,
+        "episode-num",
+        system="xmltv_ns"
+    )
 
-    ep = ET.SubElement(programme_el, "episode-num", system="xmltv_ns")
     if season_i is not None and episode_i is not None:
         ep.text = f"{s0}.{e0}."
     elif season_i is not None:
@@ -115,111 +187,267 @@ def add_episode_numbers(programme_el, item: dict):
 
 def add_date(programme_el, item: dict):
     """
-    Add <date>YYYY</date> if we can find a year.
-    Try: program.year, else pubDate epoch-ms -> year.
+    Add <date>YYYY</date> using custom.airDate when available.
     """
-    prog = get_program_obj(item)
+    custom = get_custom(item)
+    air_date = custom.get("airDate")
 
-    year = prog.get("year")
-    year_out = None
+    if air_date is None:
+        return
 
     try:
-        if year is not None:
-            y = int(year)
-            if 1800 <= y <= 2100:
-                year_out = str(y)
-    except (ValueError, TypeError):
-        year_out = None
+        dt = datetime.fromtimestamp(
+            int(air_date) / 1000,
+            tz=timezone.utc,
+        )
+
+        ET.SubElement(programme_el, "date").text = str(dt.year)
 
-    if year_out is None:
-        pub = prog.get("pubDate")
-        try:
-            if pub is not None:
-                dt = datetime.fromtimestamp(int(pub) / 1000, tz=timezone.utc)
-                year_out = str(dt.year)
     except (ValueError, TypeError, OSError):
-            year_out = None
-
-    if year_out:
-        ET.SubElement(programme_el, "date").text = year_out
+        pass
 
 
 def add_length(programme_el, item: dict):
     """
-    Add <length units="seconds">N</length> from program.runtime if present.
+    Add programme duration from custom.duration.
+
+    Univtec duration appears to be seconds.
     """
-    prog = get_program_obj(item)
-    runtime = prog.get("runtime")
-    if runtime is None:
+    custom = get_custom(item)
+    duration = custom.get("duration")
+
+    if duration is None:
         return
 
     try:
-        seconds = int(str(runtime).strip())
+        seconds = int(duration)
+
         if seconds > 0:
-            ET.SubElement(programme_el, "length", units="seconds").text = str(seconds)
+            ET.SubElement(
+                programme_el,
+                "length",
+                units="seconds"
+            ).text = str(seconds)
+
     except (ValueError, TypeError):
+        pass
+
+
+def add_rating(programme_el, item: dict):
+    """Add provider content rating."""
+    rating = clean_text(get_custom(item).get("rating"))
+
+    if not rating:
         return
 
+    rating_el = ET.SubElement(programme_el, "rating")
+    ET.SubElement(rating_el, "value").text = rating
+
+
+def add_icon(programme_el, item: dict):
+    """Add programme artwork if supplied by Univtec."""
+    custom = get_custom(item)
+
+    image = custom.get("image")
+    poster = custom.get("poster")
+
+    image_url = None
+
+    if isinstance(image, dict):
+        image_url = clean_text(image.get("downloadUrl"))
+
+    if not image_url and isinstance(poster, dict):
+        image_url = clean_text(poster.get("downloadUrl"))
+
+    if image_url:
+        ET.SubElement(
+            programme_el,
+            "icon",
+            src=image_url,
+        )
+
+
+def fetch_tvj_events():
+    """
+    Fetch the complete RJR EPG response and extract the TVJ object.
+    """
+    resp = requests.get(
+        EPG_URL,
+        headers=REQUEST_HEADERS,
+        timeout=30,
+    )
+    resp.raise_for_status()
+
+    data = resp.json()
+
+    if not isinstance(data, list) or not data:
+        raise RuntimeError(
+            "EPG response is empty or invalid "
+            "(expected a non-empty list)."
+        )
+
+    channel = next(
+        (
+            item
+            for item in data
+            if item.get("id") == CHANNEL_ID
+        ),
+        None,
+    )
+
+    if channel is None:
+        raise RuntimeError(
+            f"TVJ channel ID {CHANNEL_ID} "
+            "was not found in EPG response."
+        )
+
+    epg_object = channel.get("epgObject")
+
+    if not isinstance(epg_object, dict):
+        raise RuntimeError(
+            "TVJ EPG response has no valid epgObject."
+        )
+
+    events = epg_object.get("events")
+
+    if not isinstance(events, list):
+        raise RuntimeError(
+            "TVJ epgObject.events is missing or invalid."
+        )
+
+    return events
+
 
 def main():
     run_time_utc = datetime.now(timezone.utc)
-    print(f"[{fmt_log_time(run_time_utc)}] Script run started (TZ={TZ_NAME})")
 
-    resp = requests.get(EPG_URL, params={"id": CHANNEL_ID}, timeout=30)
-    resp.raise_for_status()
-    data = resp.json()
+    print(
+        f"[{fmt_log_time(run_time_utc)}] "
+        f"Script run started (TZ={TZ_NAME})"
+    )
 
-    if not isinstance(data, list) or not data:
-        raise RuntimeError("EPG response is empty or invalid (expected a non-empty list).")
+    events = fetch_tvj_events()
 
-    # EPG "updated until" = furthest programme end time in returned schedule
-    last_update_epoch = max(int(p["endTime"]) for p in data if "endTime" in p)
-    last_update_utc = datetime.fromtimestamp(last_update_epoch / 1000, tz=timezone.utc)
+    # Discard malformed entries and provider filler.
+    valid_events = []
 
-    tv = ET.Element("tv", attrib={"generator-info-name": "1SpotMedia TVJ EPG (enriched)"})
+    for item in events:
+        if not isinstance(item, dict):
+            continue
+
+        if not item.get("start") or not item.get("end"):
+            continue
+
+        title = clean_text(item.get("title"))
+
+        if title and title.casefold() == "no program":
+            continue
+
+        try:
+            start_dt = parse_iso8601(item["start"])
+            end_dt = parse_iso8601(item["end"])
+        except (ValueError, TypeError):
+            continue
+
+        if end_dt <= start_dt:
+            continue
+
+        valid_events.append((item, start_dt, end_dt))
+
+    if not valid_events:
+        raise RuntimeError(
+            "No valid TVJ programmes were returned."
+        )
+
+    # Sort defensively in case provider changes ordering.
+    valid_events.sort(key=lambda x: x[1])
+
+    last_update_utc = max(
+        end_dt
+        for _, _, end_dt in valid_events
+    )
+
+    tv = ET.Element(
+        "tv",
+        attrib={
+            "generator-info-name":
+                "Univtec / 1SpotMedia TVJ EPG (enriched)"
+        },
+    )
 
     # Channel block
     ch = ET.SubElement(tv, "channel", id=TVG_ID)
-    ET.SubElement(ch, "display-name").text = CHANNEL_NAME
-    ET.SubElement(ch, "icon", src=CHANNEL_ICON)
+
+    ET.SubElement(
+        ch,
+        "display-name",
+    ).text = CHANNEL_NAME
+
+    ET.SubElement(
+        ch,
+        "icon",
+        src=CHANNEL_ICON,
+    )
 
     # Programmes
-    for item in data:
-        if "startTime" not in item or "endTime" not in item:
-            continue
+    for item, start_dt, end_dt in valid_events:
 
         programme = ET.SubElement(
             tv,
             "programme",
-            start=epoch_ms_to_xmltv(int(item["startTime"])),
-            stop=epoch_ms_to_xmltv(int(item["endTime"])),
+            start=datetime_to_xmltv(start_dt),
+            stop=datetime_to_xmltv(end_dt),
             channel=TVG_ID,
         )
 
-        title = ET.SubElement(programme, "title", lang="en")
-        title.text = (
+        title = (
             clean_text(item.get("title"))
-            or clean_text(get_program_obj(item).get("title"))
+            or clean_text(get_custom(item).get("seriesTitle"))
             or "Unknown Program"
         )
 
-        ET.SubElement(programme, "desc", lang="en").text = get_description(item)
+        ET.SubElement(
+            programme,
+            "title",
+            lang="en",
+        ).text = title
 
-        cat = get_category(item)
-        if cat:
-            ET.SubElement(programme, "category", lang="en").text = cat
+        ET.SubElement(
+            programme,
+            "desc",
+            lang="en",
+        ).text = get_description(item)
 
+        add_categories(programme, item)
         add_episode_numbers(programme, item)
         add_date(programme, item)
         add_length(programme, item)
+        add_rating(programme, item)
+        add_icon(programme, item)
 
-    # Write XMLTV file (overwrites)
-    ET.ElementTree(tv).write(OUTPUT_FILE, encoding="utf-8", xml_declaration=True)
+    # Ensure output directory exists.
+    output_dir = os.path.dirname(OUTPUT_FILE)
+
+    if output_dir:
+        os.makedirs(output_dir, exist_ok=True)
+
+    # Write XMLTV file.
+    ET.ElementTree(tv).write(
+        OUTPUT_FILE,
+        encoding="utf-8",
+        xml_declaration=True,
+    )
 
     done_utc = datetime.now(timezone.utc)
-    print(f"[{fmt_log_time(done_utc)}] XMLTV written: {OUTPUT_FILE}")
+
     print(
-        f"Summary: Last run={fmt_log_time(run_time_utc)}, "
+        f"[{fmt_log_time(done_utc)}] "
+        f"XMLTV written: {OUTPUT_FILE}"
+    )
+
+    print(
+        f"Summary: Programmes={len(valid_events)}, "
+        f"Last run={fmt_log_time(run_time_utc)}, "
         f"EPG updated until={fmt_log_time(last_update_utc)}"
     )