Capture Steam Machine launch watch baseline

scripts/check_customs_shipments.sh

@@ -33,11 +33,14 @@ looks_like_importinfo_shipment_table() {
 
 looks_like_importgenius_shipment_page() {
     path=$1
-    for marker in "Bill of Lading" "GAME CONSOLE" "VALVE CORPORATION"; do
+    for marker in "Bill of Lading" "VALVE CORPORATION"; do
         if ! grep -qi "$marker" "$path"; then
             return 1
         fi
     done
+    if ! grep -Eqi "GAME CONSOLE|VIRTUAL REALITY|WIRELESS PC CONTROLLER|VR CONTROLLER|HEADSET" "$path"; then
+        return 1
+    fi
     return 0
 }
 
@@ -123,6 +126,18 @@ fetch_importgenius \
     "importgenius-ingram-valve" \
     "importgenius-ingram-valve" \
     "https://www.importgenius.com/importers/ingram-micro-c-o-valve-corporation" || true
+fetch_importgenius \
+    "importgenius-ceva-valve" \
+    "importgenius-ceva-valve" \
+    "https://www.importgenius.com/importers/ceva-c-o-valve-corporation" || true
+fetch_importgenius \
+    "importgenius-ceva-nl-valve" \
+    "importgenius-ceva-nl-valve" \
+    "https://www.importgenius.com/importers/ceva-nl-c-o-valve-corporation" || true
+fetch_importgenius \
+    "importgenius-valve-corp" \
+    "importgenius-valve-corp" \
+    "https://www.importgenius.com/importers/valve-corp" || true
 
 python3 "$SCRIPT_DIR/parse_importinfo_shipments.py" \
     --manifest "$MANIFEST" \

scripts/check_komodo.sh

@@ -33,8 +33,71 @@ load_local_env() {
   fi
 }
 
+configure_playwright_runtime() {
+  NODE_BIN="${NODE_BIN:-}"
+  if [ -z "$NODE_BIN" ]; then
+    NODE_BIN="$(command -v node 2>/dev/null || true)"
+  fi
+  bundled_node="$HOME/.cache/codex-runtimes/codex-primary-runtime/dependencies/node/bin/node"
+  if [ -z "$NODE_BIN" ] && [ -x "$bundled_node" ]; then
+    NODE_BIN="$bundled_node"
+  fi
+  export NODE_BIN
+
+  if [ -z "${NODE_PATH:-}" ]; then
+    for node_modules in \
+      "$REPO_DIR/node_modules" \
+      "$HOME/.cache/codex-runtimes/codex-primary-runtime/dependencies/node/node_modules"
+    do
+      if [ -d "$node_modules" ]; then
+        NODE_PATH="$node_modules"
+        export NODE_PATH
+        break
+      fi
+    done
+  fi
+}
+
+cdp_endpoint_alive() {
+  endpoint="${PLAYWRIGHT_CDP_ENDPOINT:-}"
+  if [ -z "$endpoint" ]; then
+    return 1
+  fi
+  curl --max-time 2 -fsS "$endpoint/json/version" >/dev/null 2>&1
+}
+
+KOMODO_BOOTSTRAP_ATTEMPTED=0
+
+bootstrap_dedicated_browser() {
+  if [ "${KOMODO_PLAYWRIGHT_FALLBACK:-0}" != "1" ] || [ "${KOMODO_AUTO_BOOTSTRAP:-1}" = "0" ]; then
+    return 0
+  fi
+  if [ "$KOMODO_BOOTSTRAP_ATTEMPTED" = "1" ]; then
+    return 0
+  fi
+
+  KOMODO_BOOTSTRAP_ATTEMPTED=1
+  bootstrap_cmd="${KOMODO_BOOTSTRAP_CMD:-$SCRIPT_DIR/bootstrap_komodo.sh}"
+  if [ ! -x "$bootstrap_cmd" ]; then
+    printf '%s\n' "Komodo bootstrap command not executable: $bootstrap_cmd" >> "$ERROR_FILE"
+    return 0
+  fi
+
+  if "$bootstrap_cmd" >/dev/null 2>>"$ERROR_FILE"; then
+    load_local_env
+  fi
+}
+
 maybe_set_cdp_endpoint() {
-  if [ -n "${PLAYWRIGHT_CDP_ENDPOINT:-}" ] || [ -z "${PLAYWRIGHT_PROFILE_DIR:-}" ]; then
+  if [ -n "${PLAYWRIGHT_CDP_ENDPOINT:-}" ]; then
+    if cdp_endpoint_alive; then
+      return 0
+    fi
+    unset PLAYWRIGHT_CDP_ENDPOINT
+  fi
+
+  if [ -z "${PLAYWRIGHT_PROFILE_DIR:-}" ]; then
+    bootstrap_dedicated_browser
     return 0
   fi
 
@@ -50,7 +113,13 @@ maybe_set_cdp_endpoint() {
   if [ -n "$port" ]; then
     PLAYWRIGHT_CDP_ENDPOINT="http://127.0.0.1:$port"
     export PLAYWRIGHT_CDP_ENDPOINT
+    if cdp_endpoint_alive; then
+      return 0
+    fi
+    unset PLAYWRIGHT_CDP_ENDPOINT
   fi
+
+  bootstrap_dedicated_browser
 }
 
 load_local_env
@@ -71,14 +140,21 @@ fetch_json() {
 
   rm -f "$tmp"
 
-  if [ "$allow_fallback" = "1" ] && [ "${KOMODO_PLAYWRIGHT_FALLBACK:-0}" = "1" ] && command -v node >/dev/null 2>&1; then
+  if [ "$allow_fallback" = "1" ] && [ "${KOMODO_PLAYWRIGHT_FALLBACK:-0}" = "1" ]; then
+    configure_playwright_runtime
+    if [ -z "${NODE_BIN:-}" ]; then
+      printf '%s\n' "Playwright fallback unavailable: node executable not found" >> "$ERROR_FILE"
+      rm -f "$tmp"
+      printf '%s\n' "failed to fetch JSON from $url" >> "$ERROR_FILE"
+      return 1
+    fi
     maybe_set_cdp_endpoint
     PLAYWRIGHT_CORE_PATH="${PLAYWRIGHT_CORE_PATH:-}"
     if command -v playwright-cli >/dev/null 2>&1; then
       PLAYWRIGHT_CLI_BIN="$(command -v playwright-cli)"
       PLAYWRIGHT_CORE_PATH="$(CDPATH= cd -- "$(dirname "$PLAYWRIGHT_CLI_BIN")/../lib/node_modules/@playwright/cli/node_modules/playwright-core" 2>/dev/null && pwd || true)"
     fi
-    PLAYWRIGHT_CORE_PATH="$PLAYWRIGHT_CORE_PATH" node "$SCRIPT_DIR/fetch_with_playwright.js" "$url" "$tmp"
+    PLAYWRIGHT_CORE_PATH="$PLAYWRIGHT_CORE_PATH" "$NODE_BIN" "$SCRIPT_DIR/fetch_with_playwright.js" "$url" "$tmp"
     if jq empty "$tmp" >/dev/null 2>&1; then
       mv "$tmp" "$OUT_DIR/$name.json"
       return 0
@@ -124,8 +200,14 @@ do
   fi
 done
 
-for id in $(jq -r '.[].id' "$OUT_DIR/sections-controller-search.json" 2>/dev/null || true); do
-  fetch_json "media-parent-section-$id" "https://komodostation.com/wp-json/wp/v2/media?parent=$id&per_page=100" || true
+for section_report in \
+  sections-controller-search \
+  sections-machine-search \
+  sections-frame-search
+do
+  for id in $(jq -r '.[].id' "$OUT_DIR/$section_report.json" 2>/dev/null || true); do
+    fetch_json "media-parent-section-$id" "https://komodostation.com/wp-json/wp/v2/media?parent=$id&per_page=100" || true
+  done
 done
 
 {
@@ -153,8 +235,24 @@ jq -r '
   | @tsv
 ' "$OUT_DIR/media-controller-search.json" > "$REPORT_DIR/komodo-controller-media-interesting.tsv"
 
-: > "$REPORT_DIR/komodo-machine-media.tsv"
-: > "$REPORT_DIR/komodo-frame-media.tsv"
+write_section_media_report() {
+  sections_file="$1"
+  output_file="$2"
+  : > "$output_file"
+  for id in $(jq -r '.[].id' "$sections_file" 2>/dev/null || true); do
+    media_file="$OUT_DIR/media-parent-section-$id.json"
+    if [ -f "$media_file" ]; then
+      jq -r '
+        .[]?
+        | [.id,.date,.modified,.slug,.mime_type,.post,.source_url]
+        | @tsv
+      ' "$media_file" >> "$output_file"
+    fi
+  done
+}
+
+write_section_media_report "$OUT_DIR/sections-machine-search.json" "$REPORT_DIR/komodo-machine-media.tsv"
+write_section_media_report "$OUT_DIR/sections-frame-search.json" "$REPORT_DIR/komodo-frame-media.tsv"
 
 jq -r '
   .[]?

scripts/check_steamdb.sh

@@ -26,8 +26,71 @@ load_local_env() {
   fi
 }
 
+configure_playwright_runtime() {
+  NODE_BIN="${NODE_BIN:-}"
+  if [ -z "$NODE_BIN" ]; then
+    NODE_BIN="$(command -v node 2>/dev/null || true)"
+  fi
+  bundled_node="$HOME/.cache/codex-runtimes/codex-primary-runtime/dependencies/node/bin/node"
+  if [ -z "$NODE_BIN" ] && [ -x "$bundled_node" ]; then
+    NODE_BIN="$bundled_node"
+  fi
+  export NODE_BIN
+
+  if [ -z "${NODE_PATH:-}" ]; then
+    for node_modules in \
+      "$REPO_DIR/node_modules" \
+      "$HOME/.cache/codex-runtimes/codex-primary-runtime/dependencies/node/node_modules"
+    do
+      if [ -d "$node_modules" ]; then
+        NODE_PATH="$node_modules"
+        export NODE_PATH
+        break
+      fi
+    done
+  fi
+}
+
+cdp_endpoint_alive() {
+  endpoint="${STEAMDB_CDP_ENDPOINT:-}"
+  if [ -z "$endpoint" ]; then
+    return 1
+  fi
+  curl --max-time 2 -fsS "$endpoint/json/version" >/dev/null 2>&1
+}
+
+STEAMDB_BOOTSTRAP_ATTEMPTED=0
+
+bootstrap_dedicated_browser() {
+  if [ "${STEAMDB_PLAYWRIGHT_FALLBACK:-0}" != "1" ] || [ "${STEAMDB_AUTO_BOOTSTRAP:-1}" = "0" ]; then
+    return 0
+  fi
+  if [ "$STEAMDB_BOOTSTRAP_ATTEMPTED" = "1" ]; then
+    return 0
+  fi
+
+  STEAMDB_BOOTSTRAP_ATTEMPTED=1
+  bootstrap_cmd="${STEAMDB_BOOTSTRAP_CMD:-$SCRIPT_DIR/bootstrap_steamdb.sh}"
+  if [ ! -x "$bootstrap_cmd" ]; then
+    printf '%s\n' "SteamDB bootstrap command not executable: $bootstrap_cmd" >> "$ERROR_FILE"
+    return 0
+  fi
+
+  if "$bootstrap_cmd" >/dev/null 2>>"$ERROR_FILE"; then
+    load_local_env
+  fi
+}
+
 maybe_set_cdp_endpoint() {
-  if [ -n "${STEAMDB_CDP_ENDPOINT:-}" ] || [ -z "${STEAMDB_PROFILE_DIR:-}" ]; then
+  if [ -n "${STEAMDB_CDP_ENDPOINT:-}" ]; then
+    if cdp_endpoint_alive; then
+      return 0
+    fi
+    unset STEAMDB_CDP_ENDPOINT
+  fi
+
+  if [ -z "${STEAMDB_PROFILE_DIR:-}" ]; then
+    bootstrap_dedicated_browser
     return 0
   fi
 
@@ -43,7 +106,13 @@ maybe_set_cdp_endpoint() {
   if [ -n "$port" ]; then
     STEAMDB_CDP_ENDPOINT="http://127.0.0.1:$port"
     export STEAMDB_CDP_ENDPOINT
+    if cdp_endpoint_alive; then
+      return 0
+    fi
+    unset STEAMDB_CDP_ENDPOINT
   fi
+
+  bootstrap_dedicated_browser
 }
 
 is_challenge_page() {
@@ -74,9 +143,11 @@ fetch_html() {
     fi
   fi
 
-  if [ "${STEAMDB_PLAYWRIGHT_FALLBACK:-0}" = "1" ] && command -v node >/dev/null 2>&1; then
+  if [ "${STEAMDB_PLAYWRIGHT_FALLBACK:-0}" = "1" ]; then
+    configure_playwright_runtime
     maybe_set_cdp_endpoint
-    if node "$SCRIPT_DIR/fetch_steamdb_with_playwright.js" "$url" "$tmp" >/dev/null 2>&1 \
+    if [ -n "${NODE_BIN:-}" ] \
+      && "$NODE_BIN" "$SCRIPT_DIR/fetch_steamdb_with_playwright.js" "$url" "$tmp" >/dev/null 2>&1 \
       && [ -s "$tmp" ] \
       && ! is_challenge_page "$tmp"; then
       mv "$tmp" "$out"

scripts/check_steamvr_depots.sh

@@ -9,6 +9,8 @@ fi
 RUN_DIR="$1"
 OUT_DIR="$RUN_DIR/api/steamvr-depots"
 REPORT_DIR="$RUN_DIR/reports"
+SCRIPT_DIR="$(CDPATH= cd -- "$(dirname "$0")" && pwd)"
+REPO_DIR="$(CDPATH= cd -- "$SCRIPT_DIR/.." && pwd)"
 mkdir -p "$OUT_DIR" "$REPORT_DIR"
 
 ERROR_FILE="$REPORT_DIR/steamvr-depots-errors.txt"
@@ -28,6 +30,101 @@ is_challenge_page() {
   rg -qi "Checking your browser|Just a moment|Cloudflare|cf-chl|cf-browser-verification" "$path"
 }
 
+load_steamdb_env() {
+  env_file="${STEAMDB_ENV_FILE:-$REPO_DIR/.local/steamdb-env.sh}"
+  if [ -f "$env_file" ]; then
+    # shellcheck disable=SC1090
+    . "$env_file"
+  fi
+}
+
+steamdb_cdp_endpoint_alive() {
+  endpoint="${STEAMDB_CDP_ENDPOINT:-}"
+  if [ -z "$endpoint" ]; then
+    return 1
+  fi
+  curl --max-time 2 -fsS "$endpoint/json/version" >/dev/null 2>&1
+}
+
+STEAMDB_BOOTSTRAP_ATTEMPTED=0
+
+bootstrap_steamdb_browser() {
+  if [ "${STEAMDB_PLAYWRIGHT_FALLBACK:-0}" != "1" ] || [ "${STEAMDB_AUTO_BOOTSTRAP:-1}" = "0" ]; then
+    return 0
+  fi
+  if [ "$STEAMDB_BOOTSTRAP_ATTEMPTED" = "1" ]; then
+    return 0
+  fi
+
+  STEAMDB_BOOTSTRAP_ATTEMPTED=1
+  bootstrap_cmd="${STEAMDB_BOOTSTRAP_CMD:-$SCRIPT_DIR/bootstrap_steamdb.sh}"
+  if [ ! -x "$bootstrap_cmd" ]; then
+    printf '%s\n' "SteamDB bootstrap command not executable: $bootstrap_cmd" >> "$ERROR_FILE"
+    return 0
+  fi
+
+  if "$bootstrap_cmd" >/dev/null 2>>"$ERROR_FILE"; then
+    load_steamdb_env
+  fi
+}
+
+ensure_steamdb_browser() {
+  if [ "${STEAMDB_PLAYWRIGHT_FALLBACK:-0}" != "1" ]; then
+    return 0
+  fi
+
+  if [ -n "${STEAMDB_CDP_ENDPOINT:-}" ]; then
+    if steamdb_cdp_endpoint_alive; then
+      return 0
+    fi
+    unset STEAMDB_CDP_ENDPOINT
+  fi
+
+  if [ -n "${STEAMDB_PROFILE_DIR:-}" ]; then
+    port="$(ps aux | awk -v profile="$STEAMDB_PROFILE_DIR" '
+      index($0, profile) && match($0, /--remote-debugging-port=([0-9]+)/) {
+        value = substr($0, RSTART, RLENGTH)
+        sub(/^--remote-debugging-port=/, "", value)
+        print value
+        exit
+      }
+    ' 2>/dev/null || true)"
+
+    if [ -n "$port" ]; then
+      STEAMDB_CDP_ENDPOINT="http://127.0.0.1:$port"
+      export STEAMDB_CDP_ENDPOINT
+      if steamdb_cdp_endpoint_alive; then
+        return 0
+      fi
+      unset STEAMDB_CDP_ENDPOINT
+    fi
+  fi
+
+  bootstrap_steamdb_browser
+}
+
+fetch_html_with_steamdb_playwright() {
+  url="$1"
+  output="$2"
+  tmp="$output.tmp.playwright"
+
+  if [ "${STEAMDB_PLAYWRIGHT_FALLBACK:-0}" != "1" ] || ! command -v node >/dev/null 2>&1; then
+    return 1
+  fi
+
+  ensure_steamdb_browser
+
+  if node "$SCRIPT_DIR/fetch_steamdb_with_playwright.js" "$url" "$tmp" >/dev/null 2>&1 \
+    && [ -s "$tmp" ] \
+    && ! is_challenge_page "$tmp"; then
+    mv "$tmp" "$output"
+    return 0
+  fi
+
+  rm -f "$tmp"
+  return 1
+}
+
 fetch_html() {
   name="$1"
   url="$2"
@@ -40,6 +137,11 @@ fetch_html() {
     return 0
   fi
 
+  rm -f "$tmp"
+  if fetch_html_with_steamdb_playwright "$url" "$out"; then
+    return 0
+  fi
+
   rm -f "$tmp" "$out"
   printf '%s\n' "failed to fetch $url: blocked or challenge page" >> "$ERROR_FILE"
   return 1
@@ -57,6 +159,8 @@ fetch_json() {
   fi
 }
 
+load_steamdb_env
+
 fetch_html "steamdb-depots.html" "https://steamdb.info/app/250820/depots/" || true
 fetch_json "steamvr-news.json" "https://api.steampowered.com/ISteamNews/GetNewsForApp/v2/?appid=250820&count=40&maxlength=12000&format=json" || true
 

scripts/download_with_playwright.js

@@ -6,13 +6,14 @@ function loadPlaywright() {
   const candidates = [];
   if (process.env.PLAYWRIGHT_CORE_PATH) candidates.push(process.env.PLAYWRIGHT_CORE_PATH);
   candidates.push("playwright-core");
+  candidates.push("playwright");
   for (const candidate of candidates) {
     try {
       return require(candidate);
     } catch (err) {
     }
   }
-  throw new Error("Unable to load playwright-core. Set PLAYWRIGHT_CORE_PATH or install playwright-core.");
+  throw new Error("Unable to load Playwright. Set PLAYWRIGHT_CORE_PATH, set NODE_PATH to a node_modules directory, or install playwright/playwright-core.");
 }
 
 function launchOptions(headless) {

scripts/fetch_steamdb_with_playwright.js

@@ -6,6 +6,7 @@ function loadPlaywright() {
   const candidates = [];
   if (process.env.PLAYWRIGHT_CORE_PATH) candidates.push(process.env.PLAYWRIGHT_CORE_PATH);
   candidates.push("playwright-core");
+  candidates.push("playwright");
   for (const candidate of candidates) {
     try {
       return require(candidate);
@@ -13,7 +14,7 @@ function loadPlaywright() {
     }
   }
   throw new Error(
-    "Unable to load playwright-core. Set PLAYWRIGHT_CORE_PATH or install playwright-core."
+    "Unable to load Playwright. Set PLAYWRIGHT_CORE_PATH, set NODE_PATH to a node_modules directory, or install playwright/playwright-core."
   );
 }
 

scripts/fetch_with_playwright.js

@@ -6,6 +6,7 @@ function loadPlaywright() {
   const candidates = [];
   if (process.env.PLAYWRIGHT_CORE_PATH) candidates.push(process.env.PLAYWRIGHT_CORE_PATH);
   candidates.push("playwright-core");
+  candidates.push("playwright");
   for (const candidate of candidates) {
     try {
       return require(candidate);
@@ -13,7 +14,7 @@ function loadPlaywright() {
     }
   }
   throw new Error(
-    "Unable to load playwright-core. Set PLAYWRIGHT_CORE_PATH or install playwright-core."
+    "Unable to load Playwright. Set PLAYWRIGHT_CORE_PATH, set NODE_PATH to a node_modules directory, or install playwright/playwright-core."
   );
 }
 

scripts/parse_importinfo_shipments.py

@@ -56,6 +56,9 @@ OUTPUT_FIELDS = [
 PARTY_TERMS = ("VALVE", "CEVA", "INGRAM MICRO", "TECH-FRONT", "CHENG UEI")
 PRODUCT_TERMS = (
     "GAME CONSOLE",
+    "VIRTUAL REALITY",
+    "VIRTUAL REALITY DEVICES",
+    "WIRELESS PC CONTROLLER",
     "VR CONTROLLER",
     "CONTROLLER",
     "STEAM",
@@ -70,6 +73,9 @@ SOURCE_URLS = {
     "tech-front-game-console": "https://www.importinfo.com/search?s=TECH-FRONT%20GAME%20CONSOLE%20VALVE",
     "valve-corporation-game-console": "https://www.importinfo.com/search?s=VALVE%20CORPORATION%20GAME%20CONSOLE",
     "importgenius-ingram-valve": "https://www.importgenius.com/importers/ingram-micro-c-o-valve-corporation",
+    "importgenius-ceva-valve": "https://www.importgenius.com/importers/ceva-c-o-valve-corporation",
+    "importgenius-ceva-nl-valve": "https://www.importgenius.com/importers/ceva-nl-c-o-valve-corporation",
+    "importgenius-valve-corp": "https://www.importgenius.com/importers/valve-corp",
 }
 
 
@@ -135,11 +141,11 @@ class TextParser(HTMLParser):
         self.parts = []
 
     def handle_starttag(self, tag, attrs):
-        if tag in ("br", "p", "tr", "li", "h1", "h2", "h3", "section"):
+        if tag in ("br", "p", "tr", "li", "h1", "h2", "h3", "section", "div"):
             self.parts.append("\n")
 
     def handle_endtag(self, tag):
-        if tag in ("p", "tr", "li", "h1", "h2", "h3", "section"):
+        if tag in ("p", "tr", "li", "h1", "h2", "h3", "section", "div"):
             self.parts.append("\n")
 
     def handle_data(self, data):
@@ -148,6 +154,10 @@ class TextParser(HTMLParser):
     def text(self):
         return html.unescape(" ".join("".join(self.parts).split()))
 
+    def lines(self):
+        raw_text = html.unescape("".join(self.parts))
+        return [" ".join(line.split()) for line in raw_text.splitlines() if line.strip()]
+
 
 def normalize_header(value):
     return FIELD_MAP.get(value.strip().lower(), "")
@@ -177,34 +187,83 @@ def parse_rows(spec):
 def parse_importgenius_rows(spec):
     parser = TextParser()
     parser.feed(spec.path.read_text(encoding="utf-8", errors="ignore"))
-    text = parser.text()
+    text_lines = parser.lines() or [parser.text()]
+    importer_pattern = (
+        r"(?:INGRAM\s+MICRO|CEVA(?:\s+NL)?)\s+C/O\s+VALVE\s+CORPORATION"
+        r"|VALVE\s+CORPORATION"
+    )
+    supplier_pattern = (
+        r"TECH-?FRONT\s+\(CHONGQING\)\s+COMPUTER\s+CO"
+        r"|CHENG\s+UEI\s+PRECISION\s+IND\.\s+CO\s+LTD"
+    )
     pattern = re.compile(
         r"(?:^|\s)(?P<rank>\d+)\s+"
         r"(?P<bol>[A-Z0-9]+)\s+"
-        r"(?P<product>GAME\s+CONSOLE\s*\.?)\s+"
-        r"(?P<importer>(?:INGRAM\s+MICRO|CEVA)\s+C/O\s+VALVE\s+CORPORATION)\s+"
-        r"(?P<supplier>TECH-?FRONT\s+\(CHONGQING\)\s+COMPUTER\s+CO)\s+"
+        r"(?P<product>[A-Z0-9][A-Z0-9 ./&()'-]+?)\s+"
+        r"(?P<importer>" + importer_pattern + r")\s+"
+        r"(?P<supplier>" + supplier_pattern + r")\s+"
         r"(?P<arrival>\d{4}-\d{2}-\d{2})\s+"
         r"(?P<country>[A-Za-z ]+?)\s+"
         r"(?P<weight>[\d,]+\s+Kgs)\s+"
         r"(?P<quantity>\d+\s+PKG)",
         re.IGNORECASE,
     )
-    rows = []
-    for match in pattern.finditer(text):
+    importer_re = re.compile(importer_pattern, re.IGNORECASE)
+    supplier_re = re.compile(supplier_pattern, re.IGNORECASE)
+    rank_re = re.compile(r"\d+")
+    bol_re = re.compile(r"[A-Z0-9]+")
+    date_re = re.compile(r"\d{4}-\d{2}-\d{2}")
+    country_re = re.compile(r"[A-Za-z ]+")
+    weight_re = re.compile(r"[\d,]+\s+Kgs", re.IGNORECASE)
+    quantity_re = re.compile(r"\d+\s+PKG", re.IGNORECASE)
+
+    def build_record(match):
         record = {field: "" for field in OUTPUT_FIELDS}
         record["source"] = "importgenius"
         record["query"] = spec.query
-        record["house_bol"] = match.group("bol").upper()
-        record["arrival_date"] = match.group("arrival")
-        record["foreign_port"] = match.group("country").strip()
-        record["quantity"] = " ".join(match.group("quantity").split())
-        record["weight"] = " ".join(match.group("weight").split())
-        record["shipper"] = normalize_company(match.group("supplier"))
-        record["consignee"] = normalize_company(match.group("importer"))
+        record["house_bol"] = match["bol"].upper()
+        record["arrival_date"] = match["arrival"]
+        record["foreign_port"] = match["country"].strip()
+        record["quantity"] = " ".join(match["quantity"].split())
+        record["weight"] = " ".join(match["weight"].split())
+        record["shipper"] = normalize_company(match["supplier"])
+        record["consignee"] = normalize_company(match["importer"])
         record["notify_party"] = record["consignee"]
-        record["commodity"] = " ".join(match.group("product").upper().split())
+        record["commodity"] = " ".join(match["product"].upper().split()).strip(" .")
         record["source_url"] = spec.source_url
+        return record
+
+    rows = []
+    for line in text_lines:
+        for match in pattern.finditer(line):
+            record = build_record(match.groupdict())
+            if is_relevant(record):
+                rows.append(record)
+    for index in range(0, max(0, len(text_lines) - 8)):
+        values = text_lines[index : index + 9]
+        if not (
+            rank_re.fullmatch(values[0])
+            and bol_re.fullmatch(values[1])
+            and importer_re.fullmatch(values[3])
+            and supplier_re.fullmatch(values[4])
+            and date_re.fullmatch(values[5])
+            and country_re.fullmatch(values[6])
+            and weight_re.fullmatch(values[7])
+            and quantity_re.fullmatch(values[8])
+        ):
+            continue
+        record = build_record(
+            {
+                "bol": values[1],
+                "product": values[2],
+                "importer": values[3],
+                "supplier": values[4],
+                "arrival": values[5],
+                "country": values[6],
+                "weight": values[7],
+                "quantity": values[8],
+            }
+        )
         if is_relevant(record):
             rows.append(record)
     return rows
@@ -350,7 +409,9 @@ def write_report(rows, output):
             "## Limitations",
             "",
             "Customs data is corroborating evidence, not a standalone confirmation. "
-            "GAME CONSOLE descriptions are medium confidence without other identifiers.",
+            "Product descriptions such as GAME CONSOLE, VIRTUAL REALITY DEVICES, "
+            "and WIRELESS PC CONTROLLER do not prove the final Valve retail SKU without "
+            "other identifiers.",
             "",
         ]
     )