mirror of
https://github.com/yogeshojha/rengine.git
synced 2026-10-18 22:33:24 +02:00
Yogesh Ojha
cd5b21e972
Finding titles become Third-party content on hostname and Dangling A record, their descriptions become label and value lines, and impact and remediation state one fact each. The internal address check, the fixable flag, the tracking kind and the referenced hostname messages drop their explanatory wording.
37 lines
1.4 KiB
Python
37 lines
1.4 KiB
Python
"""Analytics, tag manager and ads account IDs a page loads."""
|
|
|
|
from __future__ import annotations
|
|
|
|
import re
|
|
|
|
MAX_TRACKING_IDS = 20
|
|
BODY_SCAN_BYTES = 131_072
|
|
|
|
_QUOTED = r"""(?<=[\"'=/])"""
|
|
_PATTERNS: tuple[tuple[re.Pattern[str], str], ...] = (
|
|
(re.compile(_QUOTED + r"UA-(\d{4,10})-\d{1,4}(?![\w-])"), "UA-{}"),
|
|
(re.compile(_QUOTED + r"(G-[A-Z0-9]{8,12})(?![\w-])"), "{}"),
|
|
(re.compile(_QUOTED + r"(GTM-[A-Z0-9]{4,9})(?![\w-])"), "{}"),
|
|
(re.compile(_QUOTED + r"(AW-\d{8,12})(?![\w-])"), "{}"),
|
|
(re.compile(r"(ca-pub-\d{10,20})(?!\d)"), "{}"),
|
|
(re.compile(r"""fbq\(\s*['"]init['"]\s*,\s*['"](\d{10,20})['"]"""), "fb:{}"),
|
|
(re.compile(r"hjid\s*:\s*(\d{5,10})"), "hotjar:{}"),
|
|
(re.compile(r"clarity\.ms/tag/([a-z0-9]{8,12})", re.IGNORECASE), "clarity:{}"),
|
|
(re.compile(r"\bym\(\s*(\d{5,10})\s*,"), "ym:{}"),
|
|
(re.compile(r"js\.hs-scripts\.com/(\d{5,10})\.js"), "hubspot:{}"),
|
|
)
|
|
|
|
|
|
def tracking_ids(body: str | None) -> list[str]:
|
|
"""Account IDs in a page, Universal Analytics properties folded to the account."""
|
|
if not body:
|
|
return []
|
|
text = body[:BODY_SCAN_BYTES]
|
|
found: dict[str, None] = {}
|
|
for pattern, shape in _PATTERNS:
|
|
for match in pattern.findall(text):
|
|
found.setdefault(shape.format(match), None)
|
|
if len(found) >= MAX_TRACKING_IDS:
|
|
return sorted(found)
|
|
return sorted(found)
|