Files
fabula-ultima-html/scripts/strip_watermark.py
2026-06-06 13:03:00 +00:00

131 lines
4.1 KiB
Python

#!/usr/bin/env python3
"""
Remove "Guest Customer (Order #52072168)" watermark artifacts from all book pages.
Strategy:
- If a line's text content (with HTML tags stripped) consists entirely of
watermark text, remove the whole line.
- If watermark text is embedded within a line that has other content, strip
just the watermark portion and tidy the surrounding punctuation.
"""
import glob
import os
import re
HTML_DIR = os.path.join(os.path.dirname(__file__), "..", "books")
# Matches the watermark text in all the forms it appears
WATERMARK_RE = re.compile(
r"(?:"
r"Guest\s+Customer\b[^\n<]*?" # "Guest Customer" + trailing text
r"|Order\s+\w[^\n<]*?52072168" # "Order <label>: 52072168"
r"|Order\s*[#:]\s*:?\s*52072168" # "Order #52072168", "Order #: 52072168"
r"|#\s*52072168" # "#52072168" standalone
r"|\b52072168\b" # bare order number
r"|\b38246845\b"
r")",
re.IGNORECASE,
)
TAG_RE = re.compile(r"<[^>]+>")
# Punctuation and decoration that can be left behind after stripping
DECORATION_RE = re.compile(r"^[\s\-–—|,.:;!?()\[\]*&#\d/\\]+$")
def text_content(line: str) -> str:
"""Return the visible text of a line with HTML tags removed."""
return TAG_RE.sub("", line)
def is_watermark_only(line: str) -> bool:
"""True when the line's entire visible text is watermark content."""
text = text_content(line).strip()
remaining = WATERMARK_RE.sub("", text).strip()
return DECORATION_RE.match(remaining) is not None or remaining == ""
def strip_watermark_inline(line: str) -> str:
"""Remove watermark text from a line that has other real content."""
result = WATERMARK_RE.sub("", line)
# Tidy decoration left behind after removal (e.g. "— ", " —", " | ", "---")
result = re.sub(r"\s*[—–\-]{1,3}\s*$", "", result, flags=re.MULTILINE)
result = re.sub(r"^\s*[—–\-]{1,3}\s*", "", result, flags=re.MULTILINE)
result = re.sub(r"\|\s*$", "", result, flags=re.MULTILINE)
result = re.sub(r"^\s*\|\s*", "", result, flags=re.MULTILINE)
result = re.sub(r"\s{2,}", " ", result)
return result
def process_file(filepath: str) -> tuple[int, int]:
with open(filepath, encoding="utf-8") as f:
lines = f.readlines()
new_lines: list[str] = []
removed = 0
stripped = 0
for line in lines:
if not WATERMARK_RE.search(line):
new_lines.append(line)
continue
if is_watermark_only(line.strip()):
removed += 1
# Keep the newline gap only if needed for readability; skip blank result
else:
new_lines.append(strip_watermark_inline(line))
stripped += 1
with open(filepath, "w", encoding="utf-8") as f:
f.writelines(new_lines)
return removed, stripped
def main() -> None:
def sort_key(p):
m = re.search(r"(\d+)", os.path.basename(p))
return int(m.group(1)) if m else -1
from pathlib import Path
html_files = []
for root, dirs, files in Path(HTML_DIR).walk():
hf = filter(lambda fn: fn.endswith(".html"), files)
hf = [ root / fn for fn in hf ]
html_files.extend(hf)
# html_files = sorted(
# glob.glob(os.path.join(HTML_DIR, "*.html")),
# key=sort_key,
# )
breakpoint()
html_files = sorted(html_files, key=sort_key)
total_removed = 0
total_stripped = 0
affected = 0
for filepath in html_files:
removed, stripped = process_file(filepath)
if removed or stripped:
affected += 1
name = os.path.basename(filepath)
parts = []
if removed:
parts.append(f"{removed} line{'s' if removed != 1 else ''} removed")
if stripped:
parts.append(f"{stripped} inline")
print(f" {name}: {', '.join(parts)}")
total_removed += removed
total_stripped += stripped
print(
f"\nDone. {affected} files changed — "
f"{total_removed} lines removed, {total_stripped} inline occurrences stripped."
)
if __name__ == "__main__":
main()