Fix Weissman recipe time parsing

This commit is contained in:
Fizzlepoof
2026-05-22 14:23:34 +00:00
parent 9a24dd11ad
commit 1631dcd1c1
3 changed files with 49 additions and 1 deletions

View File

@@ -295,6 +295,10 @@ def enrich_recipe_node_from_dom(recipe_node: dict[str, Any], html: str) -> dict[
DURATION_RE = re.compile(r'^P(?:T(?:(?P<hours>\d+)H)?(?:(?P<minutes>\d+)M)?(?:(?P<seconds>\d+)S)?)?$')
TEXT_DURATION_PART_RE = re.compile(
r'(?P<value>\d+(?:\.\d+)?)\s*(?P<unit>hours?|hrs?|hr|minutes?|mins?|min|seconds?|secs?|sec|s)\b',
re.IGNORECASE,
)
def parse_duration_minutes(value: Any) -> int | None:
@@ -314,6 +318,20 @@ def parse_duration_minutes(value: Any) -> int | None:
seconds = int(m.group('seconds') or 0)
total = hours * 60 + minutes + (1 if seconds >= 30 else 0)
return total or None
textual_matches = list(TEXT_DURATION_PART_RE.finditer(value))
if textual_matches:
total_minutes = 0.0
for part in textual_matches:
amount = float(part.group('value'))
unit = part.group('unit').lower()
if unit.startswith(('hour', 'hr')):
total_minutes += amount * 60
elif unit.startswith(('minute', 'min')):
total_minutes += amount
else:
total_minutes += amount / 60
rounded = int(total_minutes + 0.5)
return rounded or None
num = re.search(r'(\d+)', value)
return int(num.group(1)) if num else None