[books] Fix one-extra-scrobble bug on Koreader imports
Some checks failed
build / test (push) Has been cancelled

This commit is contained in:
2026-06-12 09:52:33 -04:00
parent 935d059a20
commit c71b51fdb8
4 changed files with 170 additions and 97 deletions

View File

@ -217,7 +217,6 @@ def build_scrobbles_from_book_map(book_map: dict, user: "User") -> list["Scrobbl
pages_not_found.append(book_id)
continue
should_create_scrobble = False
scrobble_page_data = {}
playback_position_seconds = 0
prev_page_stats = {}
@ -247,32 +246,39 @@ def build_scrobbles_from_book_map(book_map: dict, user: "User") -> list["Scrobbl
end_of_reading = pages_processed == total_pages_read
big_jump_to_this_page = (page_number - last_page_number) > 10
is_session_gap = seconds_from_last_page > SESSION_GAP_SECONDS
if (is_session_gap and not big_jump_to_this_page) or end_of_reading:
should_create_scrobble = True
should_create_scrobble = (
is_session_gap and not big_jump_to_this_page
) or end_of_reading
# Always accumulate the current page first
scrobble_page_data[page_number] = stats
if should_create_scrobble:
if not scrobble_page_data:
scrobble_page_data[page_number] = stats
# For end-of-reading, the current page is already accumulated
# and belongs in this scrobble. For a session gap, remove the
# current page from this scrobble — it starts a new session.
if is_session_gap and not end_of_reading:
del scrobble_page_data[page_number]
scrobble_page_data = dict(
sorted(
scrobble_page_data.items(),
key=lambda x: x[1]["start_ts"],
)
)
try:
first_page = scrobble_page_data.get(
list(scrobble_page_data.keys())[0]
)
last_page = scrobble_page_data.get(
list(scrobble_page_data.keys())[-1]
)
except IndexError:
if not scrobble_page_data:
logger.error(
"Could not process book, no page data found",
extra={"scrobble_page_data": scrobble_page_data},
)
continue
first_page = next(iter(scrobble_page_data.values()))
last_page = scrobble_page_data[
list(scrobble_page_data.keys())[-1]
]
timestamp = user.profile.get_timestamp_with_tz(
datetime.fromtimestamp(
int(first_page.get("start_ts"))
@ -319,90 +325,18 @@ def build_scrobbles_from_book_map(book_map: dict, user: "User") -> list["Scrobbl
timezone=tz,
)
)
# Then start over
should_create_scrobble = False
# Then start over for the next session
playback_position_seconds = 0
scrobble_page_data = {}
# We accumulate pages for the scrobble until we should create a new one
scrobble_page_data[page_number] = stats
# For session gaps, re-add the current page as the
# beginning of the next session's accumulation
if is_session_gap and not end_of_reading:
scrobble_page_data[page_number] = stats
last_page_number = page_number
prev_page_stats = stats
# Handle leftover pages that never triggered a session gap
if scrobble_page_data:
scrobble_page_data = dict(
sorted(
scrobble_page_data.items(),
key=lambda x: x[1]["start_ts"],
)
)
try:
first_page = scrobble_page_data.get(
list(scrobble_page_data.keys())[0]
)
last_page = scrobble_page_data.get(
list(scrobble_page_data.keys())[-1]
)
except IndexError:
logger.error(
"Could not process book, no page data found",
extra={"scrobble_page_data": scrobble_page_data},
)
continue
playback_position_seconds = sum(
p["duration"] for p in scrobble_page_data.values()
)
timestamp = user.profile.get_timestamp_with_tz(
datetime.fromtimestamp(
int(first_page.get("start_ts"))
)
)
stop_timestamp = user.profile.get_timestamp_with_tz(
datetime.fromtimestamp(
int(last_page.get("end_ts"))
)
)
scrobble = Scrobble.objects.filter(
timestamp=timestamp,
book_id=book_id,
user_id=user.id,
).first()
if not scrobble:
logger.info(
f"Queueing scrobble for {book_id}, page {list(scrobble_page_data.keys())[0]}"
)
log_data = {
"koreader_hash": book_dict.get("hash"),
"page_data": scrobble_page_data,
"pages_read": len(scrobble_page_data.keys()),
}
if hasattr(timestamp.tzinfo, "tzname"):
tz = timestamp.tzinfo.tzname
if hasattr(timestamp.tzinfo, "name"):
tz = timestamp.tzinfo.name
scrobbles_to_create.append(
Scrobble(
book_id=book_id,
user_id=user.id,
source="KOReader",
media_type=Scrobble.MediaType.BOOK,
timestamp=timestamp,
log=log_data,
stop_timestamp=stop_timestamp,
playback_position_seconds=playback_position_seconds,
in_progress=False,
played_to_completion=True,
long_play_complete=False,
timezone=tz,
)
)
if pages_not_found:
logger.info(f"Pages not found for books: {set(pages_not_found)}")
return scrobbles_to_create