mapillary_download of a sequence creates 0 Bytes files #3
Loading…
Reference in New Issue
Block a user
No description provided.
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
when using the mapillary_download.py to get a sequence, about 30-40% of downloaded pictues are 0 Bytes (tested with multiple sequences, but all from me and the same day)
hello, thanks for this report. other people are having trouble getting some batches of photos, we did not found why it happens, usually it is not as often as what you report, more around 5%.
It suggests more of a problem on the mapillary side, are these pictures of the sequence still visible on mapillary website ?
some more detail:
find . -name '*.jpg' -size 0 -print0 | xargs -0 rm
), they are re-created with 0 Bytes when running the script againif you want to compare, the screenshot is from sequence
R59T5DwBwa9W8jH6LJs0xQ
From the API response, I could not see anything suspicious:
Download with wget works for both
thumb_original_url
s, filesize is in the range.traced it down to:
Error for all 0 Bytes files:
Directory Photo: IFD exceeds data buffer, cannot read next pointer.
Did not cross-check, but at least for this picture failing, there is a funny annotation in exif-data:
I have added two manually downloaded pictures, maybe it is easier to debug the exif extraction with these files:
Directory Photo: IFD exceeds data buffer, cannot read next pointer.
so... pyexiv2.ImageData seems to crash if jfif tags are present
I am not very experienced with this, but found a workaround by removing these with pillow.
Better solutions welcome...