Closed GoogleCodeExporter closed 9 years ago
here's the application log showing the bad behaviour:
--------------------------------------------------------------------------------
CV Scraper Version: 1.0.74
Running As: ComicRack Plugin (CR version 0.9.175)
Cache Directory: C:\Users\aza\AppData\Roaming\Comic Vine Scraper\localCache
Settings File: C:\Users\aza\AppData\Roaming\cYo\ComicRack\Scripts\Comic
Vine Scraper\settings.dat
--------------------------------------------------------------------------------
--------------------------------------------------------------------
[X] Series [X] Volume [X] Number
[X] Title [X] Published [X] Released
[X] Crossovers [X] Publisher [X] Imprint
[X] Writer [X] Penciller [X] Inker
[X] Colorist [X] Letterer [X] Cover Art
[X] Editor [X] Summary [X] Characters
[X] Teams [X] Locations [X] Webpage
-------------------------------------------------------------------
[X] Overwrite Existing [X] Ignore Blanks
[X] Convert Imprints [X] Autochoose Series
[X] Download Thumbs [X] Preserve Thumbs
[ ] Confirm Issues [X] Rescraping: Notes
[X] Fast Rescrape [X] Rescraping: Tags
[X] Summary Dialog
-------------------------------------------------------------------
Ignore folders when grouping issues into series.
Ignore all series published by 'abril'
Ignore all series published by 'marvel italia'
Ignore all series published by 'panini'
Ignore all series published by 'planeta deagostini'
Ignore all series published by 'marvel uk'
Ignore all series published by 'semic as'
Ignore all series published by 'panini comics'
-------------------------------------------------------------------
======> scraping next comic book: 'Moon Knight (2014) - 001 - [05-2014].cbz'
trying to match this book automatically...
...found a suitable match: Moon Knight (39957)
searching for the right issue in 'Moon Knight (39957)'
...identified issue number 1
querying comicvine for issue details...
setting values for this comic book ('*' = changed):
--> Series : Moon Knight
--> Issue Number : 1
--> *Title : Issue #1
--> Crossovers : --- skipped ---
--> *Summary : The wait is over! Moon Knight is here...like you've never
seen him bef ...
--> Release Date : 2011-5-4
--> *Publish Date : 2011-7-??
--> *Volume : 2011
--> Imprint : --- skipped ---
--> *Publisher : Marvel
--> *Characters : Captain America, Count Nefaria, Moon Knight, Mr. Hyde,
Spider-Man, Ult ...
--> *Teams : Avengers
--> *Locations : Los Angeles
--> *Writers : Brian Michael Bendis
--> *Pencillers : Alex Maleev
--> *Inkers : Alex Maleev
--> *Colorists : Matthew Wilson
--> *Letterers : Cory Petit
--> *CoverArtists : Bryan Hitch, Edgar Delgado, Humberto Ramos, Mark Texeira,
Paul Mounts, ...
--> *Editors : Axel Alonso, Lauren Sankovitch, Tom Brevoort
--> *Webpage :
http://www.comicvine.com/moon-knight-1-issue-1/4000-269480/
--> Rating : --- skipped ---
--> *Tags : CVDB269480
--> *Notes : Scraped metadata from ComicVine [CVDB269480].
--> *Issue Key : 269480
--> *Series Key : 39957
--> Cover Art URL :
http://static.comicvine.com/uploads/scale_small/6/67663/2515260-01a.jpg
Scraper terminated normally (scraped 1, skipped 0).
wrote debug logfile: cvs-debug-log-2014-03-13.txt
Original comment by cban...@gmail.com
on 14 Mar 2014 at 5:10
To make matters worse, I cannot seem to get this bug to occur on my own
computer. Even when I configure everything the same way (as shown in the
preferences in the application log posted above), on my computer the cover is
not recognized and I am given the option to choose the correct series.
(This is the 'correct' behaviour, since the scraper never automatically scrapes
issue 1 of any series if there are multiple available series with the same
name. This is to prevent accidentally matching a regular issue 1 to a TPD
issue 1 with the same cover.)
Original comment by cban...@gmail.com
on 14 Mar 2014 at 5:13
Well, I'm still stumped. I stepped through the code very carefully, and it
just seems to be working. :(
So I added a WHOLE BUNCH of debug statements to the code. Hopefully that will
help me isolate when the scraper is behaving differently on your end. Can you
trying installing this special build, and reproducing the error like you did
before?
https://docs.google.com/file/d/0BzIVSRpPwBMDQkQwZm8xcUx1Y0k/edit
This time, it should have a lot of extra information.
Original comment by cban...@gmail.com
on 17 Mar 2014 at 12:07
Ok. Here we go. First, my cache seems to be empty, so nothing coming from
there. I run the custom version you provide, and got the attached log while
again getting the wrong series... Hope this helps!
Original comment by theotoco...@gmail.com
on 17 Mar 2014 at 8:38
Attachments:
Thanks, that was very helpful. I've tracked the problem down to a freak
coincidence where the cover matching algorithm actually thinks these two covers
are about 85% the same:
http://static.comicvine.com/uploads/scale_large/0/40/3675846-1+moonkn2014001_dc1
1_lr.jpg
http://static.comicvine.com/uploads/scale_large/8/80205/1797285-mk01_cover03.jpg
At first glance, they don't look anything alike, but you have to realize that
the cover matching algorithm converts them to greyscale, and averages out the
brightnesses of each region of the cover (this allows it to match covers that
are similar except for different contrast ratios.)
Anyway, I tweaked the algorithm to be a bit more strict/careful about accepting
a match. This eliminates the problem you encountered, and hopefully any
similar problems in the future. Please let me know if it doesn't work for you!
The fix is available in Comic Vine Scraper 1.0.75.
Original comment by cban...@gmail.com
on 18 Mar 2014 at 3:54
Original issue reported on code.google.com by
cban...@gmail.com
on 14 Mar 2014 at 5:08