Jako wrote:Hi Pendor! congrats for the scrapper, is really really usefull.
From yesterday i cheked every gamelist.xml, i find the problem diferent that ago a month (a update between?), sorry but i didn't saved any xml with garbadge at the finish, i prefered to fix them and enjoy the few artwork saved in the list.
But if example, now i only find a incomplete last line, something like this: (i cut at the half to don't make it too much longer to show...)
Obviously, the problem (i think) is that some rom name character make es-scrapper exit, yes, it exits, don't follow with the next system listed on es_system.cfg. I never downloaded all the system in a row, my better results is one by one. The ironic is that i have most of the needed homework (excepting mame, i will ask something later).
That's weird, I've never seen anything like that.
The scraper saves the XML file after all possible games for that console have been found.
But yeah, such characters in a filename, that's never a good practice. I'll look into that.
Jako wrote:---Near Full artbox of that systems:
-Game Boy Advance (Full)
-Master System 2 (Full)
-Nes (Full 95% accuracy of the full romser, awesome!)
-Snes (Matching names in 70%, the rest change names and artbox, I have U.N. Squadron 20 times listed)
---Systems with few available/found artbox:
-Atari 2600 (4/100)
-Game Boy Color(0/~200, but all the artwork is downloaded and mostly match)
-Pc-Engine (don't find anything)
-Mame: very few but awesome artwork (20/~200), many jewels to the sight, but i prefer to see the snapshots in mame, in my mame arcade cabinet i have all the snaps, i wish to use them on retroarch, i think that this not have to be too many dificult, and less if the mame roms and the snaps always match on names.
I've had little success with Atari 2600 but haven't renamed the files yet so it might be that. Assuming all of your files have proper names, remember that thegamesdb might be missing stuff.
Then again, when I test the scraper I use a small number of games. Which brings me to my next point..
Scraping full collections or romsets is gonna take a long time, its just easier to distribute a complete list and check what's missing. I've been thinking for a while about adapting no-intro dat files to ES and merging it with boxarts packs (when available). That would require:
a) ES identifying stuff via checksums
b) Using some audit tool to rename your collection (if needed) so it matches the filenames on the list.
Jako wrote:---Learned things watching es-scrapper running:
-It works in order of stored files or folders on the sd card or medium, not alphabetically. (This is why i don't know what kind of roms or names make the problem, i never know what would follows.)
-The first option of es-scrapper (in retropie setup) is supossed that skip the forced search and storing into the list of older roms, but not always work, and anothers makes a check and don't find nothing when a folder is full off roms and pictures, it answers with "no new games added".
-The second the same that the first but force a deep search and rewrite of the gamelist (and download all again).
-The third (for what i know) is a CRC based search from another database, this is form me the worst option, overwrite the existing .xml in some lines and changes many names and pics, with a very low accuracy, is the less used option for me.
-Es-scrapper hates the "#" character, and more i think.
I dunno about that, I don't really use the script and I don't know what parameters is it using.
-f forces the script to re-scrape (and overwrite) everything, regular scraping will check for files that are not on the list and add them (if identified) to the existing file.
CRC uses another source (http://archive.vg/) and compares the checksum to the one on the database. And yeah, it's kinda slower and accuracy depends on how "clean" the files are.
Jako wrote:I found where is the es-scrapper.py (or something like this), but, how to run it in command mode? "sh" don't do nothing and i never executed something with python.
- Code: Select all
Use -h to see a list of possible arguments.