Test a new data source for sselph/scraper

DrMaxwell · 22 Sept 2016, 07:50

Nice to see you back at it and screenzone.fr are a great source for the databases. They have some excellent tools for metadata and media on their website.

Download Scraper for your respective OS

Roms Move or copy the scraper to the roms subdirectory for the system you want to scrape. For example: //roms/snes/
ImagesPlace matched images into system subfolder named images. For example: //roms//snes/images/
Use the command below to scrape metadata and match images without downloading images and keeping them in png format that Raspberry Pi plays nicely with.

scraper -add_not_found=true -append=true -download_images=false -image_dir="images" -image_path="images" -image_suffix="" -img_format="png" -no_thumb=true -thumb_only=false -thumb_suffix="" -use_gdb=true -use_ss=true -use_ovgdb=true -workers=4

lilbud · 22 Sept 2016, 10:39

Will you add the libretro thumbnails as a scraper source?

Pyjamarama · 22 Sept 2016, 11:27

Does SScraper (running in Windows) need a special parameter for FBA? (for MAME the -mame parameter is used)

sselph · 22 Sept 2016, 11:38

@DrMaxwell Nice to see you again.

@lilbud https://github.com/sselph/scraper/issues/102 The issue was that I assumed data sources would handle both metadata and images so having a source just for images was something that requires more work.

@Pyjamarama With my limited knowledge of MAME and arcade I called the flag -mame but it goes for all arcade systems like fba, neogeo, etc.

DrMaxwell · 22 Sept 2016, 12:07

@sselph I assume you have documented any and all new options in the help section of the program:

scraper --help

How can people access 3D versus 2D versions of art from the screenzone database?
I've noticed if all three databases are set to true, the metadata is more complete; is there an order or hierarchical preference that the scraper pulls this metadata from?

Thanks for the further development of this tool. Your scraper is the fastest in the business!

sselph · 22 Sept 2016, 12:35

Yes -help should have it. -console_img="b" is the default which is 2D boxart -console_img="3b"would get the 3D boxart. Right now the order is gdb, ss, ovgdb but I need to rework some of this to handle specifying the order.

DrMaxwell · 22 Sept 2016, 12:48

@sselph I think having a preference would be good. Silly question, but does the scraper check against all three sequentially; for example, if the first doesn't provide the metadata, then it goes to the next database until both image and metadata are found?

sselph · 3 Oct 2016, 11:42

@DrMaxwell Thats the idea but it might stop without getting an image and only metadata. I can't remember.

tien_huu_1408 · 24 Sept 2016, 05:55

@sselph I like this mix template for scraping. How could you do this? Could you integrate into your script option? Thanks a lot!

tien_huu_1408 · 1 Oct 2016, 09:12

@sselph Any update for the script?

paradadf · 3 Oct 2016, 11:42

@sselph it is amazing how fast SS as a source is!!!

I notice two things while using it though...

I can't scrape any arcade games (mame, fba, neogeo)
With one specific gba rom I get always the following error message and the scraper stops:

2016/10/03 13:36:15 INFO: Checking for new hash.csv.
2016/10/03 13:36:15 INFO: hash.csv "e2315a44c3d4c5d200477a109e079f29" up to date.
2016/10/03 13:36:15 INFO: Starting: gba\2 Game Pack! - Matchbox Missions - Emergency Response + Air, Land and Sea Rescue (USA).zip
2016/10/03 13:36:16 ERR: error processing gba\2 Game Pack! - Matchbox Missions - Emergency Response + Air, Land and Sea Rescue (USA).zip: image: unknown format
2016/10/03 13:36:16 INFO: Starting: gba\2 Game Pack! - Matchbox Missions - Emergency Response + Air, Land and Sea Rescue (USA).zip
2016/10/03 13:36:16 ERR: error processing gba\2 Game Pack! - Matchbox Missions - Emergency Response + Air, Land and Sea Rescue (USA).zip: image: unknown format
2016/10/03 13:36:16 INFO: Starting: gba\2 Game Pack! - Matchbox Missions - Emergency Response + Air, Land and Sea Rescue (USA).zip
2016/10/03 13:36:16 ERR: error processing gba\2 Game Pack! - Matchbox Missions - Emergency Response + Air, Land and Sea Rescue (USA).zip: image: unknown format
panic: runtime error: invalid memory address or nil pointer dereference
[signal 0xc0000005 code=0x0 addr=0x20 pc=0x407bc7]

goroutine 28 [running]:
panic(0x81cd20, 0xc0420040b0)
        /usr/local/go/src/runtime/panic.go:500 +0x1af
main.crawlROMs.func2(0xc04340c010, 0xc043408000, 0x0, 0xc042a8dfe0, 0xc0425e9980, 0xc042a971c0)
        /home/sselph/go/src/github.com/sselph/scraper/scraper.go:281 +0x247
created by main.crawlROMs
        /home/sselph/go/src/github.com/sselph/scraper/scraper.go:314 +0x8c7

lilbud · 3 Oct 2016, 14:01

Can you enable scraping games that are on an external drive.

DrMaxwell · 3 Oct 2016, 21:18

@lilbud Just put the scraper executable in the same folder as your roms; whether or not the drive is external is irrelevant.

sselph · 4 Oct 2016, 00:30

@paradadf Thanks. I found the issue and have it fixed on github. I wasn't passing the error returned when the bad image was found and caused it to crash when using -missing. I can't reproduce the image error. Maybe that game was fixed? or it was temporary. I also only set this up to scrape console games. I didn't do arcade since it uses names instead of hashes and hadn't done any testing on that.

paradadf · 4 Oct 2016, 13:29

@sselph the rom works perfectly, so I don't understand what's the problem.
Anyway, I just noticed that the letters with accents á, é, ó, etc... aren't being shown on the xml file (those letters are just erased).

Nightfire · 4 Oct 2016, 18:17

Hi @sselph, happy to hear you're working on the scraper again, I've used it every time I've worked with RetroPie. Not sure if you had seen recent development news of RetroPie, but they are working on emulators for PSP (ppsspp), Nintendo DS (desmune) and GameCube (dolphin). So, I was wondering, will you be adding support for scraping PSP/NDS/GameCube games? Or is there a way to scrape for those systems that I'm just not aware of?

robertybob · 5 Oct 2016, 11:24

@Nightfire said in Test a new data source for sselph/scraper:

Hi @sselph, happy to hear you're working on the scraper again, I've used it every time I've worked with RetroPie. Not sure if you had seen recent development news of RetroPie, but they are working on emulators for PSP (ppsspp), Nintendo DS (desmune) and GameCube (dolphin). So, I was wondering, will you be adding support for scraping PSP/NDS/GameCube games? Or is there a way to scrape for those systems that I'm just not aware of?

PSP:
https://github.com/sselph/scraper/issues/52

NDS:
https://github.com/sselph/scraper/issues/28

Gamecube would be possible, I think, as the hashes are listed on redump

Nightfire · 5 Oct 2016, 17:28

@robertybob ahh, ok, thanks for the info, guess I never went to the 'issues' when I was there.

spannernick1 · 5 Oct 2016, 21:19

Will this scraper have UK box art as well,the ones that come with ES are not good at finding Mega Drive and CD Pal box art,had to end up adding them manually ,sometimes it adds a completely different game when its not that game.

robertybob · 6 Oct 2016, 13:44

@spannernick1 Don't bother with the ES scraper, use Sselph's one.

Sselph's scraper should match (Europe) roms to Mega Drive games and (USA) roms to Genesis games. If not, first check whether the image uploaded to thegamesdb.net is actually for the correct region (more often that not it isn't). http://thegamesdb.net/browse/36/ If the Mega Drive game is listed there and the artwork is for the Europe copy, then check which region your roms are.

Failing that, segaretro.org has all the best quality Mega Drive scans :)