Re: [Wikisource-l] Djvulibre routines into Labs
There are already a lot of data in the img_metadata field of the image table. I hope all data you are looking for are in it. Thomas Le 28 nov. 2013 à 23:57, Alex Brollo alex.bro...@gmail.com a écrit : I feel uncomfortable thinking to upload large files just to use a little bit of data... I presume that djvu are saved as bundled files, have you any new about saving them as indirect files, t.i. as single pages + an index file and some rather small pieces? Who could give me some detail about djvu files storage, and about projects to develop their management? I opened a bold bug in Bugzilla asking for some API actions bridging API and djvulibre routines; who is in your opinion an API developer which could be interested about such a rough idea? Alex 2013/11/28 Federico Leva (Nemo) nemow...@gmail.com Alex Brollo, 28/11/2013 20:36: I'll try to test some routines to manage both image and text layers of our itsource djvu files. My question: have I to upload them from Commons, or there's the possibility to access to them/to a copy of them into some folder of Labs without any need of uploading (painfully) a copy? No, labs doesn't have media. Only XML dumps and pageview stats. Download from upload.wikimedia.org is supposed to be rather fast though. Nemo ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l
Re: [Wikisource-l] Djvulibre routines into Labs
Thanks Thomas, but I'm looking for something much subtler: I look for mapped text of OCR with any possible detail - t.i. I need at least the output of djvutxt, djvudump, djvused - and obviously of a copy of djvu file. Presently I can't follow the wikidata adventure nor the metadata flow - I focus my interest on tools to help user while editing/formatting text of pages. Alex 2013/11/29 Thomas Tanon thoma...@hotmail.fr There are already a lot of data in the img_metadata field of the image table. I hope all data you are looking for are in it. Thomas Le 28 nov. 2013 à 23:57, Alex Brollo alex.bro...@gmail.com a écrit : I feel uncomfortable thinking to upload large files just to use a little bit of data... I presume that djvu are saved as bundled files, have you any new about saving them as indirect files, t.i. as single pages + an index file and some rather small pieces? Who could give me some detail about djvu files storage, and about projects to develop their management? I opened a bold bug in Bugzilla asking for some API actions bridging API and djvulibre routines; who is in your opinion an API developer which could be interested about such a rough idea? Alex 2013/11/28 Federico Leva (Nemo) nemow...@gmail.com Alex Brollo, 28/11/2013 20:36: I'll try to test some routines to manage both image and text layers of our itsource djvu files. My question: have I to upload them from Commons, or there's the possibility to access to them/to a copy of them into some folder of Labs without any need of uploading (painfully) a copy? No, labs doesn't have media. Only XML dumps and pageview stats. Download from upload.wikimedia.org is supposed to be rather fast though. Nemo ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l
Re: [Wikisource-l] Djvulibre routines into Labs
For these use cases I think that download the file is the best way to do it. It’s very quick because the connection between labs and the other Wikimedia clusters is very good. Thomas Le 29 nov. 2013 à 18:25, Alex Brollo alex.bro...@gmail.com a écrit : Thanks Thomas, but I'm looking for something much subtler: I look for mapped text of OCR with any possible detail - t.i. I need at least the output of djvutxt, djvudump, djvused - and obviously of a copy of djvu file. Presently I can't follow the wikidata adventure nor the metadata flow - I focus my interest on tools to help user while editing/formatting text of pages. Alex 2013/11/29 Thomas Tanon thoma...@hotmail.fr There are already a lot of data in the img_metadata field of the image table. I hope all data you are looking for are in it. Thomas Le 28 nov. 2013 à 23:57, Alex Brollo alex.bro...@gmail.com a écrit : I feel uncomfortable thinking to upload large files just to use a little bit of data... I presume that djvu are saved as bundled files, have you any new about saving them as indirect files, t.i. as single pages + an index file and some rather small pieces? Who could give me some detail about djvu files storage, and about projects to develop their management? I opened a bold bug in Bugzilla asking for some API actions bridging API and djvulibre routines; who is in your opinion an API developer which could be interested about such a rough idea? Alex 2013/11/28 Federico Leva (Nemo) nemow...@gmail.com Alex Brollo, 28/11/2013 20:36: I'll try to test some routines to manage both image and text layers of our itsource djvu files. My question: have I to upload them from Commons, or there's the possibility to access to them/to a copy of them into some folder of Labs without any need of uploading (painfully) a copy? No, labs doesn't have media. Only XML dumps and pageview stats. Download from upload.wikimedia.org is supposed to be rather fast though. Nemo ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l
Re: [Wikisource-l] Djvulibre routines into Labs
OK, I'll do. I hate to move many Mby around the web without a real and strong need but I hope to build some tools to help users while contributing, and this, IMHO, is one from the best justifications to use band and servers time. Alex 2013/11/29 Thomas Tanon thoma...@hotmail.fr For these use cases I think that download the file is the best way to do it. It’s very quick because the connection between labs and the other Wikimedia clusters is very good. Thomas Le 29 nov. 2013 à 18:25, Alex Brollo alex.bro...@gmail.com a écrit : Thanks Thomas, but I'm looking for something much subtler: I look for mapped text of OCR with any possible detail - t.i. I need at least the output of djvutxt, djvudump, djvused - and obviously of a copy of djvu file. Presently I can't follow the wikidata adventure nor the metadata flow - I focus my interest on tools to help user while editing/formatting text of pages. Alex 2013/11/29 Thomas Tanon thoma...@hotmail.fr There are already a lot of data in the img_metadata field of the image table. I hope all data you are looking for are in it. Thomas Le 28 nov. 2013 à 23:57, Alex Brollo alex.bro...@gmail.com a écrit : I feel uncomfortable thinking to upload large files just to use a little bit of data... I presume that djvu are saved as bundled files, have you any new about saving them as indirect files, t.i. as single pages + an index file and some rather small pieces? Who could give me some detail about djvu files storage, and about projects to develop their management? I opened a bold bug in Bugzilla asking for some API actions bridging API and djvulibre routines; who is in your opinion an API developer which could be interested about such a rough idea? Alex 2013/11/28 Federico Leva (Nemo) nemow...@gmail.com Alex Brollo, 28/11/2013 20:36: I'll try to test some routines to manage both image and text layers of our itsource djvu files. My question: have I to upload them from Commons, or there's the possibility to access to them/to a copy of them into some folder of Labs without any need of uploading (painfully) a copy? No, labs doesn't have media. Only XML dumps and pageview stats. Download from upload.wikimedia.org is supposed to be rather fast though. Nemo ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l
[Wikisource-l] Djvulibre routines into Labs
I (painfully) opened a bug into Bugzilla asking for djvulibre binaries installation into Labs. Any new about? Were they already installed and I asked unusefully? Are they - as I presume - a necessary tool for wikisource? Alex ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l
Re: [Wikisource-l] Djvulibre routines into Labs
Alex Brollo, 28/11/2013 11:38: I (painfully) opened a bug into Bugzilla asking for djvulibre binaries installation into Labs. Any new about? Were they already installed and I asked unusefully? Are they - as I presume - a necessary tool for wikisource? What's mising? https://bugzilla.wikimedia.org/show_bug.cgi?id=56972 is marked FIXED, please add there if you need something more. Nemo ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l
Re: [Wikisource-l] Djvulibre routines into Labs
I feel uncomfortable thinking to upload large files just to use a little bit of data... I presume that djvu are saved as bundled files, have you any new about saving them as indirect files, t.i. as single pages + an index file and some rather small pieces? Who could give me some detail about djvu files storage, and about projects to develop their management? I opened a bold bug in Bugzilla asking for some API actions bridging API and djvulibre routines; who is in your opinion an API developer which could be interested about such a rough idea? Alex 2013/11/28 Federico Leva (Nemo) nemow...@gmail.com Alex Brollo, 28/11/2013 20:36: I'll try to test some routines to manage both image and text layers of our itsource djvu files. My question: have I to upload them from Commons, or there's the possibility to access to them/to a copy of them into some folder of Labs without any need of uploading (painfully) a copy? No, labs doesn't have media. Only XML dumps and pageview stats. Download from upload.wikimedia.org is supposed to be rather fast though. Nemo ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l ___ Wikisource-l mailing list Wikisource-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikisource-l