Hello all,
I am trying to develop a chrome extension to work with wikipedia.
Started exploring javascript for login api.
Got this example.
https://www.mediawiki.org/wiki/Example_login_code_in_JS_%28using_JQuery%29
Stored the code as test.html and opened in chrome. Got the following error.
No 'Access-Control-Allow-Origin' header is present on the requested resource. Origin 'null' is therefore not allowed access.
How to solve this error?
https://www.mediawiki.org/wiki/Manual:CORS The examples here are not helpful.
Please guide me on how to login to mediawiki via javascript?
Thanks.
Firstly, I'm afraid that page might not be up to date; there have been some changes to modernize the login API recently, and that page was last updated in 2013. The canonical documentation is https://www.mediawiki.org/wiki/API:Login (in English).
Secondly, login API is mostly intended for scripts running in the context of the wiki you want to log in to, not for scripts running on other pages or browser extensions. I'm not an expert on this, but I think you'll want to look into using OAuth: https://www.mediawiki.org/wiki/OAuth.
On Wed, Oct 26, 2016 at 12:26 PM, Bartosz Dziewoński matma.rex@gmail.com wrote:
Firstly, I'm afraid that page might not be up to date; there have been some changes to modernize the login API recently, and that page was last updated in 2013. The canonical documentation is < https://www.mediawiki.org/wiki/API:Login%3E (in English).
Secondly, login API is mostly intended for scripts running in the context of the wiki you want to log in to, not for scripts running on other pages or browser extensions. I'm not an expert on this, but I think you'll want to look into using OAuth: https://www.mediawiki.org/wiki/OAuth.
OAuth is unfortunately hard to do in a browser plugin, as its security relies on the assumption that the application secret is only known to the developer.
It's not easy to give useful advice without knowing what is it you are trying to build (see the XY problem - http://meta.stackexchange.com/questions/66377/what-is-the-xy-problem ) or at least why the CORS manual was not helpful.
As a side note, chrome extensions can authorize a domain and use it as if they're requesting while on the site, it seems, to prevent cors errors.
https://developer.chrome.com/extensions/declare_permissions
That may require that your user login directly in Wikipedia rather than through your extension - but once the login is recognized, the API calls pass through without CORS issues, as long as you add the URL format to your permission declaration in the manifest.
I've created a small extension testing out some Echo functionality using exactly that. The API calls were successful, the only issue I had was that the user had to be logged into Wikipedia already.
This may not be a perfect solution, but it might save you the trouble of working with OAuth and doing the login yourself.
On Thu, Oct 27, 2016 at 2:34 PM, Gergo Tisza gtisza@wikimedia.org wrote:
On Wed, Oct 26, 2016 at 12:26 PM, Bartosz Dziewoński matma.rex@gmail.com wrote:
Firstly, I'm afraid that page might not be up to date; there have been some changes to modernize the login API recently, and that page was last updated in 2013. The canonical documentation is < https://www.mediawiki.org/wiki/API:Login%3E (in English).
Secondly, login API is mostly intended for scripts running in the context of the wiki you want to log in to, not for scripts running on other pages or browser extensions. I'm not an expert on this, but I think you'll want to look into using OAuth: https://www.mediawiki.org/wiki/OAuth.
OAuth is unfortunately hard to do in a browser plugin, as its security relies on the assumption that the application secret is only known to the developer.
It's not easy to give useful advice without knowing what is it you are trying to build (see the XY problem - http://meta.stackexchange.com/questions/66377/what-is-the-xy-problem ) or at least why the CORS manual was not helpful. _______________________________________________ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Gergo Tisza,
Recently, we added more then 2000 books to Tamil WikiSource using Google OCR with OCR4WikiSource https://github.com/tshrinivasan/OCR4wikisource
Most of the books have 30-40% spell errors. To fix them, editing the pages manually is not easy.
I am thinking of a browser plugin with the following workflow.
1. user logins to tamil wikisource 2. reads a wiki page for any book 3. Finds a word with spelling error 4. double clicks it 5. Inline textbox opens there 6. Writes correct word and click ok button 7. The change is saved automatically in background 8. The old word and new word are sent to remote server and stored 9. Mass find and replace for the stored wrong,correct words are executed periodically
With this design plan, hope users can easily contribute to fix the spell errors.
Hope there may be even better solutions. Share your thoughts on this.
Now, I am in point 1. Cant find any javascript code login to wikipedia without getting No 'Access-Control-Allow-Origin' header is present error.
The CORS link is not useful for beginner.
Example code in javascript to login to wikipedia will be much helpful.
Moriel Schottlender,
Your solution is fine. Please share the code so that I can extend from there.
Thanks.
This sounds like a great idea (something akin to PGDP's WordCheck tool http://www.pgdp.net/wiki/WordCheck?) But I wonder if implementing it as a Gadget would be easier? That way, the user is already logged in, and can access all the normal APIs etc.
The word lists (both good and bad?) could be saved as subpages of the Index page. Could probably make sense to have site-wide word lists too, for all the really common ones.
Do you have an idea of the interface for correcting words (the right-click drop down list)?
This could be really useful for all language Wikisources I think!
—Sam
On 3 November 2016 at 06:11, Shrinivasan T tshrinivasan@gmail.com wrote:
Gergo Tisza,
Recently, we added more then 2000 books to Tamil WikiSource using Google OCR with OCR4WikiSource https://github.com/tshrinivasan/OCR4wikisource
Most of the books have 30-40% spell errors. To fix them, editing the pages manually is not easy.
I am thinking of a browser plugin with the following workflow.
- user logins to tamil wikisource
- reads a wiki page for any book
- Finds a word with spelling error
- double clicks it
- Inline textbox opens there
- Writes correct word and click ok button
- The change is saved automatically in background
- The old word and new word are sent to remote server and stored
- Mass find and replace for the stored wrong,correct words are
executed periodically
With this design plan, hope users can easily contribute to fix the spell errors.
Hope there may be even better solutions. Share your thoughts on this.
Now, I am in point 1. Cant find any javascript code login to wikipedia without getting No 'Access-Control-Allow-Origin' header is present error.
The CORS link is not useful for beginner.
Example code in javascript to login to wikipedia will be much helpful.
Moriel Schottlender,
Your solution is fine. Please share the code so that I can extend from there.
Thanks.
-- Regards, T.Shrinivasan
My Life with GNU/Linux : http://goinggnu.wordpress.com Free E-Magazine on Free Open Source Software in Tamil : http://kaniyam.com
Get Free Tamil Ebooks for Android, iOS, Kindle, Computer : http://FreeTamilEbooks.com
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Thank you very much for this idea!
+1 for a gadget or maybe, in the future, a part of the ProofreadPage extension.
An issue that we probably need to tackle when designing this tool: what about not-existing words that could be a typo of two different words? If we replace all instances of the typo by one of the two word, we will change the meaning of the text by introducing a worst typo when the other word is the correct one.
Thomas
Le 3 nov. 2016 à 01:34, Sam Wilson swilson@wikimedia.org a écrit :
This sounds like a great idea (something akin to PGDP's WordCheck tool http://www.pgdp.net/wiki/WordCheck?) But I wonder if implementing it as a Gadget would be easier? That way, the user is already logged in, and can access all the normal APIs etc.
The word lists (both good and bad?) could be saved as subpages of the Index page. Could probably make sense to have site-wide word lists too, for all the really common ones.
Do you have an idea of the interface for correcting words (the right-click drop down list)?
This could be really useful for all language Wikisources I think!
—Sam
On 3 November 2016 at 06:11, Shrinivasan T tshrinivasan@gmail.com wrote:
Gergo Tisza,
Recently, we added more then 2000 books to Tamil WikiSource using Google OCR with OCR4WikiSource https://github.com/tshrinivasan/OCR4wikisource
Most of the books have 30-40% spell errors. To fix them, editing the pages manually is not easy.
I am thinking of a browser plugin with the following workflow.
- user logins to tamil wikisource
- reads a wiki page for any book
- Finds a word with spelling error
- double clicks it
- Inline textbox opens there
- Writes correct word and click ok button
- The change is saved automatically in background
- The old word and new word are sent to remote server and stored
- Mass find and replace for the stored wrong,correct words are
executed periodically
With this design plan, hope users can easily contribute to fix the spell errors.
Hope there may be even better solutions. Share your thoughts on this.
Now, I am in point 1. Cant find any javascript code login to wikipedia without getting No 'Access-Control-Allow-Origin' header is present error.
The CORS link is not useful for beginner.
Example code in javascript to login to wikipedia will be much helpful.
Moriel Schottlender,
Your solution is fine. Please share the code so that I can extend from there.
Thanks.
-- Regards, T.Shrinivasan
My Life with GNU/Linux : http://goinggnu.wordpress.com Free E-Magazine on Free Open Source Software in Tamil : http://kaniyam.com
Get Free Tamil Ebooks for Android, iOS, Kindle, Computer : http://FreeTamilEbooks.com
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
-- User:SWilson (WMF) https://meta.wikimedia.org/wiki/User:SWilson_(WMF) Community Tech Fremantle, Western Australia _______________________________________________ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Yes, I'm not sure about the automatic replacement of words in pages other than the current one — something more along the lines of a spelling-checker, maybe? Where the dictionary can be customised per-work. Is Extension:SpellingDictionary https://www.mediawiki.org/wiki/Extension:SpellingDictionary worth looking into? But no, perhaps not, because if such a thing were built as part of ProofreadPage, it could handle scannos as well (punctuation etc.; i.e. be more specialised than just spelling). For example, the PGDP WordCheck system presents the user with a syntax-highlighted view of each page prior to saving, highlighting *all* punctuation and spaces and newlines, and making each miss-spelled word a dropdown or suggest box.
—sam
On 3 November 2016 at 18:04, Thomas PT thomaspt@hotmail.fr wrote:
Thank you very much for this idea!
+1 for a gadget or maybe, in the future, a part of the ProofreadPage extension.
An issue that we probably need to tackle when designing this tool: what about not-existing words that could be a typo of two different words? If we replace all instances of the typo by one of the two word, we will change the meaning of the text by introducing a worst typo when the other word is the correct one.
Thomas
Le 3 nov. 2016 à 01:34, Sam Wilson swilson@wikimedia.org a écrit :
This sounds like a great idea (something akin to PGDP's WordCheck tool http://www.pgdp.net/wiki/WordCheck?) But I wonder if implementing it
as a
Gadget would be easier? That way, the user is already logged in, and can access all the normal APIs etc.
The word lists (both good and bad?) could be saved as subpages of the
Index
page. Could probably make sense to have site-wide word lists too, for all the really common ones.
Do you have an idea of the interface for correcting words (the
right-click
drop down list)?
This could be really useful for all language Wikisources I think!
—Sam
On 3 November 2016 at 06:11, Shrinivasan T tshrinivasan@gmail.com
wrote:
Gergo Tisza,
Recently, we added more then 2000 books to Tamil WikiSource using Google OCR with OCR4WikiSource https://github.com/tshrinivasan/OCR4wikisource
Most of the books have 30-40% spell errors. To fix them, editing the pages manually is not easy.
I am thinking of a browser plugin with the following workflow.
- user logins to tamil wikisource
- reads a wiki page for any book
- Finds a word with spelling error
- double clicks it
- Inline textbox opens there
- Writes correct word and click ok button
- The change is saved automatically in background
- The old word and new word are sent to remote server and stored
- Mass find and replace for the stored wrong,correct words are
executed periodically
With this design plan, hope users can easily contribute to fix the spell errors.
Hope there may be even better solutions. Share your thoughts on this.
Now, I am in point 1. Cant find any javascript code login to wikipedia without getting No 'Access-Control-Allow-Origin' header is present error.
The CORS link is not useful for beginner.
Example code in javascript to login to wikipedia will be much helpful.
Moriel Schottlender,
Your solution is fine. Please share the code so that I can extend from there.
Thanks.
-- Regards, T.Shrinivasan
My Life with GNU/Linux : http://goinggnu.wordpress.com Free E-Magazine on Free Open Source Software in Tamil :
Get Free Tamil Ebooks for Android, iOS, Kindle, Computer : http://FreeTamilEbooks.com
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
-- User:SWilson (WMF) https://meta.wikimedia.org/wiki/User:SWilson_(WMF) Community Tech Fremantle, Western Australia _______________________________________________ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On Thu, Nov 3, 2016 at 9:24 PM, Sam Wilson swilson@wikimedia.org wrote:
[...] For example, the PGDP WordCheck system presents the user with a syntax-highlighted view of each page prior to saving, highlighting *all* punctuation and spaces and newlines, and making each miss-spelled word a dropdown or suggest box.
Screenshot of PGDP Wordcheck interface: http://imagizer.imageshack.com/img921/2270/gtQCda.png and PGDP standard interface: http://imagizer.imageshack.com/img923/6371/hFWOnQ.png (And here it's having problems with curved-lines in the scan: http://imagizer.imageshack.com/img921/2090/8vvhpi.png or: http://imagizer.imageshack.com/img924/3403/sHomJS.png )
The few pages I tried didn't have any dropdown boxes, so there might be additional functionality not captured in those images? HTH.
wikitech-l@lists.wikimedia.org