Accessing main picture of wikipedia page by API [closed]

https://stackoverflow.com/questions/8363531

27-10-2019
|

Question

Is there any way I can access the thumbnail picture of any wikipedia page by using an API? I mean the image on the top right side in box. Is there any APIs for that?

Solution

http://en.wikipedia.org/w/api.php

Look at prop=images.

It returns an array of image filenames that are used in the parsed page. You then have the option of making another API call to find out the full image URL, e.g.: action=query&titles=Image:INSERT_EXAMPLE_FILE_NAME_HERE.jpg&prop=imageinfo&iiprop=url

or to calculate the URL via the filename's hash.

Unfortunately, while the array of images returned by prop=images is in the order they are found on the page, the first can not be guaranteed to be the image in the info box because sometimes a page will include an image before the infobox (most of the time icons for metadata about the page: e.g. "this article is locked").

Searching the array of images for the first image that includes the page title is probably the best guess for the infobox image.

OTHER TIPS

You can get the thumbnail of any wikipedia page using prop=pageimages. For example:

http://en.wikipedia.org/w/api.php?action=query&titles=Al-Farabi&prop=pageimages&format=json&pithumbsize=100

And you will get the thumbnail full URL.

This is good way to get the Main Image of a page in wikipedia

http://en.wikipedia.org/w/api.php?action=query&prop=pageimages&format=json&piprop=original&titles=India

Check out the MediaWiki API example for getting the main picture of a wikipedia page: https://www.mediawiki.org/wiki/API:Page_info_in_search_results.

As other's have mentioned, you would use prop=pageimages in your API query.

If you also want the image description, you would use prop=pageimages|pageterms instead in your API query.

You can get the original image using piprop=original. Or you can get a thumbnail image with a specified width/height. For a thumbnail with width/height=600, piprop=thumbnail&pithumbsize=600. If you omit either, the image returned in the API callback will default to a thumbnail with width/height of 50px.

If you are requesting results in JSON format, you should always use formatversion=2 in your API query (i.e., format=json&formatversion=2) because it makes retrieving the image from the query easier.

Original Size Image:

https://en.wikipedia.org/w/api.php?action=query&format=json&formatversion=2&prop=pageimages|pageterms&piprop=original&titles=Albert Einstein

Thumbnail Size (600px width/height) Image:

https://en.wikipedia.org/w/api.php?action=query&format=json&formatversion=2&prop=pageimages|pageterms&piprop=thumbnail&pithumbsize=600&titles=Albert Einstein

Way 1: You can try some query like this:

http://en.wikipedia.org/w/api.php?action=opensearch&limit=5&format=xml&search=italy&namespace=0

in the response, you can see the Image tag.

<Item>
<Text xml:space="preserve">Italy national rugby union team</Text>
<Description xml:space="preserve">
The Italy national rugby union team represent the nation of Italy in the sport of rugby union.
</Description>
<Url xml:space="preserve">
http://en.wikipedia.org/wiki/Italy_national_rugby_union_team
</Url>
<Image source="http://upload.wikimedia.org/wikipedia/en/thumb/4/46/Italy_rugby.png/43px-Italy_rugby.png" width="43" height="50"/>
</Item>

Way 2: use query http://en.wikipedia.org/w/index.php?action=render&title=italy

then you can get a raw html code, you can get the image use something like PHP Simple HTML DOM Parser http://simplehtmldom.sourceforge.net

I have no time write it to you. just give you some advice, thanks.

I'm sorry for not answering specifically your question about the main image. But here's some code to get a list of all images:

function makeCall($url) {
    $curl = curl_init();
    curl_setopt($curl, CURLOPT_URL, $url);
    curl_setopt($curl, CURLOPT_RETURNTRANSFER, 1);
    return curl_exec($curl);
}

function wikipediaImageUrls($url) {
    $imageUrls = array();
    $pathComponents = explode('/', parse_url($url, PHP_URL_PATH));
    $pageTitle = array_pop($pathComponents);
    $imagesQuery = "http://en.wikipedia.org/w/api.php?action=query&titles={$pageTitle}&prop=images&format=json";
    $jsonResponse = makeCall($imagesQuery);
    $response = json_decode($jsonResponse, true);
    $imagesKey = key($response['query']['pages']);
    foreach($response['query']['pages'][$imagesKey]['images'] as $imageArray) {
        if($imageArray['title'] != 'File:Commons-logo.svg' && $imageArray['title'] != 'File:P vip.svg') {
            $title = str_replace('File:', '', $imageArray['title']);
            $title = str_replace(' ', '_', $title);
            $imageUrlQuery = "http://en.wikipedia.org/w/api.php?action=query&titles=Image:{$title}&prop=imageinfo&iiprop=url&format=json";
            $jsonUrlQuery = makeCall($imageUrlQuery);
            $urlResponse = json_decode($jsonUrlQuery, true);
            $imageKey = key($urlResponse['query']['pages']);
            $imageUrls[] = $urlResponse['query']['pages'][$imageKey]['imageinfo'][0]['url'];
        }
    }
    return $imageUrls;
}
print_r(wikipediaImageUrls('http://en.wikipedia.org/wiki/Saturn_%28mythology%29'));
print_r(wikipediaImageUrls('http://en.wikipedia.org/wiki/Hans-Ulrich_Rudel'));

I got this for http://en.wikipedia.org/wiki/Saturn_%28mythology%29:

Array
(
    [0] => http://upload.wikimedia.org/wikipedia/commons/1/10/Arch_of_SeptimiusSeverus.jpg
    [1] => http://upload.wikimedia.org/wikipedia/commons/8/81/Ivan_Akimov_Saturn_.jpg
    [2] => http://upload.wikimedia.org/wikipedia/commons/d/d7/Lucius_Appuleius_Saturninus.jpg
    [3] => http://upload.wikimedia.org/wikipedia/commons/2/2c/Polidoro_da_Caravaggio_-_Saturnus-thumb.jpg
    [4] => http://upload.wikimedia.org/wikipedia/commons/b/bd/Porta_Maggiore_Alatri.jpg
    [5] => http://upload.wikimedia.org/wikipedia/commons/6/6a/She-wolf_suckles_Romulus_and_Remus.jpg
    [6] => http://upload.wikimedia.org/wikipedia/commons/4/45/Throne_of_Saturn_Louvre_Ma1662.jpg
)

And for the second URL (http://en.wikipedia.org/wiki/Hans-Ulrich_Rudel):

Array
(
    [0] => http://upload.wikimedia.org/wikipedia/commons/e/e9/BmRKEL.jpg
    [1] => http://upload.wikimedia.org/wikipedia/commons/3/3f/BmRKELS.jpg
    [2] => http://upload.wikimedia.org/wikipedia/commons/2/2c/Bundesarchiv_Bild_101I-655-5976-04%2C_Russland%2C_Sturzkampfbomber_Junkers_Ju_87_G.jpg
    [3] => http://upload.wikimedia.org/wikipedia/commons/6/62/Bundeswehr_Kreuz_Black.svg
    [4] => http://upload.wikimedia.org/wikipedia/commons/9/99/Flag_of_German_Reich_%281935%E2%80%931945%29.svg
    [5] => http://upload.wikimedia.org/wikipedia/en/6/64/HansUlrichRudel.jpeg
    [6] => http://upload.wikimedia.org/wikipedia/commons/8/82/Heinkel_He_111_during_the_Battle_of_Britain.jpg
    [7] => http://upload.wikimedia.org/wikipedia/commons/6/66/Regulation_WW_II_Underwing_Balkenkreuz.png
)

Note that the URL changed a bit on the 6th element of the second array. It's what @JosephJaber was warning about in his comment above.

Hope this helps someone.

I have written some code that gets main image (full URL) by Wikipedia article title. It's not perfect, but overall I'm very pleased with the results.

The challenge was that when queried for a specific title, Wikipedia returns multiple image filenames (without path). Furthermore, the secondary search (I used the code varatis posted in this thread - thanks!) returns URLs of all images found based on the image filename that was searched, regardless of the original article title. After all this, we may end up with a generic image irrelevant to the search, so we filter those out. The code iterates over filenames and URLs until it finds (hopefully the best) match... a bit complicated, but it works :)

Note on the generic filter: I've been compiling a list of generic image strings for the isGeneric() function, but the list just keeps growing. I am considering maintaining it as a public list - if there is any interest let me know.

Pre:

protected static $baseurl = "http://en.wikipedia.org/w/api.php";

Main function - get image URL from title:

public static function getImageURL($title)
{
    $images = self::getImageFilenameObj($title); // returns JSON object
    if (!$images) return '';

    foreach ($images as $image)
    {
        // get object of image URL for given filename
        $imgjson = self::getFileURLObj($image->title);

        // return first image match
        foreach ($imgjson as $img)
        {
            // get URL for image
            $url = $img->imageinfo[0]->url;

            // no image found               
            if (!$url) continue;

            // filter generic images
            if (self::isGeneric($url)) continue;

            // match found
            return $url;
        }
    }
    // match not found
    return '';          
}

== The following functions are called by the main function above ==

Get JSON object (filenames) by title:

public static function getImageFilenameObj($title)
{
    try     // see if page has images
    {
        // get image file name
        $json = json_decode(
            self::retrieveInfo(
                self::$baseurl . '?action=query&titles=' .
                urlencode($title) . '&prop=images&format=json'
            ))->query->pages;

        /** The foreach is only to get around
         *  the fact that we don't have the id.
         */
        foreach ($json as $id) { return $id->images; }
    }
    catch(exception $e) // no images
    {
        return NULL;
    }
}

Get JSON object (URLs) by filename:

public static function getFileURLObj($filename)
{
    try                     // resolve URL from filename
    {
        return json_decode(
            self::retrieveInfo(
                self::$baseurl . '?action=query&titles=' .
                urlencode($filename) . '&prop=imageinfo&iiprop=url&format=json'
            ))->query->pages;
    }
    catch(exception $e)     // no URLs
    {
        return NULL;
    }
}

Filter out generic images:

public static function isGeneric($url)
{
    $generic_strings = array(
        '_gray.svg',
        'icon',
        'Commons-logo.svg',
        'Ambox',
        'Text_document_with_red_question_mark.svg',
        'Question_book-new.svg',
        'Canadese_kano',
        'Wiki_letter_',
        'Edit-clear.svg',
        'WPanthroponymy',
        'Compass_rose_pale',
        'Us-actor.svg',
        'voting_box',
        'Crystal_',
        'transportation_inv',
        'arrow.svg',
        'Quill_and_ink-US.svg',
        'Decrease2.svg',
        'Rating-',
        'template',
        'Nuvola_apps_',
        'Mergefrom.svg',
        'Portal-',
        'Translation_to_',
        '/School.svg',
        'arrow',
        'Symbol_',
        'stub',
        'Unbalanced_scales.svg',
        '-logo.',
        'P_vip.svg',
        'Books-aj.svg_aj_ashton_01.svg',
        'Film',
        '/Gnome-',
        'cap.svg',
        'Missing',
        'silhouette',
        'Star_empty.svg',
        'Music_film_clapperboard.svg',
        'IPA_Unicode',
        'symbol',
        '_highlighting_',
        'pictogram',
        'Red_pog.svg',
        '_medal_with_cup',
        '_balloon',
        'Feature',
        'Aiga_'
    );

    foreach ($generic_strings as $str)
    {
        if (stripos($url, $str) !== false) return true;
    }

    return false;
}

Comments welcome.

I there is a way to reliably get a main image for a wikipedia page - the Extension called PageImages

The PageImages extension collects information about images used on a page.

Its aim is to return the single most appropriate thumbnail associated with an article, attempting to return only meaningful images, e.g. not those from maintenance templates, stubs or flag icons. Currently it uses the first non-meaningless image used in the page.

https://www.mediawiki.org/wiki/Extension:PageImages

Just add the prop pageimages to your API Query:

/w/api.php?action=query&prop=pageimages&titles=Somepage&format=xml

This reliably filters out annoying default images and prevents you from having to filter them yourself! The extension is installed on all the main wikipedia pages...

See this related question on an API for Wikipedia. However, I would not know if it is possible to retrieve the thumbnail picture through an API.

You can also consider just parsing the web page to find the image URL, and retrieve the image that way.

Lets take Example of Page http://en.wikipedia.org/wiki/index.html?curid=57570 to get Main Pic

Check out

prop=pageprops

action=query&pageids=57570&prop=pageprops&format=json

Results Page Data Eg.

{ "pages" : { "57570":{
                    "pageid":57570,
                    "ns":0,
                    "title":"Sachin Tendulkar",
                    "pageprops" : {
                         "defaultsort":"Tendulkar,Sachin",
                         "page_image":"Sachin_at_Castrol_Golden_Spanner_Awards_(crop).jpg",
                         "wikibase_item":"Q9488"
                    }
            }
          }
 }}

We get main Pic file name this result as

** (wikiId).pageprops.page_image = Sachin_at_Castrol_Golden_Spanner_Awards_(crop).jpg**

Now as we have Image file name we will have to make another Api Call to get full image path from file name as follows

action=query&titles=Image:INSERT_EXAMPLE_FILE_NAME_HERE.jpg&prop=imageinfo&iiprop=url

Eg.

action=query&titles=Image:Sachin_at_Castrol_Golden_Spanner_Awards_(crop).jpg&prop=imageinfo&iiprop=url

Returns Array of Image Data having url in it as http://upload.wikimedia.org/wikipedia/commons/3/35/Sachin_at_Castrol_Golden_Spanner_Awards_%28crop%29.jpg

Like Anuraj mentioned, the pageimages parameter is it. Look at the following url that'll bring about some nifty stuff:

https://en.wikipedia.org/w/api.php?action=query&prop=info|extracts|pageimages|images&inprop=url&exsentences=1&titles=india

Her are some interesting parameters:

The two parameters extracts and exsentences gives you a short description you can use. (exsentences is the number of sentences you want to include in the excerpt)
The info and the inprop=url parameters gives you the url of the page
The prop property has multiple parameters separated by a bar symbol
And if you insert the format=json in there, it is even better

I think not, but you can capture the image using a link parser HTML documents

Here is my list of XPaths I have found work for 95 percent of articles. the main ones are 1, 2 3 and 4. A lot of articles are not formatted correctly and these would be edge cases:

You can use a DOM parsing lib to fetch image using the XPath.

static NSString   *kWikipediaImageXPath2    =   @"//*[@id=\"mw-content-text\"]/div[1]/div/table/tr[2]/td/a/img";
static NSString   *kWikipediaImageXPath3    =   @"//*[@id=\"mw-content-text\"]/div[1]/table/tr[1]/td/a/img";
static NSString   *kWikipediaImageXPath1    =   @"//*[@id=\"mw-content-text\"]/div[1]/table/tr[2]/td/a/img";
static NSString   *kWikipediaImageXPath4    =   @"//*[@id=\"mw-content-text\"]/div[2]/table/tr[2]/td/a/img";
static NSString   *kWikipediaImageXPath5    =   @"//*[@id=\"mw-content-text\"]/div[1]/table/tr[2]/td/p/a/img";
static NSString   *kWikipediaImageXPath6    =   @"//*[@id=\"mw-content-text\"]/div[1]/table/tr[2]/td/div/div/a/img";
static NSString   *kWikipediaImageXPath7    =   @"//*[@id=\"mw-content-text\"]/div[1]/table/tr[1]/td/div/div/a/img";

I used a ObjC wrapper called Hpple around libxml2.2 to pull out the image url. Hope this helps

Licensed under: CC-BY-SA with attribution

Not affiliated with StackOverflow