User Guide
Unleash the power of KSearch!
User Guide
How does KSearch retrieve Search Results?
KSearch searches for all the different forms of an Arabic word (called inflected forms or inflections); this kind of search is called morphological search. In addition, it is possible to filter the search results based on the meanings of words. This combination of morphological search and meaning filtering is termed here ‘Linguistic Search’. Linguistic search depends on the words being included in a comprehensive dictionary.
In addition to linguistic search, it is possible to search for Arabic words that are not included in the dictionary, such as some Arabic proper nouns, foreign (non-Arabic) words and words written using Latin characters (e.g. English words). This type of search is called ‘Wildcard search’. Wildcards are special characters (* and ? or ؟), where * denotes from zero to any number of characters and ? or ؟ denotes a single character. ?? or ؟؟ denotes exactly two characters, etc.
For example the query (search) word اسكتلندا* searches for اسكتلندا (Scotland) with any prefix (including no prefixes at all) and produces results containing words such as واسكتلندا ([and] Scotland), etc. The query word ؟؟سكتلندا finds words with prefixes two characters long only, such as وكاسكتلندا ([and] [like] Scotland).
It is also possible, in a multiple word query, to mix freely between linguistic search words and words with wildcards. For example, the query “رجل *مودرن” (“*modern man”), might return results containing “رجل مودرن” (“[a] modern man”), “الرجال المودرن” ([the]modern men), etc.; the determiner [the] is a prefix in Arabic.
How to Search using Simple Search
Query words (also called search words) are entered in KSearch's search box (see Fig(1) below). The user interface related to the search demo. in this site has been designed to merge pleasantly with the website. Our technical team can, if you wish, redesign your search page (on your website or the internal web application in your intranet), to suit the nature of your business, when installing KSearch in your website or enterprise:
Fig(1): Simple Search Screen
To use simple search, just enter the search words in the search box and click the search button. For example, if you enter the query سياسة أمريكا (Politics America), you will obtain results which contain both سياسة and أمريكا in the same web page/document. The results will also include all possible inflections of these words.
To search for a phrase (i.e. consecutive words), include the search words that make up the phrase between double quotes; for example “سياسة أمريكا” will retrieve search results that contain that phrase, including all the inflections of these words.
It is also possible to mix single words and phrases in the same search query; for example, the query:
"سياسة أمريكا" الزراعة
“Politics America” agriculture
Will retrieve documents or web pages that contain both all inflections of the phrase and all inflections of the single word.
How to Search using Advanced Search
In order to invoke the advanced search settings, click the link ‘بحث متقدم’ (Advanced Search) from the simple search screen; the advanced search page will then be displayed (see Fig(2) below).
Fig(2) Advanced Search Screen
It is important to note here that once a set of search settings has been selected from the advanced search page, these settings remain in effect throughout the browser session. The advanced search page contains the following settings:
- نوع البحث (Search Type):
[1] البحث بالجذر (Root Search)
This is the default search type; using this type, all the inflections of the search words that have the same morphological root as the search word will be retrieved. Note that the search results will be filtered by the meaning selected for the word. For example, for the query word اجتماع (meeting), its root is ج م ع and so root search would display results such as والاجتماع ([and the] meeting) and اجتماعات (meetings), as well as words like اجتمعوا ([they] met).
[2] البحث بالمشتقات (Stem Search)
Using this search type, all the inflections of the search word’s stem, rather than its root, will be retrieved. For example, for the query word الاجتماع ([the] meeting), its stem is اجتماع and so stem search would only display results such as والاجتماع ([and the] meeting) and اجتماعات (meetings), and not words like اجتمعوا ([they] met).
[3] البحث بالمرادفات (Thesaurus Search)
With this search type enabled, all the inflections of the search words that have the same meaning will be retrieved, whether they have the same morphological root as the search word or not. For example, for the query word بترول (oil), the search results would also contain the inflections of the word نفط (a synonym of oil, in Arabic), as well as those of the query word itself, of course.
- خيارات البحث (Search Options):
[1] "بكل الكلمات" (ALL Words)
This is the default search option when using simple search. This search option searches for web pages/documents that contain ALL the query words in the same page/document.
[2] "بأي كلمة" (ANY Word)
This option searches for web pages/documents that contain ANY one of the query words in the same page/document.
[3] "منطقيا"(Boolean)
This option allows the user to include Boolean or logical relations (AND, OR, NOT), as well as brackets to separate logical groups. The following capabilities are possible under this option:
(a) Phrase Search:
{Example: السياسة الداخلية (domestic policy)}. The search results will contain all inflections of the words in this phrase.
(b) Phrase Search with Adjacency, in Order of Query Words:
{Example: سياسة ق/3 خارجية (policy ق/3 foreign)}. The results will return all inflections of the word policy that occur before the word foreign, with a maximum of 3 intervening words.
(c) Phrase Search with Adjacency, not necessarily in Order of Query Words:
{Example: سياسة م/3 خارجية (policy م/3 foreign)}. The results will return all inflections of the word policy that occur in the vicinity of (either before or after) the word foreign, with a maximum of 3 intervening words.
(d) Phrase Search using Logical Relations و (AND), أو (OR), ليس (NOT):
{Example: السياسة الدولية و المصرية ليس أمريكا (International Politics AND [the]Egyptian NOT America)}. This query would retrieve results containing all inflections of the words in the exact phrase(International Politics) as well as all inflections of the word Egyptian, but not words matching any inflections of the query word America. Note that there must be at least one space between a query word and a logical relation.
- طريقة العرض (Display Mode):
This setting determines the way in which the search results are displayed inside a text document only (by clicking the appropriate search result hyperlink); it does not affect the way in which web pages (.html files) are displayed. There are two settings:
[1] على فقرات (Paragraph Mode)
This is the default setting. Using this setting, only paragraphs that contain search results (target words) in the text file are displayed, with these words highlighted.
[2] الملف كاملا (Complete File Mode)
Using this setting, the complete file is displayed in pages. When navigating through the pages, if a page happens to contain search results, then the relevant words are highlighted.
- اختيار الفهرس (Index selection):
Certain websites (or parts thereof), as well as text documents have been pre-indexed using KIndexer. These indexes are shown under the title اختيار الفهرس (Index Selection). When the check box next to the index name is checked, this means that it will be included in the search and if it is unchecked, it will be excluded from the search. KIndexer (not included in the online demo.) allows you to group indexes into groups; only one group of indexes has been used here and is called G1.
Examples: A Brief Tutorial
- A 'Simple Search' Example:
From the 'Simple Search' page, enter the search words سياسة أمريكا and click the search button. Fig(3) shows the first 10 search results. Note that under the title معاني الكلمات (Word Meanings), a combo box contains the search words, with a list of possible meanings (shown as hyperlinks) next to the selected word.
Fig(3): A Simple Search Example
If you wish to search using a different meaning for a particular word, click the combo box until the desired word is displayed, then select the required meaning by clicking it and finally click the search button to perform the search and display the search results, filtered by the new meaning.
If you click on any of the search results, the relevant web page will be displayed, with the various inflections of the target search words highlighted, as shown in Fig(4). You may then navigate to the next (التالي) and previous (السابق) highlighted words using the arrows at the bottom of the page.
Fig(4): A Simple Search Example: Targets inside Web Page
- A 'Simple Search' Example with Spelling Correction:
KSearch includes spelling error detection and correction for the query words entered in the search box. In the simple search page, enter the word شياسة (politics, misspelled). You will immediately be informed that the word is incorrectly spelled and a list of spelling suggestions will be presented (see Fig(5)).
Fig(5): A Simple Search Example: Misspelled Arabic Word
If you click on any of the spelling suggestions, the search will automatically be executed using the selected suggestion. In this case, click on the suggestion سياسة (politics, correctly spelled) and the search results will be accordingly displayed.
In some cases, one of the words you entered might not be a 'misspelled' Arabic word, but a foreign word that has not been included in the linguistic dictionary. For example, enter the word جونز (Jones); here, instead of choosing one of the spelling suggestions, you would click on الكلمة كما هي (search for word 'as is') and the search results will be displayed, as shown in Fig(6).
Fig(6): A Simple Search Example: Misspelled Foreign Word
- An 'Advanced Search' Example:
Go to the 'Advanced Search' page and enter the search words البترول الدولي و سياسة أمريكا (International Oil AND Politics America). Select the search type البحث بالمرادفات (Thesaurus Search) and the search option منطقيا (Boolean Search) – see Fig(7).
Fig(7): An Advanced Search Example
Click the search button and the search results are displayed as shown in Fig(8); these show the web pages that possibly contain all inflected forms of the phrase البترول الدولي (as well as the inflected forms of its synonyms) AND all inflected forms of the phrase سياسة أمريكا (as well as the inflected forms of its synonyms).
Fig(8): An Advanced Search Example - Results
- A Search Example for a Text Document:
Go to the 'Advanced Search' page and check only the index called 'The Holy Quran', which is the index of the Holy Quran as a text document. Select the search type البحث بالجذر (Root Search) and the search option كل الكلمات (All Words). Then enter the search query "غفور رحيم" ("forgiving merciful") and click the search button. The search results will then be displayed, as shown in Fig(9).
Fig(9): Search Results from Text Document
If you click on the search result, the results inside the text document will be displayed in paragraphs, as shown in Fig(10).
Fig(10): Search Results Inside Text Document (Paragraph Mode)
Close the window shown in Fig(10) and return to the Advanced Search page again. With the same search words, type and option, this time select the setting to display the results inside the text file in 'Complete File' mode (الملف كاملا), then click the search button again. When the search result appears, click on it and the first page containing search results inside the file will be displayed, as shown in Fig(11).
Fig(11): Search Results Inside Text Document ('Complete File' Mode)
You may navigate through the file at will; one of the navigation controls (the top one) is labelled صفحات الملف (Document Pages) and allows you to navigate freely to any page, whether the page contains a search result or not. If you wish to navigate only to a page containing results, then use the lower navigation control labelled صفحات النتائج (Results Pages) – see Fig(12).
Fig(12): Search Results Inside Text Document ('Complete File' Mode) – Navigation Controls