Search

Match word forms by

The word itself, the forms it is an inflection of, then words sharing two letters in a row. Best for most searches.

Words sharing any two letters in a row: ܟܬܒ also finds ܬܒ.

Words sharing three letters in a row: ܟܬܒ finds ܟܬܒܐ but not ܬܒ.

Words ranked by how many letters must change to turn them into yours.

A pattern: [ܟܝ]ܬܒ finds ܟܬܒ and ܝܬܒ.

The matching Sedra uses to tag texts: the words your form can be an inflection of.

Bi-Gram finds words that have 2 consecutive letters in common with 2 consecutive letters in the searched term: searching ܟܬܒ 'to write' will also find ܘܟܬܒ 'and ...', ܟܬܒܐ 'book', ܒܟܬܒܗ 'in his book', but will also find ܬܶܒ 'sit' and ܬܳܒ 'to repent'.

Tri-Gram finds words that have 3 consecutive letters in common with 3 consecutive letters in the searched term: searching ܟܬܒ 'to write' will also find all its conjugations, ܟܬܒܐ 'book' and its conjugations, but will not find ܬܶܒ 'sit' or ܬܳܒ 'to repent'.

Levenshtein will find words that are similar based on the number of single character edits that must be applied to the found word in order to change that word into the searched term.

Dynamic Matching will find words based upon the fuzzy search algorithm used by Sedra for dynamic text tagging. This algorithm uses lexical information to determine likely word matches.

Regular Expression searching [ܟܝ]ܬܒ will find words that begin with either ܟ‍ or ܝ followed by ܬ followed by ܒ such as ܟܬܒ 'to write' and ܝܬܒ 'to sit'.

Notes on Search Results

The above search algorithms report the similarity found between the search term and the found words. For instance, searching for ܟܬܒ 'to write' using bigrams will give find the exact word with a 1.0 (i.e. exact match) similarity, but will also find ܬܒ 'sit' and ܬܒ 'repent' with 0.6667 similarity. It will also find ܟܬܒܐ 'book' with 0.8 similarity. The higher the number, the closer it is to the search term. Trigram search works the same way. The results also give the Levenshtein Distance which is the number of edits that must be applied to the found word to make it match the search term. The number 0 represents no edits (i.e. exact match) while a higher number indicates that the found word is further away from the search term.

Regular Expression

Regular expressions allow for powerful search. Instead of searching for a word, you can search for a pattern. For example, if you are looking for words that begin with either ܒ or ܡ‍ and end with ܬ you can search for [ܒܡ]*ܬ. This will find words like ܒܝܬ 'house', ܒܗܬ 'ashamed', ܒܪܬ 'daughter' etc. You can visit the Postgres Regular Expression page to read more about the syntax of regular expressions accepted by Sedra.