Class JaspellTernarySearchTrie
Implementation of a Ternary Search Trie, a data structure for storing strings that combines the compact size of a binary search tree with the speed of a digital search trie, and is therefore ideal for practical use in sorting and searching data.
This data structure is faster than hashing for many typical search problems, and supports a broader range of useful problems and operations. Ternary searches are faster than hashing and more powerful, too.
The theory of ternary search trees was described at a symposium in 1997 (see "Fast Algorithms for Sorting and Searching Strings," by J.L. Bentley and R. Sedgewick, Proceedings of the 8th Annual ACM-SIAM Symposium on Discrete Algorithms, January 1997). Algorithms in C, Third Edition, by Robert Sedgewick (Addison-Wesley, 1998) provides yet another view of ternary search trees.
Inherited Members
Namespace: Lucene.Net.Search.Suggest.Jaspell
Assembly: Lucene.Net.Suggest.dll
Syntax
public class JaspellTernarySearchTrieConstructors
JaspellTernarySearchTrie()
Constructs an empty Ternary Search Trie.
Declaration
public JaspellTernarySearchTrie()JaspellTernarySearchTrie(CultureInfo)
Constructs an empty Ternary Search Trie, specifying the CultureInfo used for lowercasing.
Declaration
public JaspellTernarySearchTrie(CultureInfo culture)Parameters
| Type | Name | Description |
|---|---|---|
| CultureInfo | culture |
JaspellTernarySearchTrie(FileInfo)
Constructs a Ternary Search Trie and loads data from a FileInfo into the Trie. The file is a normal text document, where each line is of the form word TAB float.
Uses the culture of the current thread to lowercase words before comparing.
Declaration
public JaspellTernarySearchTrie(FileInfo file)Parameters
| Type | Name | Description |
|---|---|---|
| FileInfo | file | The FileInfo with the data to load into the Trie. |
Exceptions
| Type | Condition |
|---|---|
| IOException | A problem occured while reading the data. |
JaspellTernarySearchTrie(FileInfo, bool)
Constructs a Ternary Search Trie and loads data from a FileInfo into the Trie. The file is a normal text document, where each line is of the form "word TAB float".
Uses the culture of the current thread to lowercase words before comparing.
Declaration
public JaspellTernarySearchTrie(FileInfo file, bool compression)Parameters
| Type | Name | Description |
|---|---|---|
| FileInfo | file | The FileInfo with the data to load into the Trie. |
| bool | compression | If true, the file is compressed with the GZIP algorithm, and if false, the file is a normal text document. |
Exceptions
| Type | Condition |
|---|---|
| IOException | A problem occured while reading the data. |
JaspellTernarySearchTrie(FileInfo, bool, CultureInfo)
Constructs a Ternary Search Trie and loads data from a FileInfo into the Trie. The file is a normal text document, where each line is of the form "word TAB float".
Uses the supplied culture to lowercase words before comparing.
NOTE for subclasses: this constructor calls a virtual method, which could result in your override of it being called before the class is properly initialized. To overcome the issue, you could override JaspellTernarySearchTrie(CultureInfo) constructor and then call the logic in a way that suits your needs.
Declaration
public JaspellTernarySearchTrie(FileInfo file, bool compression, CultureInfo culture)Parameters
| Type | Name | Description |
|---|---|---|
| FileInfo | file | The FileInfo with the data to load into the Trie. |
| bool | compression | If true, the file is compressed with the GZIP algorithm, and if false, the file is a normal text document. |
| CultureInfo | culture | The culture used for lowercasing. |
Exceptions
| Type | Condition |
|---|---|
| IOException | A problem occured while reading the data. |
JaspellTernarySearchTrie(FileInfo, CultureInfo)
Constructs a Ternary Search Trie and loads data from a FileInfo into the Trie. The file is a normal text document, where each line is of the form word TAB float.
Uses the supplied culture to lowercase words before comparing.
Declaration
public JaspellTernarySearchTrie(FileInfo file, CultureInfo culture)Parameters
| Type | Name | Description |
|---|---|---|
| FileInfo | file | The FileInfo with the data to load into the Trie. |
| CultureInfo | culture | The culture used for lowercasing. |
Exceptions
| Type | Condition |
|---|---|
| IOException | A problem occured while reading the data. |
Properties
MatchAlmostDiff
Sets the number of characters by which words can differ from target word when calling the MatchAlmost(string, int) method.
Arguments less than 0 will set the char difference to 0, and arguments greater than 3 will set the char difference to 3.
Declaration
public virtual int MatchAlmostDiff { get; set; }Property Value
| Type | Description |
|---|---|
| int |
NumReturnValues
Sets the default maximum number of values returned from the MatchPrefix(string, int) and MatchAlmost(string, int) methods.
The value should be set this to -1 to get an unlimited number of return values. note that the methods mentioned above provide overloaded versions that allow you to specify the maximum number of return values, in which case this value is temporarily overridden.
Declaration
public virtual int NumReturnValues { get; set; }Property Value
| Type | Description |
|---|---|
| int |
Methods
Get(string)
Retrieve the object indexed by a key.
Declaration
public virtual object Get(string key)Parameters
| Type | Name | Description |
|---|---|---|
| string | key | A string index. |
Returns
| Type | Description |
|---|---|
| object | The object retrieved from the Ternary Search Trie. |
GetAndIncrement(string)
Retrieve the Nullable{float} indexed by key, increment it by one unit and store the new Nullable{float}.
Declaration
public virtual float? GetAndIncrement(string key)Parameters
| Type | Name | Description |
|---|---|---|
| string | key | A string index. |
Returns
| Type | Description |
|---|---|
| float? | The Nullable{float} retrieved from the Ternary Search Trie. |
GetKey(TSTNode)
Returns the key that indexes the node argument.
Declaration
protected virtual string GetKey(JaspellTernarySearchTrie.TSTNode node)Parameters
| Type | Name | Description |
|---|---|---|
| JaspellTernarySearchTrie.TSTNode | node | The node whose index is to be calculated. |
Returns
| Type | Description |
|---|---|
| string | The string that indexes the node argument. |
GetNode(string)
Returns the node indexed by key, or null if that node doesn't
exist. Search begins at root node.
Declaration
public virtual JaspellTernarySearchTrie.TSTNode GetNode(string key)Parameters
| Type | Name | Description |
|---|---|---|
| string | key | A string that indexes the node that is returned. |
Returns
| Type | Description |
|---|---|
| JaspellTernarySearchTrie.TSTNode | The node object indexed by key. This object is an instance of an inner class named JaspellTernarySearchTrie.TSTNode. |
GetNode(string, TSTNode)
Returns the node indexed by key, or null if that node doesn't
exist. The search begins at root node.
Declaration
protected virtual JaspellTernarySearchTrie.TSTNode GetNode(string key, JaspellTernarySearchTrie.TSTNode startNode)Parameters
| Type | Name | Description |
|---|---|---|
| string | key | A string that indexes the node that is returned. |
| JaspellTernarySearchTrie.TSTNode | startNode | The top node defining the subtrie to be searched. |
Returns
| Type | Description |
|---|---|
| JaspellTernarySearchTrie.TSTNode | The node object indexed by key. This object is an instance of an inner class named JaspellTernarySearchTrie.TSTNode. |
GetOrCreateNode(string)
Returns the node indexed by key, creating that node if it doesn't exist, and creating any required intermediate nodes if they don't exist.
Declaration
protected virtual JaspellTernarySearchTrie.TSTNode GetOrCreateNode(string key)Parameters
| Type | Name | Description |
|---|---|---|
| string | key | A string that indexes the node that is returned. |
Returns
| Type | Description |
|---|---|
| JaspellTernarySearchTrie.TSTNode | The node object indexed by key. This object is an instance of an inner class named JaspellTernarySearchTrie.TSTNode. |
Exceptions
| Type | Condition |
|---|---|
| ArgumentNullException | If the key is |
| ArgumentException | If the key is an empty string. |
GetSizeInBytes()
Return an approximate memory usage for this trie.
Declaration
public virtual long GetSizeInBytes()Returns
| Type | Description |
|---|---|
| long |
MatchAlmost(string)
Returns a IList<T> of keys that almost match the argument key. Keys returned will have exactly diff characters that do not match the target key, where diff is equal to the last value set to the MatchAlmostDiff property.
If the MatchAlmost(string, int) method is called before the MatchAlmostDiff property has been called for the first time, then diff = 0.
Declaration
public virtual IList<string> MatchAlmost(string key)Parameters
| Type | Name | Description |
|---|---|---|
| string | key | The target key. |
Returns
| Type | Description |
|---|---|
| IList<string> | A IList<T> with the results. |
MatchAlmost(string, int)
Returns a IList<T> of keys that almost match the argument key. Keys returned will have exactly diff characters that do not match the target key, where diff is equal to the last value set to the MatchAlmostDiff property.
If the MatchAlmost(string, int) method is called before the MatchAlmostDiff property has been called for the first time, then diff = 0.
Declaration
public virtual IList<string> MatchAlmost(string key, int numReturnValues)Parameters
| Type | Name | Description |
|---|---|---|
| string | key | The target key. |
| int | numReturnValues | The maximum number of values returned by this method. |
Returns
| Type | Description |
|---|---|
| IList<string> | A IList<T> with the results |
MatchPrefix(string)
Returns an alphabetical IList<T> of all keys in the trie that begin with a given prefix. Only keys for nodes having non-null data are included in the IList<T>.
Declaration
public virtual IList<string> MatchPrefix(string prefix)Parameters
| Type | Name | Description |
|---|---|---|
| string | prefix | Each key returned from this method will begin with the characters in prefix. |
Returns
| Type | Description |
|---|---|
| IList<string> | A IList<T> with the results. |
MatchPrefix(string, int)
Returns an alphabetical IList<T> of all keys in the trie that begin with a given prefix. Only keys for nodes having non-null data are included in the IList<T>.
Declaration
public virtual IList<string> MatchPrefix(string prefix, int numReturnValues)Parameters
| Type | Name | Description |
|---|---|---|
| string | prefix | Each key returned from this method will begin with the characters in prefix. |
| int | numReturnValues | The maximum number of values returned from this method. |
Returns
| Type | Description |
|---|---|
| IList<string> | A IList<T> with the results |
NumDataNodes()
Returns the number of nodes in the trie that have non-null data.
Declaration
public virtual int NumDataNodes()Returns
| Type | Description |
|---|---|
| int | The number of nodes in the trie that have non-null data. |
NumDataNodes(TSTNode)
Returns the number of nodes in the subtrie below and including the starting node. The method counts only nodes that have non-null data.
Declaration
protected virtual int NumDataNodes(JaspellTernarySearchTrie.TSTNode startingNode)Parameters
| Type | Name | Description |
|---|---|---|
| JaspellTernarySearchTrie.TSTNode | startingNode | The top node of the subtrie. the node that defines the subtrie. |
Returns
| Type | Description |
|---|---|
| int | The total number of nodes in the subtrie. |
NumNodes()
Returns the total number of nodes in the trie. The method counts nodes whether or not they have data.
Declaration
public virtual int NumNodes()Returns
| Type | Description |
|---|---|
| int | The total number of nodes in the trie. |
NumNodes(TSTNode)
Returns the total number of nodes in the subtrie below and including the starting Node. The method counts nodes whether or not they have data.
Declaration
protected virtual int NumNodes(JaspellTernarySearchTrie.TSTNode startingNode)Parameters
| Type | Name | Description |
|---|---|---|
| JaspellTernarySearchTrie.TSTNode | startingNode | The top node of the subtrie. The node that defines the subtrie. |
Returns
| Type | Description |
|---|---|
| int | The total number of nodes in the subtrie. |
Put(string, object)
Stores a value in the trie. The value may be retrieved using the key.
Declaration
public virtual void Put(string key, object value)Parameters
| Type | Name | Description |
|---|---|---|
| string | key | A string that indexes the object to be stored. |
| object | value | The object to be stored in the Trie. |
Remove(string)
Removes the value indexed by key. Also removes all nodes that are rendered unnecessary by the removal of this data.
Declaration
public virtual void Remove(string key)Parameters
| Type | Name | Description |
|---|---|---|
| string | key | A string that indexes the object to be removed from the Trie. |
SortKeys(TSTNode, int)
Returns keys sorted in alphabetical order. This includes the start Node and all nodes connected to the start Node.
The number of keys returned is limited to numReturnValues. To get a list that isn't limited in size, set numReturnValues to -1.
Declaration
protected virtual IList<string> SortKeys(JaspellTernarySearchTrie.TSTNode startNode, int numReturnValues)Parameters
| Type | Name | Description |
|---|---|---|
| JaspellTernarySearchTrie.TSTNode | startNode | The top node defining the subtrie to be searched. |
| int | numReturnValues | The maximum number of values returned from this method. |
Returns
| Type | Description |
|---|---|
| IList<string> | A IList<T> with the results. |