serializable-bptree 1.1.0 → 3.0.0

package/README.md

@@ -7,9 +7,13 @@ This is a B+tree that's totally okay with duplicate values. If you need to keep
 
 ```typescript
 import { readFileSync, writeFileSync, existsSync } from 'fs'
-import { BPTree, SerializeStrategy, NumericComparator } from 'serializable-bptree'
+import {
+  BPTreeSync,
+  SerializeStrategySync,
+  NumericComparator
+} from 'serializable-bptree'
 
-class FileStoreStrategy extends SerializeStrategy<K, V> {
+class FileStoreStrategySync extends SerializeStrategySync<K, V> {
   id(): number {
     const random = Math.ceil(Math.random()*1000000)
     return random
@@ -40,11 +44,12 @@ class FileStoreStrategy extends SerializeStrategy<K, V> {
 }
 
 const order = 5
-const tree = new BPTree(
-  new FileStoreStrategy(order),
+const tree = new BPTreeSync(
+  new FileStoreStrategySync(order),
   new NumericComparator()
 )
 
+tree.init()
 tree.insert('a', 1)
 tree.insert('b', 2)
 tree.insert('c', 3)
@@ -59,7 +64,9 @@ tree.where({ gt: 0, lt: 4 }) // [{ key: 'a', value: 1 }, { key: 'c', value: 3 }]
 
 ## Why use a `serializable-bptree`?
 
-Firstly, in most cases, there is no need to use a B+tree in JavaScript. This is because there is a great alternative, the Map object. Nonetheless, if you need to retrieve values in a sorted order, a B+tree can be a good solution. These cases are often related to databases, and you may want to store this state not just in memory, but on a remote server or in a file. In this case, `serializable-bptree` can help you.
+Firstly, in most cases, there is no need to use a B+tree in JavaScript. This is because there is a great alternative, the Map object. Nonetheless, if you need to retrieve values in a sorted order, a B+tree can be a good solution. These cases are often related to databases, and you may want to store this state not just in memory, but on a remote server or in a file. In this case, **serializable-bptree** can help you.
+
+Additionally, this library supports asynchronous operations. Please refer to the section below for instructions on using it asynchronously.
 
 ## How to use
 
@@ -70,14 +77,29 @@ npm i serializable-bptree
 ```
 
 ```typescript
-import { BPTree } from 'serializable-bptree'
+import {
+  BPTreeSync,
+  BPTreeAsync,
+  SerializeStrategySync,
+  SerializeStrategyAsync,
+  NumericComparator,
+  StringComparator
+} from 'serializable-bptree'
 ```
 
 ### Browser (esm)
 
 ```html
 <script type="module">
-  import { BPTree } from 'https://cdn.jsdelivr.net/npm/serializable-bptree@1.x.x/dist/esm/index.min.js'
+  import {
+    BPTreeSync,
+    BPTreeAsync,
+    InMemoryStoreStrategySync,
+    InMemoryStoreStrategyAsync,
+    ValueComparator,
+    NumericComparator,
+    StringComparator
+  } from 'https://cdn.jsdelivr.net/npm/serializable-bptree@3.x.x/dist/esm/index.min.js'
 </script>
 ```
 
@@ -85,15 +107,15 @@ import { BPTree } from 'serializable-bptree'
 
 ### Value comparator
 
-B+tree needs to keep values in sorted order. Therefore, a process to compare the sizes of values is needed, and that role is played by the `ValueComparator`.
+B+tree needs to keep values in sorted order. Therefore, a process to compare the sizes of values is needed, and that role is played by the **ValueComparator**.
 
-Commonly used numerical and string comparisons are natively supported by the `serializable-bptree` library. Use it as follows:
+Commonly used numerical and string comparisons are natively supported by the **serializable-bptree** library. Use it as follows:
 
 ```typescript
 import { NumericComparator, StringComparator } from 'serializable-bptree'
 ```
 
-However, you may want to sort complex objects other than numbers and strings. For example, if you want to sort by the `age` property order of an object, you need to create a new class that inherits from the `ValueComparator` class. Use it as follows:
+However, you may want to sort complex objects other than numbers and strings. For example, if you want to sort by the **age** property order of an object, you need to create a new class that inherits from the **ValueComparator** class. Use it as follows:
 
 ```typescript
 import { ValueComparator } from 'serializable-bptree'
@@ -103,13 +125,15 @@ interface MyObject {
   name: string
 }
 
-class AgeComparator {
+class AgeComparator extends ValueComparator<MyObject> {
   asc(a: MyObject, b: MyObject): number {
     return a.age - b.age
   }
 }
 ```
 
+The **asc** method should return values in ascending order. If the return value is negative, it means that the parameter **a** is smaller than **b**. If the return value is positive, it means that **a** is greater than **b**. If the return value is **0**, it indicates that **a** and **b** are of the same size.
+
 ### Serialize strategy
 
 A B+tree instance is made up of numerous nodes. You would want to store this value when such nodes are created or updated. Let's assume you want to save it to a file.
@@ -117,9 +141,9 @@ A B+tree instance is made up of numerous nodes. You would want to store this val
 You need to construct a logic for input/output from the file by inheriting the SerializeStrategy class. Look at the class structure below:
 
 ```typescript
-import { SerializeStrategy } from 'serializable-bptree'
+import { SerializeStrategySync } from 'serializable-bptree'
 
-class MyFileIOStrategy extends SerializeStrategy {
+class MyFileIOStrategySync extends SerializeStrategySync {
   id(): number
   read(id: number): BPTreeNode<K, V>
   write(id: number, node: BPTreeNode<K, V>): void
@@ -132,7 +156,7 @@ What does this method mean? And why do we need to construct such a method?
 
 #### id(): `number`
 
-When a node is created in the B+tree, the node needs a unique value to represent itself. This is the `node.id` attribute, and you can specify this attribute yourself. For example, it could be implemented like this.
+When a node is created in the B+tree, the node needs a unique value to represent itself. This is the **node.id** attribute, and you can specify this attribute yourself. For example, it could be implemented like this.
 
 ```typescript
 id(): number {
@@ -142,7 +166,9 @@ id(): number {
 }
 ```
 
-Or, you could use file input/output to save and load the value of the `before` variable.
+Or, you could use file input/output to save and load the value of the **before** variable.
+
+This method is called before a node is created in the tree. Therefore, it can also be used to allocate space for storing the node.
 
 #### read(id: `number`): `BPTreeNode<K, V>`
 
@@ -158,7 +184,7 @@ read(id: number): BPTreeNode<K, V> {
 }
 ```
 
-This method is called only once when loading a node from a tree instance.
+This method is called only once when loading a node from a tree instance. The loaded node is loaded into memory, and subsequently, when the tree references the node, it operates based on the values in memory **without** re-invoking this method.
 
 #### write(id: `number`, node: `BPTreeNode<K, V>`): `void`
 
@@ -181,7 +207,7 @@ function writeBack(id: number, node: BPTreeNode<K, V>, timer: number) {
 
 ...
 write(id: number, node: BPTreeNode<K, V>): void {
-  const writeBackInterval = 100
+  const writeBackInterval = 10
   writeBack(id, node, writeBackInterval)
 }
 ```
@@ -190,32 +216,32 @@ This kind of delay writing should ideally occur within a few milliseconds. If th
 
 #### readHead(): `SerializeStrategyHead`|`null`
 
-This method is called only once when the tree is created. It's a method to restore the saved tree information. If it is the initial creation and there is no stored root node, it should return `null`.
+This method is called only once when the tree is created. It's a method to restore the saved tree information. If it is the initial creation and there is no stored root node, it should return **null**.
 
-This method should return the value stored in the `writeHead` method.
+This method should return the value stored in the **writeHead** method.
 
 #### writeHead(head: `SerializeStrategyHead`): `void`
 
-This method is called whenever the head information of the tree changes, typically when the root node changes.
+This method is called whenever the head information of the tree changes, typically when the root node changes. This method also works when the tree's **setHeadData** method is called. This is because the method attempts to store head data in the root node.
 
-As a parameter, it receives the header information of the tree. This value should be serialized and stored. Later, the `readHead` method should convert this serialized value into a json format and return it.
+As a parameter, it receives the header information of the tree. This value should be serialized and stored. Later, the **readHead** method should convert this serialized value into a json format and return it.
 
 ### The Default `ValueComparator` and `SerializeStrategy`
 
-To utilize `serializable-bptree`, you need to implement certain functions. However, a few basic helper classes are provided by default.
+To utilize **serializable-bptree**, you need to implement certain functions. However, a few basic helper classes are provided by default.
 
 #### ValueComparator
 
 * `NumericComparator`
 * `StringComparator`
 
-If the values being inserted into the tree are numeric, please use the `NumericComparator` class.
+If the values being inserted into the tree are numeric, please use the **NumericComparator** class.
 
 ```typescript
 import { NumericComparator } from 'serializable-bptree'
 ```
 
-If the values being inserted into the tree can be strings, you can use the `StringComparator` class in this case.
+If the values being inserted into the tree can be strings, you can use the **StringComparator** class in this case.
 
 ```typescript
 import { StringComparator } from 'serializable-bptree'
@@ -223,14 +249,134 @@ import { StringComparator } from 'serializable-bptree'
 
 #### SerializeStrategy
 
-* `InMemoryStoreStrategy`
+* `InMemoryStoreStrategySync`
+* `InMemoryStoreStrategyAsync`
+
+As of now, the only class supported by default is the **InMemoryStoreStrategy**. This class is suitable for use when you prefer to operate the tree solely in-memory, similar to a typical B+ tree.
+
+```typescript
+import {
+  InMemoryStoreStrategySync,
+  InMemoryStoreStrategyAsync
+} from 'serializable-bptree'
+```
+
+## Data Query Condition Clause
+
+This library supports various conditional clauses. Currently, it supports **gte**, **gt**, **lte**, **lt**, **equal**, **notEqual**, and **like** conditions. Each condition is as follows:
+
+### `gte`
+
+Queries values that are greater than or equal to the given value.
+
+### `gt`
+
+Queries values that are greater than the given value.
+
+### `lte`
+
+Queries values that are less than or equal to the given value.
 
-As of now, the only class supported by default is the `InMemoryStoreStrategy`. This class is suitable for use when you prefer to operate the tree solely in-memory, similar to a typical B+ tree.
+### `lt`
+
+Queries values that are less than the given value.
+
+### `equal`
+
+Queries values that match the given value.
+
+### `notEqual`
+
+Queries values that do not match the given value.
+
+### `like`
+
+Queries values that contain the given value in a manner similar to regular expressions. Special characters such as % and _ can be used.
+
+**%** matches zero or more characters. For example, **%ada%** means all strings that contain "ada" anywhere in the string. **%ada** means strings that end with "ada". **ada%** means strings that start with **"ada"**.
+
+**_** matches exactly one character.
+Using **p_t**, it can match any string where the underscore is replaced by any character, such as "pit", "put", etc.
+
+You can obtain matching data by combining these condition clauses. If there are multiple conditions, an **AND** operation is used to retrieve only the data that satisfies all conditions.
+
+## Using Asynchronously
+
+Support for asynchronous trees has been available since version 3.0.0. Asynchronous is useful for operations with delays, such as file input/output and remote storage. Here is an example of how to use it:
 
 ```typescript
-import { InMemoryStoreStrategy } from 'serializable-bptree'
+import { existsSync } from 'fs'
+import { readFile, writeFile } from 'fs/promises'
+import {
+  BPTreeAsync,
+  SerializeStrategyAsync,
+  NumericComparator,
+  StringComparator
+} from 'serializable-bptree'
+
+class FileStoreStrategyAsync extends SerializeStrategyAsync<K, V> {
+  async id(): Promise<number> {
+    const random = Math.ceil(Math.random()*1000000)
+    return random
+  }
+
+  async read(id: number): Promise<BPTreeNode<K, V>> {
+    const raw = await readFile(id.toString(), 'utf8')
+    return JSON.parse(raw)
+  }
+
+  async write(id: number, node: BPTreeNode<K, V>): Promise<void> {
+    const stringify = JSON.stringify(node)
+    await writeFile(id.toString(), stringify, 'utf8')
+  }
+
+  async readHead(): Promise<SerializeStrategyHead|null> {
+    if (!existsSync('head')) {
+      return null
+    }
+    const raw = await readFile('head', 'utf8')
+    return JSON.parse(raw)
+  }
+
+  async writeHead(head: SerializeStrategyHead): Promise<void> {
+    const stringify = JSON.stringify(head)
+    await writeFile('head', stringify, 'utf8')
+  }
+}
+
+const order = 5
+const tree = new BPTreeAsync(
+  new FileStoreStrategyAsync(order),
+  new NumericComparator()
+)
+
+await tree.init()
+await tree.insert('a', 1)
+await tree.insert('b', 2)
+await tree.insert('c', 3)
+
+await tree.delete('b', 2)
+
+await tree.where({ equal: 1 }) // [{ key: 'a', value: 1 }]
+await tree.where({ gt: 1 }) // [{ key: 'c', value: 3 }]
+await tree.where({ lt: 2 }) // [{ key: 'a', value: 1 }]
+await tree.where({ gt: 0, lt: 4 }) // [{ key: 'a', value: 1 }, { key: 'c', value: 3 }]
 ```
 
+The implementation method for asynchronous operations is not significantly different. The **-Async** suffix is used instead of the **-Sync** suffix in the **BPTree** and **SerializeStrategy** classes. The only difference is that the methods become asynchronous. The **ValueComparator** class and similar value comparators do not use asynchronous operations.
+
+## Precautions for Use
+
+### Synchronization Issue
+
+The serializable-bptree minimizes file I/O by storing loaded nodes in-memory. This approach works well in situations where there is a 1:1 relationship between the remote storage and the client. However, in a 1:n scenario, where multiple clients read from and write to a single remote storage, data inconsistency between the remote storage and the clients can occur.
+
+To solve this issue, it's necessary to update the cached nodes. The forceUpdate method was created for this purpose. It allows for the update of specific nodes, but when updating a node ID, a signal must be sent to all clients connected to the remote storage, informing them that the node has been updated. Clients must receive this signal and call the forceUpdate method to update the node. This logic goes beyond the scope of the library, so it must be implemented directly.
+
+### Concurrency Issue in Asynchronous Trees
+
+This issue occurs only in asynchronous trees and can also occur in a 1:1 relationship between remote storage and client. During the process of inserting/removing data asynchronously, querying the data can result in inconsistent data. To prevent concurrency issues, do not query data while inserting/removing it.
+
 ## LICENSE
 
 MIT LICENSE