本說明書實施例旨在提供一種更有效的下拉提示方法，以解決現有技術中的不足。
為實現上述目的，本說明書一個態樣提供一種建立下拉提示索引的方法，所述下拉提示用於使用者終端APP中所包括的至少一個搜尋框，每個搜尋框設置於所述APP中的對應業務中，所述方法在伺服器端執行，包括：獲取所述使用者終端APP中每個業務的用於下拉提示的多個項目；獲取每個項目的至少一個屬性；建立前綴索引，其中，所述前綴索引包括每個項目的前綴到其自身的反向索引；以及建立屬性索引，其中所述屬性索引包括每個項目與其各個屬性之間的索引。
在一個實施例中，在上述建立下拉提示索引的方法中，所述至少一個屬性包括以下至少一個：業務類型、查詢次數、結果數、點擊次數、點擊通過率、地理位置、城市以及地址。
在一個實施例中，在上述建立下拉提示索引的方法中，所述建立屬性索引包括，建立第一項目到其第一屬性的正向索引。
在一個實施例中，在上述建立下拉提示索引的方法中，所述建立屬性索引包括，建立第二項目的第二屬性到第二項目的反向索引。
在一個實施例中，在上述建立下拉提示索引的方法中，所述建立前綴索引包括離線建立前綴索引；以及所述建立屬性索引包括離線建立屬性索引。
在一個實施例中，在上述建立下拉提示索引的方法中，所述建立第一項目到其第一屬性的正向索引包括，對所述第一項目設置多個項目子文件，用於分別建立所述第一項目到在不同城市的第一屬性的正向索引。
在一個實施例中，在上述建立下拉提示索引的方法中，所述建立前綴索引包括，獲取第三項目包括的最大語義粒度詞，以及建立所述最大語義粒度詞及其各自的前綴到所述第三項目的反向索引。
本說明書另一態樣提供一種用於下拉提示的搜尋方法，所述下拉提示用於使用者終端APP中所包括的至少一個搜尋框，每個搜尋框設置於所述APP中的對應業務中，所述方法在伺服器端執行，包括：從使用者終端APP接收使用者在搜尋框中輸入的查詢內容；從使用者終端APP接收與使用者終端、使用者終端APP或使用者相關的至少一個屬性；獲取與所述搜尋框對應的業務的搜尋條件；通過根據上述建立索引的方法獲得的前綴索引，搜尋出與該查詢內容對應的多個項目；以及根據所述搜尋條件和所述至少一個屬性，通過根據上述建立索引的方法獲得的屬性索引，對所述多個項目進行過濾，從而獲得搜尋結果。
在一個實施例中，所述用於下拉提示的搜尋方法還包括，在獲得搜尋結果之後，獲取與所述搜尋框對應的業務的排序方法；以及通過所述排序方法對所述搜尋結果進行排序。
在一個實施例中，在上述用於下拉提示的搜尋方法中，所述至少一個屬性包括以下至少一個：搜尋框業務類型、使用者終端地理位置、使用者終端城市和使用者查詢時間。
在一個實施例中，在上述用於下拉提示的搜尋方法中，根據所述搜尋條件和所述至少一個屬性，通過根據上述建立索引的方法獲得的屬性索引，對所述多個項目進行過濾包括：根據將項目業務類型限定為所述搜尋框業務類型的搜尋條件，過濾掉不屬於所述搜尋框業務類型的項目。
本說明書另一態樣提供一種下拉提示方法，所述下拉提示用於使用者終端APP中所包括的至少一個搜尋框，每個搜尋框設置於所述APP中的對應業務中，所述方法在使用者終端APP中執行，包括：獲取使用者在搜尋框中輸入的查詢內容；獲取與使用者終端、使用者終端APP或使用者相關的至少一個第一屬性；將所述查詢內容和所述至少一個第一屬性發送給伺服器端；從伺服器端接收通過根據上述搜尋方法獲得的搜尋結果；以及在所述搜尋框的下拉提示位置顯示所述搜尋結果。
在一個實施例中，所述下拉提示方法還包括，在從伺服器端接收通過根據上述搜尋方法獲得的搜尋結果之後，獲取與使用者相關的至少一個第二屬性，獲取與所述搜尋框對應的業務的排序方法，以及，根據所述至少一個第二屬性，通過所述排序方法對所述搜尋結果進行排序。
在一個實施例中，在上述下拉提示方法中，所述至少一個第二屬性包括以下至少一個：使用者年齡、性別、消費偏好以及消費能力。
本說明書另一態樣提供一種建立下拉提示索引的裝置，所述下拉提示用於使用者終端APP中所包括的至少一個搜尋框，每個搜尋框設置於所述APP中的對應業務中，所述裝置在伺服器端實施，包括：項目獲取單元，配置為，獲取所述使用者終端APP中每個業務的用於下拉提示的多個項目；屬性獲取單元，配置為，獲取每個項目的至少一個屬性；前綴索引建立單元，配置為，建立前綴索引，其中所述前綴索引包括每個項目的前綴到其自身的反向索引；以及屬性索引建立單元，配置為，建立屬性索引，其中所述屬性索引包括每個項目與其各個屬性之間的索引。
本說明書另一態樣提供一種用於下拉提示的搜尋裝置，所述下拉提示用於使用者終端APP中所包括的至少一個搜尋框，每個搜尋框設置於所述APP中的對應業務中，所述裝置在伺服器端實施，包括：查詢內容接收單元，配置為，從使用者終端APP接收使用者在搜尋框中輸入的查詢內容；屬性接收單元，配置為，從使用者終端APP接收與使用者終端、使用者終端APP或使用者相關的至少一個屬性；獲取單元，配置為，獲取與所述搜尋框對應的業務的搜尋條件；
第一搜尋單元，配置為，通過根據上述建立索引的方法獲得的前綴索引，搜尋出與該查詢內容對應的多個項目；以及第二搜尋單元， 配置為，根據所述搜尋條件和所述至少一個屬性，通過根據上述建立索引的方法獲得的屬性索引，對所述多個項目進行過濾，從而獲得搜尋結果。
本說明書另一態樣提供一種下拉提示裝置，所述下拉提示用於使用者終端APP中所包括的至少一個搜尋框，每個搜尋框設置於所述APP中的對應業務中，所述裝置在使用者終端APP中實施，包括：查詢內容獲取單元，配置為，獲取使用者在搜尋框中輸入的查詢內容；屬性獲取單元，配置為，獲取與使用者終端、使用者終端APP或使用者相關的至少一個第一屬性；發送單元，配置為，將所述查詢內容和所述至少一個第一屬性發送給伺服器端；接收單元，配置為，從伺服器端接收通過根據上述搜尋方法獲得的搜尋結果；以及顯示單元，配置為，在所述搜尋框的下拉提示位置顯示所述搜尋結果。
通過根據本說明書實施例的下拉提示方案，通過使用搜尋引擎中的技術實現下拉提示系統，並且通過一套通用下拉提示系統支援多個下拉提示業務，節省了人力和資源成本，提高了迭代效率，並且有足夠的靈活性，可以支援不同業務的不同效果需求。The embodiments of the present specification aim to provide a more effective drop-down prompt method to solve the deficiencies in the prior art.
To achieve the above purpose, one aspect of this specification provides a method for creating a pull-down prompt index, the pull-down prompt is used for at least one search box included in a user terminal APP, and each search box is provided in a corresponding one in the APP In the business, the method is executed on the server side, and includes: acquiring a plurality of items used for pull-down prompts for each business in the user terminal APP; acquiring at least one attribute of each project; establishing a prefix index, wherein, The prefix index includes the reverse index of each item's prefix to itself; and the establishment of an attribute index, where the attribute index includes the index between each item and its respective attributes.
In one embodiment, in the above method of establishing a drop-down prompt index, the at least one attribute includes at least one of the following: business type, query times, result number, clicks, click-through rate, geographic location, city, and address.
In one embodiment, in the above method of creating a drop-down prompt index, the establishing an attribute index includes establishing a forward index from the first item to its first attribute.
In one embodiment, in the above method of creating a drop-down prompt index, the establishing an attribute index includes establishing an inverse index of the second attribute of the second item to the second item.
In one embodiment, in the above method of creating a drop-down prompt index, the establishing a prefix index includes establishing a prefix index offline; and the establishing an attribute index includes establishing an attribute index offline.
In one embodiment, in the above method of creating a drop-down prompt index, the establishment of a positive index of a first item to its first attribute includes setting a plurality of item sub-files for the first item, which are used to establish Forward index of the first item to the first attribute in different cities.
In one embodiment, in the above method for creating a drop-down prompt index, the prefix index establishment includes obtaining the maximum semantic granularity words included in the third item, and establishing the maximum semantic granularity words and their respective prefixes to the The reverse index of the third item.
Another aspect of this specification provides a search method for a pull-down prompt, the pull-down prompt is used for at least one search box included in a user terminal APP, each search box is set in a corresponding service in the APP, The method is executed on the server side, and includes: receiving the query content input by the user in the search box from the user terminal APP; receiving at least one related to the user terminal, the user terminal APP, or the user from the user terminal APP Attributes; obtaining search conditions for services corresponding to the search box; searching for multiple items corresponding to the query content through the prefix index obtained according to the indexing method described above; and according to the search conditions and the at least one Attribute, through the attribute index obtained according to the above indexing method, filtering the plurality of items to obtain a search result.
In one embodiment, the search method for the drop-down prompt further includes, after obtaining the search result, a method for sorting the business corresponding to the search box; and sorting the search result by the method .
In one embodiment, in the above search method for drop-down prompts, the at least one attribute includes at least one of the following: a search box service type, a user terminal geographic location, a user terminal city, and a user query time.
In one embodiment, in the above search method for pull-down prompts, filtering the plurality of items according to the search condition and the at least one attribute through an attribute index obtained according to the above indexing method includes : Filtering out items that do not belong to the search box business type according to the search conditions that limit the item business type to the search box business type.
Another aspect of this specification provides a pull-down prompt method, where the pull-down prompt is used for at least one search box included in a user terminal APP, and each search box is set in a corresponding service in the APP. The execution in the user terminal APP includes: acquiring query content entered by the user in the search box; acquiring at least one first attribute related to the user terminal, the user terminal APP, or the user; combining the query content and the At least one first attribute is sent to the server side; the search result obtained by the above search method is received from the server side; and the search result is displayed at the drop-down prompt position of the search box.
In one embodiment, the pull-down prompt method further includes, after receiving the search result obtained by the search method from the server side, acquiring at least one second attribute related to the user, and acquiring corresponding to the search box Sorting method of the business, and, according to the at least one second attribute, sorting the search results by the sorting method.
In one embodiment, in the above pull-down prompt method, the at least one second attribute includes at least one of the following: the user's age, gender, consumption preference, and consumption ability.
Another aspect of this specification provides a device for creating a pull-down prompt index, where the pull-down prompt is used for at least one search box included in a user terminal APP, and each search box is set in a corresponding service in the APP. The device is implemented on the server side, and includes: an item acquisition unit configured to acquire multiple items for each drop-down prompt in the user terminal APP; an attribute acquisition unit configured to acquire each item At least one attribute; a prefix index establishment unit configured to establish a prefix index, wherein the prefix index includes the reverse index of each item's prefix to itself; and an attribute index establishment unit configured to establish an attribute index, where all The attribute index includes the index between each item and its various attributes.
Another aspect of this specification provides a search device for a pull-down prompt, the pull-down prompt is used for at least one search box included in a user terminal APP, each search box is set in a corresponding service in the APP, The device is implemented on the server side and includes: a query content receiving unit configured to receive the query content entered by the user in the search box from the user terminal APP; an attribute receiving unit configured to receive and receive from the user terminal APP At least one attribute related to the user terminal, the user terminal APP, or the user; the obtaining unit is configured to obtain the search condition of the service corresponding to the search box;
The first search unit is configured to search for a plurality of items corresponding to the query content through the prefix index obtained according to the above indexing method; and the second search unit is configured to be based on the search condition and the at least For an attribute, the multiple items are filtered by the attribute index obtained according to the above indexing method to obtain a search result.
Another aspect of this specification provides a pull-down prompt device, the pull-down prompt is used for at least one search box included in the user terminal APP, each search box is set in a corresponding service in the APP, the device is Implemented in the user terminal APP, including: query content acquisition unit configured to acquire the query content entered by the user in the search box; attribute acquisition unit configured to acquire user terminal, user terminal APP or user related At least one first attribute; a sending unit configured to send the query content and the at least one first attribute to the server side; a receiving unit configured to receive from the server side the data obtained by the search method described above Search results; and a display unit configured to display the search results at the drop-down prompt position of the search box.
Through the pull-down prompting scheme according to the embodiment of the present specification, the pull-down prompting system is implemented by using the technology in the search engine, and the multiple pull-down prompting services are supported through a set of universal pull-down prompting systems, which saves labor and resource costs and improves iteration efficiency. And there is enough flexibility to support the different effect requirements of different businesses.
下面將結合附圖描述本說明書實施例。
圖1示出了根據本說明書實施例的下拉提示系統示意圖。如圖中所示，系統包括用戶端(101、102、103)和伺服器端104。用戶端可以是使用者終端APP，在APP中可能包括至少一個搜尋框，不同的搜尋框對應於不同的業務。例如，用戶端101在使用對應於APP中的業務1的搜尋框，用戶端102在使用對應於APP中的業務2的搜尋框，以及用戶端103在使用對應於APP中的業務3的搜尋框。
根據本說明書實施例的下拉提示系統通過伺服器端104中的搜尋實例11提供下拉提示服務。搜尋實例11包括前綴索引111、屬性索引112以及排序外掛程式113。其中，前綴索引是指項目中包括的前綴字或詞或其拼音到該項目的反向索引，前綴索引111中包括APP中全部業務的項目。所述屬性索引包括每個項目與其各個屬性之間的索引。當使用者在搜尋框中輸入搜尋內容時，用戶端(101、102、103)將查詢內容及與使用者終端、使用者終端APP以及使用者相關的屬性通過網絡一起發送給伺服器端104。在APP中包括多個搜尋框的情況中，上述相關屬性中包括業務類型，通過使用者輸入的搜尋框所在的業務確定業務類型。在伺服器端104接收到查詢內容和相關屬性之後，伺服器端104獲取對業務類型設置的搜尋條件，並將其與查詢內容和相關屬性一起傳送給搜尋實例11。搜尋實例11根據前綴索引111搜尋出與查詢內容相關的項目，根據屬性索引112，並依據搜尋條件，對搜尋出的項目進行過濾，從而獲得搜尋結果。例如，搜尋條件中包括只搜尋本業務類型的項目，從而根據項目與業務類型屬性之間的索引，將不屬於該業務類型的項目過濾出去。搜尋實例11還包括排序外掛程式113，排序外掛程式113中設置有適用於各個業務的排序方法，從而以對應於業務類型的排序方法對搜尋結果進行排序。伺服器端104中還包括資料庫12，其中儲存有各個業務的項目及與項目相關的屬性，搜尋實例11通過從資料庫12獲取資料，建立前綴索引111和屬性索引112。
下面結合實例描述根據本說明書的下拉提示系統。
圖2示出了一種建立下拉提示索引的方法。所述下拉提示用於使用者終端APP中所包括的至少一個搜尋框，每個搜尋框設置於所述APP中的對應業務中。所述方法在伺服器端執行，更具體地，如圖1所示，在伺服器端104的搜尋實例11中執行。所述方法包括以下步驟：在步驟S21， 獲取所述使用者終端APP中每個業務的用於下拉提示的多個項目；在步驟S22，獲取每個項目的至少一個屬性；在步驟S23，建立前綴索引，其中，所述前綴索引包括每個項目的前綴到其自身的反向索引；以及在步驟S24，建立屬性索引，其中所述屬性索引包括每個項目與其各個屬性之間的索引。
首先，在步驟S21，獲取所述使用者終端APP中每個業務的用於下拉提示的多個項目。例如，在支付寶APP中，包括口碑業務、財富業務等。其中每個業務都在伺服器端包括對應的資料庫，用於儲存與業務相關的資料，這些資料中的一些資料可以用作下拉提示的項目。例如，在口碑業務中，與業務相關的資料包括業務類型、店鋪名、菜名、店鋪地址、店鋪地理位置、點擊次數、查詢次數等。其中，店鋪名和菜名可用作下拉提示的項目。例如，在財富業務中，基金名可用作下拉提示的項目。從而，搜尋實例11通過讀取資料庫12，可獲取各個業務中可用於下拉提示的項目。
在步驟S22，獲取每個項目的至少一個屬性。不同業務的項目具有與業務對應的屬性。例如，口碑業務的項目包括的屬性有：業務類型(口碑)、店鋪地址、店鋪地理位置、城市、結果數、點擊次數、點擊通過率、查詢次數等。財富業務包括的屬性有：業務類型(財富)、結果數、總點擊次數、總點擊通過率、不同城市的點擊次數、不同城市的點擊通過率等。可以以不同的欄位分別表示這些屬性。例如，以欄位“type”表示業務類型。屬於口碑業務的項目都包括“type=口碑”的屬性。屬於財富業務的項目都包括“type=財富”的屬性。在這些屬性中，有些屬性資料是由店鋪提交的，如店鋪地址、店鋪地理位置、城市等，有些屬性資料是由伺服器端從APP端接收獲得的，例如點擊次數、查詢次數等。伺服器端將這些屬性與項目關聯地儲存在資料庫12中，從而搜尋實例11可通過讀取資料庫12，獲取每個項目的至少一個屬性。
在步驟S23，建立前綴索引，其中，所述前綴索引包括每個項目的前綴到其自身的反向索引。項目的前綴是包括的項目尾字之前的單個字或連續多個字。例如，項目“農夫山泉”的前綴包括“農”、“農夫”、“農夫山”、“夫”、“夫山”和“山”。從而可以獲取“農”、“農夫”、“農夫山”、“夫”、“夫山”和“山”到“農夫山泉”的反向索引。
在獲取了APP中全部業務的可用於下拉提示的項目之後，可以如上所述獲取這些項目的全部前綴作為索引表中的key，以相應的項目作為索引表中與key對應的值，從而建立全部業務通用的一個前綴索引。在一個實施例中，前綴索引中還包括上述前綴的拼音到項目的索引。
在一個實施例中，可以建立項目所包括的最大語義粒度詞及其各自的前綴到項目的前綴索引。例如，對於項目“農夫山泉有點甜”，可根據語義詞典將其分為“農夫山泉”和“有點甜”兩個最大語義粒度詞。從而，可以建立“農夫山泉”及其前綴“農”、“農夫”、“農夫山”、 “夫”、“夫山”和“山”到“農夫山泉有點甜”的反向索引，還可以建立“有點甜”及其前綴“有”、“有點”和“點”到“農夫山泉有點甜”的反向索引。
在步驟S24，建立屬性索引，其中所述屬性索引包括每個項目與其各個屬性之間的索引。項目與屬性之間的索引可以是屬性到項目的反向索引，也可以是項目到屬性正向索引。在一個實施例中，建立業務類型到項目的反向索引。在另一個實施例中，建立地理位置到項目的反向索引。在另一個實施例中，建立項目到點擊次數的正向索引。
例如，對於口碑業務中的店鋪項目可以建立如表1所示的正向索引表。
表1
而口碑業務中的一個菜名項目可能與多個城市相關。在該情況中，可對該菜名項目設置多個子文件，以分別索引其在不同城市的屬性。例如，菜名項目關於屬性的正向索引表可以如表2所示。
表2
在一個實施例中，如圖1所示，搜尋實例11在從伺服器端104中的資料庫12中獲取業務資料之後，在預定時段離線建立前綴索引和屬性索引。通常，在每天淩晨2-5點使用者較少的時段，進行對索引的離線全量更新。在一個實施例中，搜尋實例11在從資料庫12獲取使用者實時更新資料之後，對索引進行實時更新。例如，在口碑業務中，當新增店鋪，或者店鋪更新菜名時，搜尋實例通過訂閱店鋪資訊，可以實時從資料庫12獲取該店鋪更改資訊，從而在索引中進行對資訊的秒級更新。
圖3示出了根據本說明書實施例的一種用於下拉提示的搜尋方法。所述下拉提示用於使用者終端APP中所包括的至少一個搜尋框，每個搜尋框設置於所述APP中的對應業務中。所述方法在伺服器端執行，包括以下步驟：在步驟S31，從使用者終端APP接收使用者在所述搜尋框中輸入的查詢內容；在步驟S32，從使用者終端APP接收與使用者終端、使用者終端APP或使用者相關的至少一個屬性；在步驟S33，獲取與所述搜尋框對應的業務的搜尋條件；在步驟S34，通過根據上述建立索引的方法獲得的前綴索引，搜尋出與該查詢內容對應的多個項目；以及在步驟S35，根據所述搜尋條件和所述至少一個屬性，通過根據上述建立索引的方法獲得的屬性索引，對所述多個項目進行過濾，從而獲得搜尋結果。
首先，在步驟S31，從使用者終端APP接收使用者在所述搜尋框中輸入的查詢內容。使用者在APP中的特定搜尋框中輸入的查詢內容可能是單個字、連續幾個字，或者可能是拼音。使用者每進行一次輸入動作，APP都會將這次輸入的內容同步發送給伺服器端。
在步驟S32，從使用者終端APP接收與使用者終端、使用者終端APP或使用者相關的至少一個屬性。這些屬性可以是與使用者終端相關的使用者終端所在的地理位置(經緯度)和城市、與使用者終端APP相關的搜尋框業務類型、與使用者相關的使用者查詢時間等。這些屬性可以分別以欄位表示。伺服器端從使用者終端APP接收的具體屬性與搜尋框業務類型相關。例如，當搜尋框業務類型為口碑業務時，伺服器端從APP端接收搜尋框業務類型(口碑)、使用者終端所在的地理位置等屬性。而當搜尋框業務類型為財富業務時，伺服器端從APP端接收搜尋框業務類型(財富)等屬性。其中，業務類型通過使用者查詢使用的搜尋框在APP中的位置確定。例如，使用者在支付寶APP中的口碑業務中的搜尋框輸入查詢內容，則搜尋框業務類型為“type=口碑”。在一個實施例中，使用者在支付寶APP首頁的搜尋框中進行查詢，這裡的搜尋框業務類型為全域業務，即“type=all”。
在步驟S33，獲取與所述搜尋框對應的業務的搜尋條件。對於不同的業務，設置有相應的搜尋條件。例如，對於口碑業務來說，搜尋條件可能包括：業務類型為“口碑”，到使用者終端距離在預定距離之內等。對於財富業務，搜尋條件可能包括產品的風險等級在預定級別等。對於全域業務，其對應的搜尋條件為APP包括的各個業務的全部搜尋條件。這些搜尋條件可以設置在伺服器端，伺服器端從其自身的儲存位置獲取與業務類型對應的搜尋條件並傳送給搜尋實例。在一個實施例中，可以在使用者終端APP中設置所述搜尋條件，伺服器端根據搜尋框業務類型從APP接收對應的搜尋條件，並將其傳送給搜尋實例。
在步驟S34，通過根據上述建立索引的方法獲得的前綴索引，搜尋出與該查詢內容對應的多個項目。例如，在支付寶APP中的一個搜尋框中，輸入查詢內容“阿里”，則根據這裡的前綴索引，搜尋出的項目可能包括：阿里影業、阿里健康、阿里納製藥、阿里郎、阿里食府、阿里拌麵等。其中“阿里影業、阿里健康、阿里納製藥”是屬於財富業務的項目，其中“阿里郎、阿里食府、阿里拌麵”是屬於口碑業務的項目。
在步驟S35，根據所述搜尋條件和所述至少一個屬性，通過根據上述建立索引的方法獲得的屬性索引，對所述多個項目進行過濾，從而獲得搜尋結果。例如，在上一步驟S34中搜尋獲得多個項目“阿里影業、阿里健康、阿里納製藥、阿里郎、阿里食府、阿里拌麵”之後，在使用者在財富業務的搜尋框中輸入查詢內容“阿里”的情況中，伺服器端從APP接收的至少一個屬性中包括：搜尋框業務類型type=財富，而財富業務的搜尋條件中包括將項目業務類型限定為搜尋框業務類型“財富”。從而根據業務類型到項目的反向索引，搜尋出滿足業務類型是“財富”的項目。通過取兩次搜尋的結果的交集，將上述通過前綴索引搜尋出的項目中業務類型不是“財富”的項目過濾出去，從而獲得搜尋結果“阿里影業、阿里健康、阿里納製藥”。
再例如，使用者在口碑業務的搜尋框中輸入“阿里”，伺服器端從APP接收的屬性包括搜尋框業務類型(口碑)和使用者終端的地理位置，獲取的與口碑業務對應的搜尋條件包括：項目的業務類型為口碑，並且項目的地理位置距離使用者終端為五公里之內。在上一步驟S34中搜尋獲得多個項目“阿里影業、阿里健康、阿里納製藥、阿里郎、阿里食府、阿里拌麵”之後，根據口碑業務的搜尋條件及業務類型屬性和使用者地理位置屬性，通過業務類型到項目的反向索引，搜尋獲得符合“type=口碑”的項目包括“阿里郎、阿里食府、阿里拌麵、巴貝拉、巴西烤肉…”，根據地理位置到項目的反向索引，搜尋獲得地理位置屬性符合距離使用者終端為五公里之內項目包括“阿里食府、阿里拌麵、巴西烤肉、肯德基、必勝客…”，取前綴索引和上述兩種屬性索引的索引結果的交集，從而達到對前綴索引結果的過濾，從而獲得搜尋結果“阿里食府、阿里拌麵”。
在另一個例子中，對於上述搜尋結果“阿里食府、阿里拌麵”，口碑業務對應的搜尋條件中可能還包括：例如，“查詢次數超過300次”。從而，根據該搜尋條件，並且根據口碑業務中的店鋪項目到查詢次數屬性的正向索引，獲取項目“阿里食府”、“阿里拌麵”各自的查詢次數，並根據查詢次數對其進行過濾。
圖4示意示出上述搜尋過程，其中為了方便示意表示，在搜尋結果中，以大寫字母表示屬於一個業務類型(例如財富業務)的項目，以小寫字母表示屬於另一個業務類型(例如口碑業務)的項目。在圖4中，搜尋實例中包括的前綴索引41、第一屬性索引42(例如業務類型索引)、第二屬性索引43(例如地理位置索引)都是反向索引，通過這三個索引的搜尋，可獲得其相應的搜尋結果{A、B、C、a、b、c}、{a、b、c、d、e}和{b、c、e、f、g}。對三個搜尋結果取交集，獲得搜尋結果{b、c}。在一個實施例中，可能存在與第三屬性(例如查詢次數)相關的搜尋，第三屬性索引44可以是正向索引。從而，根據搜尋條件和第三屬性索引44，對上述搜尋結果{b、c}進行進一步正排過濾，從而獲得最終的搜尋結果，例如{b}。
在一個實施例中，在上述步驟S35中獲得的搜尋結果的數量可能非常大，可能有幾千個。在該情況中，可以先對該搜尋結果進行排序，然後擷取排在前面的預定數目(例如500個)的結果發送給使用者終端APP。
在一個實施例中，如圖1所示，根據本說明書實施例的搜尋實例11中包括排序外掛程式113。在排序外掛程式113中針對不同的業務設置有不同的排序方法。搜尋實例11在將所述搜尋結果發送給所述使用者終端APP之前，從排序外掛程式113獲取與所述搜尋框對應的業務的排序方法，並通過該排序方法對所述搜尋結果進行排序。例如，對應於口碑業務中店鋪項目的排序方法可以包括：首先根據項目地理位置屬性到使用者終端的距離進行排序，在該排序的下一層，通過項目查詢次數屬性進行排序，在下一層，通過項目的點擊次數屬性進行排序等。而對應於財富業務的項目的排序方法可以包括：根據項目的收益率屬性進行排序、根據項目的風險等級屬性進行排序、根據項目的點擊次數屬性進行排序等。
圖5示出了一種下拉提示方法，所述方法在使用者終端APP中執行，所述使用者終端APP包括至少一個搜尋框，每個搜尋框設置於所述APP中的對應業務中。所述方法包括以下步驟：在步驟S51，獲取使用者在搜尋框中輸入的查詢內容；在步驟S52，獲取與使用者終端、使用者終端APP或使用者相關的至少一個屬性；在步驟S53，將所述查詢內容和所述至少一個第一屬性發送給伺服器端；在步驟S54，從伺服器端接收通過根據上述搜尋方法獲得的搜尋結果；以及在步驟S55，在所述搜尋框的下拉提示位置顯示所述搜尋結果。
首先，在步驟S51，獲取使用者在搜尋框中輸入的查詢內容。使用者每次對搜尋框中進行一次輸入動作，APP都會進行該步驟，獲取查詢內容。例如，使用者將要輸入“農夫山泉”，當使用者輸入“農”時，APP獲取的查詢內容為“農”。當使用者再輸入“夫”時，APP獲取的查詢內容為“農夫”。
在步驟S52，獲取與使用者終端、使用者終端APP或使用者相關的至少一個屬性。這裡的至少一個屬性如上文參考圖3中所述，可以是與使用者終端相關的使用者終端所在的地理位置(經緯度)和城市、與使用者終端APP相關的搜尋框業務類型、與使用者相關的使用者查詢時間等。這些屬性可以分別以欄位表示。同樣地，使用者終端APP對伺服器端傳送的具體屬性與搜尋框業務類型相關，在此不再贅述。對於與使用者終端相關的屬性，APP可通過具體應用獲取這些屬性資料，例如，通過定位應用獲取使用者終端地理位置。對於與APP相關的屬性，可通過APP自身的設置獲取，例如，通過對各個業務的搜尋框設置欄位以對搜尋框業務類型進行限定，從而直接從搜尋框獲取其對應的業務類型。對於與使用者相關的屬性，根據不同的屬性可採用不同的獲取方法。例如使用者查詢時間，可從APP中的時鐘獲取，再例如，對於使用者的一些個人特徵，可從伺服器接收獲取。
在步驟S53，將所述查詢內容和所述至少一個第一屬性發送給伺服器端。與步驟S51對應地，使用者每次對搜尋框中進行一次輸入動作，APP都會進行該步驟，即，發送查詢內容及相關屬性。例如，當使用者輸入“農”時，APP將獲取的查詢內容“農”發送給伺服器端。當使用者再輸入“夫”時，APP將獲取的查詢內容“農夫”發送給伺服器端。
在步驟S54，從伺服器端接收通過根據上述搜尋方法獲得的搜尋結果。在一個實施例中，該搜尋結果是經過排序的結果。在另一個實施例中，可在APP中對該搜尋結果進行二次排序。APP中也可以包括排序外掛程式，以用於設置對應於不同業務的排序方法。該排序方法相比於搜尋實例中的排序方法，例如可以基於一些使用者特徵參數，基於一些更複雜的排序模型。例如，對於使用者在口碑業務的查詢，APP可以從伺服器端的資料庫中獲取使用者的與消費相關的肖像特徵，例如年齡、性別、消費偏好以及消費能力等，並根據使用者的肖像特徵及對應的排序方法，對從伺服器端接收的搜尋結果進行進一步排序。再例如，當搜尋框位於APP首頁時，對應的業務為全域業務，即，APP中包括的全部業務，其也具有與其對應的排序方法。在全域業務中，會依據對應的排序方法對各個業務中的項目計算全域分數，並依據全域分數對APP中的各個業務項目進行排序。
在步驟S55，在所述搜尋框的下拉提示位置顯示所述搜尋結果。在APP中，搜尋框的下拉提示位置一般為搜尋框的下方。可以設置下拉提示的項目個數、或頁數，從而根據個數或頁數限定，對上述排序結果從頭擷取。在一個實施例中，還可以在下拉提示中顯示項目的結果數、距離、地址等屬性。
圖6示意示出了根據本說明書實施例的下拉提示效果示意圖。其中，搜尋框為口碑業務中設置的搜尋框，使用者輸入的查詢內容為“農夫”。在對該詞的下拉提示中，包括結果數、距離、地址等屬性。
圖7示出了一種建立下拉提示索引的裝置700。所述下拉提示用於使用者終端APP中所包括的至少一個搜尋框，每個搜尋框設置於所述APP中的對應業務中。所述裝置700在伺服器端實施，包括：項目獲取單元71，配置為，獲取所述使用者終端APP中每個業務的用於下拉提示的多個項目；屬性獲取單元72，配置為，獲取每個項目的至少一個屬性；前綴索引建立單元73，配置為，建立前綴索引，其中所述前綴索引包括每個項目的前綴到其自身的反向索引；以及屬性索引建立單元74，配置為，建立屬性索引，其中所述屬性索引包括每個項目與其各個屬性之間的索引。
在一個實施例中，在所述建立下拉提示索引的裝置中，所述屬性索引建立單元74還配置為，建立第一項目到其第一屬性的正向索引。
在一個實施例中，在所述建立下拉提示索引的裝置中，所述屬性索引建立單元74還配置為，建立第二項目的第二屬性到第二項目的反向索引。
在一個實施例中，在所述建立下拉提示索引的裝置中，所述前綴索引建立單元73還配置為離線建立前綴索引；以及所述屬性索引建立單元74還配置為離線建立屬性索引。
在一個實施例中，在所述建立下拉提示索引的裝置中，所述屬性索引建立單元74還配置為，對所述第一項目設置多個項目子文件，用於分別建立所述第一項目到在不同城市的第一屬性的正向索引。
在一個實施例中，在所述建立下拉提示索引的裝置中，所述前綴索引建立單元73還配置為，獲取第三項目包括的最大語義粒度詞，以及建立所述最大語義粒度詞及其各自的前綴到所述第三項目的反向索引。
圖8示出一種用於下拉提示的搜尋裝置800。所述下拉提示用於使用者終端APP中所包括的至少一個搜尋框，每個搜尋框設置於所述APP中的對應業務中。所述裝置800在伺服器端實施，包括：查詢內容接收單元81，配置為，從使用者終端APP接收使用者在搜尋框中輸入的查詢內容；屬性接收單元82，配置為，從使用者終端APP接收與使用者終端、使用者終端APP或使用者相關的至少一個屬性；獲取單元83，配置為，獲取與所述搜尋框對應的業務的搜尋條件；第一搜尋單元84，配置為，通過根據上述建立索引的方法獲得的前綴索引，搜尋出與該查詢內容對應的多個項目；以及第二搜尋單元85， 配置為，根據所述搜尋條件和所述至少一個屬性，通過根據上述建立索引的方法獲得的屬性索引，對所述多個項目進行過濾，從而獲得搜尋結果。
在一個實施例中，上述搜尋裝置800還包括排序單元，配置為，在獲得搜尋結果之後，獲取與所述搜尋框對應的業務的排序方法；以及通過所述排序方法對所述搜尋結果進行排序。
圖9示出了一種下拉提示裝置900。所述下拉提示用於使用者終端APP中所包括的至少一個搜尋框，每個搜尋框設置於所述APP中的對應業務中。所述裝置900在使用者終端APP中實施，包括：查詢內容獲取單元91，配置為，獲取使用者在搜尋框中輸入的查詢內容；屬性獲取單元92，配置為，獲取與使用者終端、使用者終端APP或使用者相關的至少一個第一屬性；發送單元93，配置為，將所述查詢內容和所述至少一個第一屬性發送給伺服器端；接收單元94，配置為，從伺服器端接收通過根據上述搜尋方法獲得的搜尋結果；以及顯示單元95，配置為，在所述搜尋框的下拉提示位置顯示所述搜尋結果。
在一個實施例中，上述下拉提示裝置900還包括排序單元，配置為，在從伺服器端接收通過根據上述搜尋方法獲得的搜尋結果之後，獲取與使用者相關的至少一個第二屬性，獲取與所述搜尋框對應的業務的排序方法，以及，根據所述至少一個第二屬性，通過所述排序方法對所述搜尋結果進行排序。
通過根據本說明書實施例的下拉提示方案，通過使用搜尋引擎中的技術實現下拉提示系統，可以更好的支援多種檢索條件(前綴召回，地理位置，以及其他過濾條件)、以及靈活的排序策略(支援不同業務實現不同的排序外掛程式)。同時，本說明書實施例中通過一套通用下拉提示系統支援多個下拉提示業務，其中只需要一份下拉提示索引資料，通過子文件和排序外掛程式的方案，支援業務的靈活性。從而節省了人力和資源成本，提高了迭代效率，並且有足夠的靈活性，可以支援不同業務的不同效果需求。
本領域普通技術人員應該還可以進一步意識到，結合本文中所公開的實施例描述的各示例的單元及演算法步驟，能夠以電子硬體、電腦軟體或者二者的結合來實現，為了清楚地說明硬體和軟體的可互換性，在上述說明中已經按照功能一般性地描述了各示例的組成及步驟。這些功能究竟以硬體還是軟體方式來執行，取決於技術方案的特定應用和設計約束條件。本領域普通技術人員可以對每個特定的應用來使用不同方法來實現所描述的功能，但是這種實現不應認為超出本發明的範圍。
結合本文中所公開的實施例描述的方法或演算法的步驟可以用硬體、處理器執行的軟體模組，或者二者的結合來實施。軟體模組可以置於隨機存取記憶體(RAM)、內部記憶體、唯讀記憶體(ROM)、電可程式化ROM、電可抹除可程式化ROM、暫存器、硬碟、可移動磁碟、CD-ROM、或技術領域內所公知的任意其它形式的儲存媒體中。
以上所述的具體實施方式，對本發明的目的、技術方案和有益效果進行了進一步詳細說明，所應理解的是，以上所述僅為本發明的具體實施方式而已，並不用於限定本發明的保護範圍，凡在本發明的精神和原則之內，所做的任何修改、等同替換、改進等，均應包含在本發明的保護範圍之內。The embodiments of this specification will be described below with reference to the drawings. FIG. 1 shows a schematic diagram of a pull-down prompt system according to an embodiment of the present specification. As shown in the figure, the system includes a user terminal (101, 102, 103) and a server terminal 104. The user terminal may be a user terminal APP, and the APP may include at least one search box, and different search boxes correspond to different services. For example, the user terminal 101 is using the search box corresponding to the service 1 in the APP, the user terminal 102 is using the search box corresponding to the service 2 in the APP, and the user terminal 103 is using the search box corresponding to the service 3 in the APP . The pull-down prompt system according to the embodiment of the present specification provides the pull-down prompt service through the search instance 11 in the server terminal 104. The search example 11 includes a prefix index 111, an attribute index 112, and a sorting plugin 113. The prefix index refers to the reverse index of the prefix words or words included in the project or their pinyin to the project, and the prefix index 111 includes all business items in the APP. The attribute index includes an index between each item and its respective attributes. When the user enters the search content in the search box, the client (101, 102, 103) sends the query content and the attributes related to the user terminal, the user terminal APP, and the user to the server 104 through the network. In the case where the APP includes multiple search boxes, the above-mentioned relevant attributes include the service type, and the service type is determined by the service where the search box entered by the user is located. After the server terminal 104 receives the query content and related attributes, the server terminal 104 obtains the search condition set for the service type and transmits it to the search instance 11 together with the query content and related attributes. The search example 11 searches for items related to the query content according to the prefix index 111, filters the searched items according to the attribute index 112, and according to the search conditions, thereby obtaining search results. For example, the search condition includes searching only items of the business type, so that items that do not belong to the business type are filtered out according to the index between the item and the business type attribute. The search example 11 further includes a sorting plug-in 113, and a sorting method suitable for each business is set in the sorting plug-in 113, so that the search results are sorted by a sorting method corresponding to the type of business. The server side 104 also includes a database 12, which stores items of each business and attributes related to the items. The search instance 11 establishes a prefix index 111 and an attribute index 112 by obtaining data from the database 12. The pull-down prompt system according to this specification will be described below with examples. FIG. 2 shows a method for establishing a drop-down prompt index. The pull-down prompt is used for at least one search box included in the user terminal APP, and each search box is set in a corresponding service in the APP. The method is executed on the server side, more specifically, as shown in FIG. 1, in the search instance 11 of the server side 104. The method includes the following steps: in step S21, obtain a plurality of items for pull-down prompts for each service in the user terminal APP; in step S22, obtain at least one attribute of each item; in step S23, establish A prefix index, wherein the prefix index includes the reverse index of each item's prefix to itself; and in step S24, an attribute index is established, where the attribute index includes an index between each item and its respective attributes. First, in step S21, a plurality of items for pull-down prompts for each service in the user terminal APP are acquired. For example, the Alipay APP includes word-of-mouth business and wealth business. Each business includes a corresponding database on the server side, which is used to store business-related data, and some of these data can be used as drop-down prompt items. For example, in word-of-mouth business, business-related materials include business type, store name, dish name, store address, store geographic location, clicks, and query times. Among them, the shop name and the dish name can be used as items for the drop-down prompt. For example, in wealth business, the fund name can be used as a drop-down reminder item. Therefore, the search example 11 can obtain items that can be used for drop-down prompts in various businesses by reading the database 12. In step S22, at least one attribute of each item is acquired. Items of different businesses have attributes corresponding to the business. For example, the attributes of the word-of-mouth business include: business type (word of mouth), store address, store geographic location, city, number of results, number of clicks, click-through rate, number of queries, etc. The attributes included in the wealth business are: business type (wealth), number of results, total clicks, total click-through rate, clicks in different cities, click-through rates in different cities, etc. These attributes can be represented in different fields. For example, the field "type" indicates the type of business. The items belonging to the word-of-mouth business include the attribute of "type=word of mouth". All projects belonging to the wealth business include the attribute of "type=wealth". Among these attributes, some attribute information is submitted by the store, such as store address, store geographic location, city, etc., and some attribute data is received by the server side from the APP side, such as the number of clicks, the number of queries, etc. The server side stores these attributes in the database 12 in association with the items, so that the search instance 11 can obtain at least one attribute of each item by reading the database 12. In step S23, a prefix index is established, where the prefix index includes the reverse index of each item's prefix to its own. The prefix of an item is a single word before the end of the included item or multiple words in a row. For example, the prefix of the project "Nongfu Mountain Spring" includes "Nong", "Farmer", "Nongfu Mountain", "Husband", "Husband Mountain" and "Mountain". Thus, the reverse index of "agriculture", "farmer", "farmer's mountain", "husband", "husband" and "mountain" to "farmer's mountain spring" can be obtained. After obtaining the items that can be used for the drop-down prompt for all the services in the APP, you can obtain all the prefixes of these items as the key in the index table as described above, and use the corresponding item as the value corresponding to the key in the index table, thereby establishing all A prefix index common to services. In one embodiment, the prefix index also includes the phonetic-to-item index of the prefix. In one embodiment, the maximum semantic granularity words included in the project and their respective prefix-to-project prefix indexes can be established. For example, for the item "Nongfu Mountain Spring is a little sweet", it can be divided into two largest semantic granularity words "Nongfu Mountain Spring" and "A little sweet" according to the semantic dictionary. Therefore, the reverse index of "Nongfu Mountain Spring" and its prefixes "Nong", "Farmer", "Nongfu Mountain", "Husband", "Fushan" and "Mountain" to "Nongfu Mountain Spring is a little sweet" can also be established. Establish a reverse index of "a bit sweet" and its prefixes "yes", "a bit" and "point" to "Nongfu Mountain Spring is a bit sweet" In step S24, an attribute index is established, where the attribute index includes an index between each item and its respective attributes. The index between an item and an attribute can be a reverse index from attribute to item, or a forward index from item to attribute. In one embodiment, a reverse index of business type to project is established. In another embodiment, a reverse index of geographic location to items is established. In another embodiment, a positive index of items to clicks is established. For example, for store items in word-of-mouth business, a positive index table as shown in Table 1 may be established. Table 1 A dish name item in word-of-mouth business may be related to multiple cities. In this case, multiple sub-files can be set for the dish name item to separately index its attributes in different cities. For example, the positive index table for the attribute of the dish name item may be as shown in Table 2. Table 2 In one embodiment, as shown in FIG. 1, after acquiring the business data from the database 12 in the server side 104, the search example 11 establishes a prefix index and an attribute index offline during a predetermined period of time. Generally, the offline full update of the index is performed at a time when there are few users at 2-5 am every day. In one embodiment, the search instance 11 updates the index in real time after acquiring the user real-time update data from the database 12. For example, in the word-of-mouth business, when a new store is added or the store name is updated, the search instance can obtain the store change information from the database 12 in real time by subscribing to the store information, so as to update the information in the index in seconds. FIG. 3 shows a search method for pull-down prompts according to an embodiment of the present specification. The pull-down prompt is used for at least one search box included in the user terminal APP, and each search box is set in a corresponding service in the APP. The method is executed on the server side and includes the following steps: in step S31, receive the query content entered by the user in the search box from the user terminal APP; in step S32, receive the user terminal from the user terminal APP , At least one attribute related to the user terminal APP or the user; in step S33, the search conditions of the service corresponding to the search box are obtained; in step S34, the prefix index obtained according to the above indexing method is used to search for and Multiple items corresponding to the query content; and in step S35, based on the search condition and the at least one attribute, the multiple items are filtered by the attribute index obtained according to the above indexing method to obtain a search result. First, in step S31, the query content input by the user in the search box is received from the user terminal APP. The query content entered by the user in a specific search box in the APP may be a single word, several consecutive characters, or may be pinyin. Every time the user performs an input action, the APP will synchronously send the input content to the server side. In step S32, at least one attribute related to the user terminal, the user terminal APP, or the user is received from the user terminal APP. These attributes may be the geographic location (latitude and longitude) and city where the user terminal is related to the user terminal, the search box service type related to the user terminal APP, and the user query time related to the user. These attributes can be represented by fields. The specific attributes received by the server terminal from the user terminal APP are related to the search box service type. For example, when the search box service type is word-of-mouth service, the server receives attributes such as the search box service type (word of mouth) and the geographic location of the user terminal from the APP. When the search box business type is wealth business, the server receives the search box business type (wealth) and other attributes from the APP. Among them, the service type is determined by the position of the search box used by the user inquiring in the APP. For example, if the user enters the query content in the search box in the word-of-mouth service in the Alipay APP, the search box business type is "type=word of mouth". In one embodiment, the user makes a query in the search box on the homepage of Alipay APP, where the search box service type is a global service, that is, "type=all". In step S33, the search conditions of the service corresponding to the search box are obtained. For different businesses, corresponding search conditions are set. For example, for word-of-mouth services, the search conditions may include: the service type is "word of mouth", and the distance to the user terminal is within a predetermined distance. For wealth business, the search conditions may include that the risk level of the product is at a predetermined level, etc. For a global service, the corresponding search conditions are all search conditions of each service included in the APP. These search conditions can be set on the server side, and the server side obtains the search conditions corresponding to the service type from its own storage location and sends them to the search instance. In one embodiment, the search condition may be set in the user terminal APP, and the server end receives the corresponding search condition from the APP according to the service type of the search box and transmits it to the search instance. In step S34, a plurality of items corresponding to the query content are searched out through the prefix index obtained according to the above indexing method. For example, in a search box in Alipay APP, enter the query content "Ali", then according to the prefix index here, the searched items may include: Alibaba Pictures, Alibaba Health, Arina Pharmaceuticals, Arirang, Alibaba , Ali instant noodles and so on. Among them, "Ali Pictures, Ali Health, and Alina Pharmaceuticals" are projects that belong to the wealth business, and "Arirang, Ali Restaurant, and Ali Noodles" are projects that belong to the word-of-mouth business. In step S35, according to the search condition and the at least one attribute, the multiple items are filtered by the attribute index obtained according to the above indexing method, thereby obtaining a search result. For example, after searching and obtaining multiple items in the previous step S34, "Ali Pictures, Alibaba Health, Arina Pharmaceuticals, Arirang, Ari Restaurant, Ari Noodles", enter a query in the search box of the wealth business by the user In the case of content "Ali", at least one attribute received by the server from the APP includes: search box business type type=wealth, and the search condition of the wealth business includes limiting the project business type to the search box business type "wealth" . Therefore, according to the reverse index of the business type to the project, the project that meets the business type of "wealth" is searched. By taking the intersection of the results of the two searches, the items whose business type is not “wealth” among the items searched through the prefix index above are filtered out, thereby obtaining the search results “Ali Pictures, Alibaba Health, and Arina Pharmaceuticals”. For another example, the user enters "Ali" in the search box of the word-of-mouth service. The attributes received by the server from the APP include the search box service type (word of mouth) and the geographic location of the user terminal, and the search conditions corresponding to the word-of-mouth service are obtained. Including: the business type of the project is word of mouth, and the geographic location of the project is within five kilometers from the user terminal. After searching in the previous step S34 to obtain multiple items "Ali Pictures, Alibaba Health, Arina Pharmaceuticals, Arirang, Ari Restaurant, Ari Noodles", according to the search conditions of word-of-mouth business and the type of business attributes and user geography Location attribute, through the reverse index of the business type to the project, search for items that meet "type=Word of mouth" including "Arirang, Ali restaurant, Ali noodles, Barbera, Brazilian barbecue...", according to the geographical location to the project The reverse index of the search, the location attribute obtained within five kilometers from the user terminal, including "Ali restaurant, Ali noodles, Brazilian barbecue, KFC, Pizza Hut...", take the prefix index and the above two attribute index The intersection of the index results, so as to filter the prefix index results, so as to obtain the search results "Ali food, Ali noodles." In another example, for the above search result "Ali Restaurant, Ali Noodles", the search conditions corresponding to the word-of-mouth business may also include: For example, "the number of queries exceeds 300". Therefore, according to the search condition, and according to the positive index of the store items in the word-of-mouth business to the query times attribute, the respective query times of the items "Ali Restaurant" and "Ali Noodles" are obtained and filtered according to the query times . FIG. 4 schematically shows the above search process. For convenience, in the search results, items that belong to one business type (for example, wealth business) are indicated by capital letters, and another business type (for example, word-of-mouth business) are indicated by lower case letters. s project. In FIG. 4, the prefix index 41, the first attribute index 42 (such as the service type index), and the second attribute index 43 (such as the geographic location index) included in the search example are all reverse indexes. , You can get the corresponding search results {A, B, C, a, b, c}, {a, b, c, d, e} and {b, c, e, f, g}. Take the intersection of the three search results and get the search results {b, c}. In one embodiment, there may be searches related to the third attribute (eg, the number of queries), and the third attribute index 44 may be a forward index. Therefore, according to the search conditions and the third attribute index 44, the above search results {b, c} are further filtered in an up-front row to obtain the final search results, such as {b}. In one embodiment, the number of search results obtained in the above step S35 may be very large, and may be several thousand. In this case, the search results may be sorted first, and then a predetermined number (for example, 500) of the results ranked above may be retrieved and sent to the user terminal APP. In one embodiment, as shown in FIG. 1, the search example 11 according to the embodiment of the present specification includes a ranking plugin 113. In the sorting plug-in 113, different sorting methods are set for different businesses. Before sending the search result to the user terminal APP, the search example 11 obtains a sorting method of services corresponding to the search box from the sorting plugin 113, and sorts the search result by the sorting method. For example, the sorting method corresponding to the store items in the word-of-mouth business may include: first sorting according to the distance between the geographic location attribute of the item and the user terminal, at the next level of the sorting, sorting by the attribute of the number of item queries, and at the next level, passing the item Sort of click attributes. The sorting method of the project corresponding to the wealth business may include: sorting according to the project's rate of return attribute, sorting according to the project's risk level attribute, sorting according to the item's click count attribute, etc. FIG. 5 shows a pull-down prompt method. The method is executed in a user terminal APP. The user terminal APP includes at least one search box, and each search box is set in a corresponding service in the APP. The method includes the following steps: in step S51, obtain the query content entered by the user in the search box; in step S52, obtain at least one attribute related to the user terminal, the user terminal APP, or the user; in step S53, Sending the query content and the at least one first attribute to the server side; in step S54, receiving the search result obtained by the above search method from the server side; and in step S55, pulling down in the search box The prompt position displays the search result. First, in step S51, the query content entered by the user in the search box is obtained. Each time the user performs an input action in the search box, the APP will perform this step to obtain the query content. For example, the user will input "Nongfu Spring". When the user inputs "Nong", the query content obtained by the APP is "Nong". When the user enters "husband" again, the query content obtained by the APP is "farmer". In step S52, at least one attribute related to the user terminal, the user terminal APP, or the user is acquired. At least one attribute here is as described above with reference to FIG. 3, and may be the geographic location (latitude and longitude) and city where the user terminal is related to the user terminal, the search box service type related to the user terminal APP, and the user Related user query time, etc. These attributes can be represented by fields. Similarly, the specific attributes sent by the user terminal APP to the server end are related to the search box service type, and will not be repeated here. For the attributes related to the user terminal, the APP can obtain these attribute data through specific applications, for example, the location of the user terminal is obtained through the positioning application. The attributes related to the APP can be obtained through the settings of the APP itself, for example, by setting a field in the search box of each business to limit the type of the search box business, so as to obtain its corresponding service type directly from the search box. For attributes related to users, different acquisition methods can be used according to different attributes. For example, the user can query the time from the clock in the APP, and for example, some personal characteristics of the user can be obtained from the server. In step S53, the query content and the at least one first attribute are sent to the server. Corresponding to step S51, every time the user performs an input action in the search box, the APP will perform this step, that is, send the query content and related attributes. For example, when the user enters "Nong", the APP sends the acquired query content "Nong" to the server. When the user enters "husband" again, the APP sends the acquired query content "farmer" to the server. In step S54, the search result obtained by the search method described above is received from the server side. In one embodiment, the search results are sorted results. In another embodiment, the search results may be sorted twice in the APP. The APP may also include a sorting plug-in for setting a sorting method corresponding to different services. Compared with the ranking method in the search example, the ranking method can be based on some user characteristic parameters and some more complicated ranking models, for example. For example, for the user’s query on word-of-mouth business, the APP can obtain the user’s consumption-related portrait characteristics, such as age, gender, consumption preferences, and consumption capabilities, from the database on the server side, and according to the user’s portrait characteristics And the corresponding sorting method, to further sort the search results received from the server. For another example, when the search box is located on the homepage of the APP, the corresponding service is a global service, that is, all the services included in the APP also have a ranking method corresponding thereto. In the global business, the global score is calculated for the items in each business according to the corresponding ranking method, and each business item in the APP is sorted according to the global score. In step S55, the search result is displayed at the drop-down prompt position of the search box. In the APP, the drop-down prompt position of the search box is generally below the search box. You can set the number of items or pages in the drop-down prompt, so that the above sorting results can be retrieved from scratch based on the number or page limit. In one embodiment, attributes such as the result number, distance, and address of the item can also be displayed in the drop-down prompt. FIG. 6 schematically shows a schematic diagram of a pull-down prompt effect according to an embodiment of the present specification. Among them, the search box is a search box set in the word-of-mouth business, and the query content entered by the user is "farmer". The drop-down prompt for the word includes attributes such as the number of results, distance, and address. FIG. 7 shows an apparatus 700 for establishing a pull-down prompt index. The pull-down prompt is used for at least one search box included in the user terminal APP, and each search box is set in a corresponding service in the APP. The device 700 is implemented on the server side, and includes: an item acquisition unit 71 configured to acquire multiple items for each business drop-down prompt in the user terminal APP; an attribute acquisition unit 72 configured to acquire At least one attribute of each item; the prefix index establishing unit 73, configured to establish a prefix index, wherein the prefix index includes the reverse index of each item's prefix to itself; and the attribute index establishing unit 74, configured to, Establish an attribute index, where the attribute index includes an index between each item and its respective attributes. In one embodiment, in the device for creating a pull-down prompt index, the attribute index establishing unit 74 is further configured to establish a forward index from the first item to its first attribute. In one embodiment, in the device for creating a drop-down prompt index, the attribute index establishing unit 74 is further configured to establish a reverse index of the second attribute of the second item to the second item. In one embodiment, in the device for creating a drop-down prompt index, the prefix index establishing unit 73 is further configured to establish a prefix index offline; and the attribute index establishing unit 74 is further configured to establish an attribute index offline. In one embodiment, in the device for creating a drop-down prompt index, the attribute index establishing unit 74 is further configured to set a plurality of project sub-files for the first project, for establishing the first project respectively Forward index to the first property in different cities. In one embodiment, in the device for establishing a drop-down prompt index, the prefix index establishing unit 73 is further configured to obtain the maximum semantic granularity word included in the third item, and establish the maximum semantic granularity word and their respective To the reverse index of the third item. FIG. 8 shows a search device 800 for pull-down prompts. The pull-down prompt is used for at least one search box included in the user terminal APP, and each search box is set in a corresponding service in the APP. The device 800 is implemented on the server side and includes: a query content receiving unit 81 configured to receive the query content input by the user in the search box from the user terminal APP; an attribute receiving unit 82 configured to receive the user terminal The APP receives at least one attribute related to the user terminal, the user terminal APP, or the user; the acquiring unit 83 is configured to acquire the search condition of the service corresponding to the search box; the first searching unit 84 is configured to pass According to the prefix index obtained by the above indexing method, multiple items corresponding to the query content are searched out; and the second search unit 85 is configured to, according to the search condition and the at least one attribute, create the index according to the above The attribute index obtained by the method, filtering the multiple items, thereby obtaining search results. In one embodiment, the search device 800 further includes a sorting unit configured to obtain a sorting method of services corresponding to the search box after obtaining a search result; and sort the search results by the sorting method . FIG. 9 shows a pull-down prompt device 900. The pull-down prompt is used for at least one search box included in the user terminal APP, and each search box is set in a corresponding service in the APP. The device 900 is implemented in the user terminal APP and includes: a query content acquisition unit 91 configured to acquire the query content entered by the user in the search box; an attribute acquisition unit 92 configured to acquire the user terminal and use At least one first attribute related to the user terminal APP or user; the sending unit 93 is configured to send the query content and the at least one first attribute to the server; the receiving unit 94 is configured to slave the server The terminal receives the search result obtained by the above search method; and the display unit 95 is configured to display the search result at the drop-down prompt position of the search box. In one embodiment, the pull-down prompt device 900 further includes a sorting unit configured to obtain at least one second attribute related to the user after receiving the search result obtained by the search method from the server side. A method for sorting services corresponding to the search box, and, according to the at least one second attribute, sorting the search results by the sorting method. Through the pull-down prompting scheme according to the embodiment of the present specification, and using the technology in the search engine to implement the pull-down prompting system, it is possible to better support multiple search conditions (prefix recall, geographic location, and other filtering conditions) and flexible sorting strategies ( Support different services to achieve different sequencing plug-ins). At the same time, in the embodiment of the present specification, a set of general pull-down prompt system is used to support multiple pull-down prompt services, of which only one copy of the pull-down prompt index data is needed, and the flexibility of the business is supported through the scheme of sub-documents and sorting plug-ins. Thereby saving manpower and resource costs, improving iteration efficiency, and having enough flexibility to support different effects of different business needs. Those of ordinary skill in the art should also be further aware that the example units and algorithm steps described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two, for clarity The interchangeability of hardware and software is described. In the above description, the composition and steps of each example have been described generally according to functions. Whether these functions are executed in hardware or software depends on the specific application of the technical solution and design constraints. A person of ordinary skill in the art may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention. The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented by hardware, a software module executed by a processor, or a combination of both. Software modules can be placed in random access memory (RAM), internal memory, read only memory (ROM), electrically programmable ROM, electrically erasable and programmable ROM, registers, hard drives, A removable disk, CD-ROM, or any other form of storage medium known in the art. The specific embodiments described above further describe the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above are only specific embodiments of the present invention and are not intended to limit the scope of the present invention. The scope of protection, within the spirit and principle of the present invention, any modification, equivalent replacement, improvement, etc., should be included in the scope of protection of the present invention.
