Documents - Search Post
搜尋索引中的文件。
POST {endpoint}/indexes('{indexName}')/docs/search.post.search?api-version=2025-11-01-preview
URI 參數
| 名稱 | 位於 | 必要 | 類型 | Description |
|---|---|---|---|---|
|
endpoint
|
path | True |
string (uri) |
搜尋服務的端點 URL。 |
|
index
|
path | True |
string |
索引的名稱。 |
|
api-version
|
query | True |
string minLength: 1 |
用於此作業的 API 版本。 |
要求標頭
| 名稱 | 必要 | 類型 | Description |
|---|---|---|---|
| Accept |
接受標頭。 |
||
| x-ms-query-source-authorization |
string |
識別正在執行查詢的使用者的權杖。 此權杖可用來對文件強制執行安全性限制。 |
|
| x-ms-enable-elevated-read |
boolean |
啟用提高許可權讀取的值,可略過查詢作業的文件層級許可權檢查。 |
|
| x-ms-client-request-id |
string (uuid) |
要求不透明、全域唯一、用戶端產生的字串標識碼。 |
要求本文
| 名稱 | 類型 | Description |
|---|---|---|
| answers |
值,指定是否應該在搜尋回應中傳回答案。 |
|
| captions |
值,指定是否應該在搜尋回應中傳回標題。 |
|
| count |
boolean |
值,指定是否要擷取結果的總計數。 預設值為 false。 將此值設定為 true 可能會對效能造成影響。 請注意,傳回的計數是近似值。 |
| debug |
啟用偵錯工具,可用來進一步探索重新排名的結果。 |
|
| facets |
string[] |
要套用至搜尋查詢的Facet表達式清單。 每個 Facet 運算式都包含功能變數名稱,選擇性地後面接著以逗號分隔的名稱:值組清單。 |
| filter |
string |
要套用至搜尋查詢的 OData $filter表示式。 |
| highlight |
string |
要用於點擊醒目提示的功能變數名稱逗號分隔清單。 只有可搜尋的欄位可用於點擊醒目提示。 |
| highlightPostTag |
string |
附加至叫用醒目提示的字串標記。 必須使用 highlightPreTag 進行設定。 預設值為 </em>。 |
| highlightPreTag |
string |
在叫用醒目提示前面加上的字串標記。 必須使用 highlightPostTag 進行設定。 預設值為 em <>。 |
| hybridSearch |
設定混合式搜尋行為的查詢參數。 |
|
| minimumCoverage |
number (double) |
介於 0 到 100 之間的數位,表示搜尋查詢必須涵蓋的索引百分比,以便將查詢回報為成功。 此參數對於確保搜尋可用性,即使是只有一個複本的服務,也很有用。 預設值為 100。 |
| orderby |
string |
以逗號分隔的 OData 清單$orderby用來排序結果的運算式。 每個運算式可以是功能變數名稱或對 geo.distance() 或 search.score() 函式的呼叫。 每個表達式後面可以接著 asc 表示遞增,或 desc 表示遞減。 預設值為遞增順序。 系結會因檔比對分數而中斷。 如果未指定任何$orderby,則預設排序順序會依檔比對分數遞減。 最多可以有 32 個$orderby子句。 |
| queryLanguage |
指定搜尋查詢語言的值。 |
|
| queryRewrites |
指定是否應該產生查詢重寫以擴充搜尋查詢的值。 |
|
| queryType |
值,指定搜尋查詢的語法。 預設值為 『simple』。 如果您的查詢使用 Lucene 查詢語法,請使用 『full』。 |
|
| scoringParameters |
string[] |
要用於評分函式的參數值清單(例如 referencePointParameter),其格式為 name-values。 例如,如果評分配置檔使用名為 'mylocation' 的參數定義函式,則參數位符串會是 “mylocation--122.2,44.8” (不含引號)。 |
| scoringProfile |
string |
要評估比對檔比對分數的評分配置檔名稱,以便排序結果。 |
| scoringStatistics |
值,指定我們是要全域計算評分統計數據(例如文件頻率),以取得更一致的評分,還是針對較低的延遲在本機計算。 預設值為「本機」。 在評分之前,使用 'global' 來彙總全域評分統計資料。 使用全域評分統計資料可能會增加搜尋查詢的延遲。 |
|
| search |
string |
全文檢索搜尋查詢運算式;使用「*」或省略此參數來比對所有文件。 |
| searchFields |
string |
要設定全文搜索範圍之功能變數名稱的逗號分隔清單。 在完整的 Lucene 查詢中使用字段搜尋 (fieldName:searchExpression) 時,每個字段搜尋表達式的功能變數名稱優先於此參數中列出的任何功能變數名稱。 |
| searchMode |
值,指定是否必須比對任何或所有搜尋字詞,才能將檔計算為相符專案。 |
|
| select |
string |
要擷取的欄位以逗點分隔的清單。 如果未指定,則會包含標示為可在架構中擷取的所有欄位。 |
| semanticConfiguration |
string |
處理語意類型查詢文件時將使用的語意配置名稱。 |
| semanticErrorHandling |
允許使用者選擇語意呼叫是否應該完全失敗 (預設/目前行為) ,或傳回部分結果。 |
|
| semanticFields |
string |
用於語意排名的欄位名稱的逗點分隔清單。 |
| semanticMaxWaitInMilliseconds |
integer (int32) minimum: 700 |
容許使用者設定語意擴充在要求失敗之前完成處理所需的時間量上限。 |
| semanticQuery |
string |
允許設定單獨的搜尋查詢,該查詢將僅用於語意重新排序、語意標題和語意答案。 適用於需要在基底擷取和排名階段與 L2 語意階段之間使用不同查詢的案例。 |
| sessionId |
string |
用來建立黏性工作階段的值,有助於取得更一致的結果。 只要使用相同的 sessionId,就會嘗試以相同的副本集為目標。 請謹慎地重複重複使用相同的 sessionID 值,可能會干擾跨複本的要求負載平衡,並對搜尋服務的效能造成負面影響。 做為 sessionId 的值不能以 '_' 字元開頭。 |
| skip |
integer (int32) |
要略過的搜尋結果數目。 這個值不能大於 100,000。 如果您需要依序掃描檔,但因這項限制而無法使用skip,請考慮改為在完全排序的索引鍵上使用orderby,並改用範圍查詢篩選。 |
| speller |
一個用來指定拼字者用來修正個別搜尋詞的類型值。 |
|
| top |
integer (int32) |
要擷取的搜尋結果數目。 這可與$skip搭配使用,以實作搜尋結果的用戶端分頁。 如果結果因為伺服器端分頁而遭到截斷,回應會包含接續令牌,可用來針對下一頁的結果發出另一個搜尋要求。 |
| vectorFilterMode |
決定是否在執行向量搜尋之前或之後套用篩選器。 新索引的預設值為 'preFilter'。 |
|
| vectorQueries | VectorQuery[]: |
向量和混合式搜尋查詢的查詢參數。 |
回應
| 名稱 | 類型 | Description |
|---|---|---|
| 200 OK |
要求已成功。 |
|
| Other Status Codes |
成功 |
|
| Other Status Codes |
未預期的錯誤回應。 |
安全性
api-key
類型:
apiKey
位於:
header
OAuth2Auth
類型:
oauth2
Flow:
implicit
授權 URL:
https://login.microsoftonline.com/common/oauth2/v2.0/authorize
範圍
| 名稱 | Description |
|---|---|
| https://search.azure.com/.default |
範例
|
Search |
|
Search |
SearchIndexSearchDocumentsPost
範例要求
POST https://previewexampleservice.search.windows.net/indexes('preview-test')/docs/search.post.search?api-version=2025-11-01-preview
{
"count": true,
"facets": [
"ownerId",
"price,metric:sum,default:10"
],
"filter": "category eq 'purple' or category eq 'pink'",
"highlight": "category",
"highlightPostTag": "</em>",
"highlightPreTag": "</em>",
"minimumCoverage": 100,
"queryType": "semantic",
"scoringStatistics": "global",
"sessionId": "mysessionid",
"scoringParameters": [
"categoryTag:desiredCategoryValue"
],
"scoringProfile": "stringFieldBoost",
"debug": "vector",
"search": "purple",
"searchFields": "id,name,description,category,ownerId",
"searchMode": "any",
"queryLanguage": "en-us",
"speller": "lexicon",
"select": "id,name,description,category,ownerId",
"skip": 0,
"top": 10,
"semanticConfiguration": "testconfig",
"semanticErrorHandling": "partial",
"semanticMaxWaitInMilliseconds": 5000,
"semanticQuery": "find all purple",
"answers": "extractive",
"captions": "extractive",
"queryRewrites": "generative",
"vectorQueries": [
{
"vector": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9
],
"kind": "vector",
"k": 50,
"fields": "vector22, vector1b",
"oversampling": 20,
"weight": 1,
"threshold": {
"value": 0.984,
"kind": "vectorSimilarity"
},
"filterOverride": "ownerId eq 'sam'"
}
],
"vectorFilterMode": "preFilter",
"hybridSearch": {
"maxTextRecallSize": 100,
"countAndFacetMode": "countAllResults"
}
}
範例回覆
{
"@odata.count": 27,
"@search.coverage": 100,
"@search.facets": {
"ownerId": [
{
"count": 16,
"value": "sam"
},
{
"count": 8,
"value": "ryan"
},
{
"count": 3,
"value": "benny"
}
],
"price": [
{
"sum": 320
}
]
},
"@search.answers": [],
"value": [
{
"@search.score": 0.015625,
"@search.highlights": {
"category": [
"</em>purple</em>"
]
},
"@search.rerankerScore": 1.7178425788879395,
"@search.rerankerBoostedScore": 1.7178425788879395,
"@search.captions": [
{
"text": "test10 hello.",
"highlights": ""
}
],
"@search.documentDebugInfo": {
"vectors": {
"subscores": {
"documentBoost": 1,
"text": {
"searchScore": 27.725889205932617
},
"vectors": [
{}
]
}
}
},
"id": "10",
"name": "test",
"description": "test10 hello",
"category": "purple",
"ownerId": "benny"
},
{
"@search.score": 0.012820512987673283,
"@search.highlights": {
"category": [
"</em>purple</em>"
]
},
"@search.rerankerScore": 1.7079174518585205,
"@search.rerankerBoostedScore": 1.7079174518585205,
"@search.captions": [
{
"text": "no vector.",
"highlights": "</em>no vector.</em>"
}
],
"@search.documentDebugInfo": {
"vectors": {
"subscores": {
"documentBoost": 1,
"text": {
"searchScore": 21.559860229492188
},
"vectors": [
{}
]
}
}
},
"id": "empty-vectors",
"name": "test",
"description": "no vector",
"category": "purple",
"ownerId": "sam"
},
{
"@search.score": 0.011627906933426857,
"@search.highlights": {
"category": [
"</em>purple</em>"
]
},
"@search.rerankerScore": 1.7079174518585205,
"@search.rerankerBoostedScore": 1.7079174518585205,
"@search.captions": [
{
"text": "no vector.",
"highlights": "</em>no vector.</em>"
}
],
"@search.documentDebugInfo": {
"vectors": {
"subscores": {
"documentBoost": 1,
"text": {
"searchScore": 6.500757217407227
},
"vectors": [
{}
]
}
}
},
"id": "no-vectors",
"name": "test",
"description": "no vector",
"category": "purple",
"ownerId": "sam"
},
{
"@search.score": 0.011904762126505375,
"@search.highlights": {
"category": [
"</em>purple</em>"
]
},
"@search.rerankerScore": 1.6994102001190186,
"@search.rerankerBoostedScore": 1.6994102001190186,
"@search.captions": [
{
"text": "test4 hello.",
"highlights": ""
}
],
"@search.documentDebugInfo": {
"vectors": {
"subscores": {
"documentBoost": 1,
"text": {
"searchScore": 6.500757217407227
},
"vectors": [
{}
]
}
}
},
"id": "4",
"name": "test",
"description": "test4 hello",
"category": "purple",
"ownerId": "sam"
},
{
"@search.score": 0.012345679104328156,
"@search.highlights": {
"category": [
"</em>purple</em>"
]
},
"@search.rerankerScore": 1.6937386989593506,
"@search.rerankerBoostedScore": 1.6937386989593506,
"@search.captions": [
{
"text": "test2 hello.",
"highlights": ""
}
],
"@search.documentDebugInfo": {
"vectors": {
"subscores": {
"documentBoost": 1,
"text": {
"searchScore": 6.500757217407227
},
"vectors": [
{}
]
}
}
},
"id": "2",
"name": "test",
"description": "test2 hello",
"category": "purple",
"ownerId": "sam"
},
{
"@search.score": 0.0117647061124444,
"@search.highlights": {
"category": [
"</em>purple</em>"
]
},
"@search.rerankerScore": 1.6923208236694336,
"@search.rerankerBoostedScore": 1.6923208236694336,
"@search.captions": [
{
"text": "test5 hello.",
"highlights": ""
}
],
"@search.documentDebugInfo": {
"vectors": {
"subscores": {
"documentBoost": 1,
"text": {
"searchScore": 6.500757217407227
},
"vectors": [
{}
]
}
}
},
"id": "5",
"name": "test",
"description": "test5 hello",
"category": "purple",
"ownerId": "sam"
},
{
"@search.score": 0.01666666753590107,
"@search.highlights": {
"category": [
"</em>purple</em>"
]
},
"@search.rerankerScore": 1.691611886024475,
"@search.rerankerBoostedScore": 1.691611886024475,
"@search.captions": [
{
"text": "test7 hello.",
"highlights": ""
}
],
"@search.documentDebugInfo": {
"vectors": {
"subscores": {
"documentBoost": 1,
"text": {
"searchScore": 39.23316955566406
},
"vectors": [
{}
]
}
}
},
"id": "7",
"name": "test",
"description": "test7 hello",
"category": "purple",
"ownerId": "sam"
},
{
"@search.score": 0.013698630034923553,
"@search.highlights": {
"category": [
"</em>purple</em>"
]
},
"@search.rerankerScore": 1.6781420707702637,
"@search.rerankerBoostedScore": 1.6781420707702637,
"@search.captions": [
{
"text": "test0 hello.",
"highlights": ""
}
],
"@search.documentDebugInfo": {
"vectors": {
"subscores": {
"documentBoost": 1,
"text": {
"searchScore": 21.559860229492188
},
"vectors": [
{}
]
}
}
},
"id": "0",
"name": "test",
"description": "test0 hello",
"category": "purple",
"ownerId": "benny"
},
{
"@search.score": 0.013888888992369175,
"@search.highlights": {
"category": [
"</em>purple</em>"
]
},
"@search.rerankerScore": 1.6653810739517212,
"@search.rerankerBoostedScore": 1.6653810739517212,
"@search.captions": [
{
"text": "test8 hello.",
"highlights": ""
}
],
"@search.documentDebugInfo": {
"vectors": {
"subscores": {
"documentBoost": 1,
"text": {
"searchScore": 23.01456642150879
},
"vectors": [
{}
]
}
}
},
"id": "8",
"name": "test",
"description": "test8 hello",
"category": "purple",
"ownerId": "sam"
},
{
"@search.score": 0.012658228166401386,
"@search.highlights": {
"category": [
"</em>purple</em>"
]
},
"@search.rerankerScore": 1.6618363857269287,
"@search.rerankerBoostedScore": 1.6618363857269287,
"@search.captions": [
{
"text": "test11 hello.",
"highlights": ""
}
],
"@search.documentDebugInfo": {
"vectors": {
"subscores": {
"documentBoost": 1,
"text": {
"searchScore": 6.500757217407227
},
"vectors": [
{}
]
}
}
},
"id": "11",
"name": "test",
"description": "test11 hello",
"category": "purple",
"ownerId": "sam"
}
]
}
{
"@odata.count": 27,
"@search.coverage": 100,
"@search.facets": {
"ownerId": [
{
"count": 16,
"value": "sam"
},
{
"count": 8,
"value": "ryan"
},
{
"count": 3,
"value": "benny"
}
],
"price": [
{
"sum": 320
}
]
},
"@search.answers": [],
"value": [
{
"@search.score": 0.015625,
"@search.highlights": {
"category": [
"</em>purple</em>"
]
},
"@search.rerankerScore": 1.7178425788879395,
"@search.rerankerBoostedScore": 1.7178425788879395,
"@search.captions": [
{
"text": "test10 hello.",
"highlights": ""
}
],
"@search.documentDebugInfo": {
"vectors": {
"subscores": {
"documentBoost": 1,
"text": {
"searchScore": 27.725889205932617
},
"vectors": [
{}
]
}
}
},
"id": "10",
"name": "test",
"description": "test10 hello",
"category": "purple",
"ownerId": "benny"
},
{
"@search.score": 0.012820512987673283,
"@search.highlights": {
"category": [
"</em>purple</em>"
]
},
"@search.rerankerScore": 1.7079174518585205,
"@search.rerankerBoostedScore": 1.7079174518585205,
"@search.captions": [
{
"text": "no vector.",
"highlights": "</em>no vector.</em>"
}
],
"@search.documentDebugInfo": {
"vectors": {
"subscores": {
"documentBoost": 1,
"text": {
"searchScore": 21.559860229492188
},
"vectors": [
{}
]
}
}
},
"id": "empty-vectors",
"name": "test",
"description": "no vector",
"category": "purple",
"ownerId": "sam"
},
{
"@search.score": 0.011627906933426857,
"@search.highlights": {
"category": [
"</em>purple</em>"
]
},
"@search.rerankerScore": 1.7079174518585205,
"@search.rerankerBoostedScore": 1.7079174518585205,
"@search.captions": [
{
"text": "no vector.",
"highlights": "</em>no vector.</em>"
}
],
"@search.documentDebugInfo": {
"vectors": {
"subscores": {
"documentBoost": 1,
"text": {
"searchScore": 6.500757217407227
},
"vectors": [
{}
]
}
}
},
"id": "no-vectors",
"name": "test",
"description": "no vector",
"category": "purple",
"ownerId": "sam"
},
{
"@search.score": 0.011904762126505375,
"@search.highlights": {
"category": [
"</em>purple</em>"
]
},
"@search.rerankerScore": 1.6994102001190186,
"@search.rerankerBoostedScore": 1.6994102001190186,
"@search.captions": [
{
"text": "test4 hello.",
"highlights": ""
}
],
"@search.documentDebugInfo": {
"vectors": {
"subscores": {
"documentBoost": 1,
"text": {
"searchScore": 6.500757217407227
},
"vectors": [
{}
]
}
}
},
"id": "4",
"name": "test",
"description": "test4 hello",
"category": "purple",
"ownerId": "sam"
},
{
"@search.score": 0.012345679104328156,
"@search.highlights": {
"category": [
"</em>purple</em>"
]
},
"@search.rerankerScore": 1.6937386989593506,
"@search.rerankerBoostedScore": 1.6937386989593506,
"@search.captions": [
{
"text": "test2 hello.",
"highlights": ""
}
],
"@search.documentDebugInfo": {
"vectors": {
"subscores": {
"documentBoost": 1,
"text": {
"searchScore": 6.500757217407227
},
"vectors": [
{}
]
}
}
},
"id": "2",
"name": "test",
"description": "test2 hello",
"category": "purple",
"ownerId": "sam"
},
{
"@search.score": 0.0117647061124444,
"@search.highlights": {
"category": [
"</em>purple</em>"
]
},
"@search.rerankerScore": 1.6923208236694336,
"@search.rerankerBoostedScore": 1.6923208236694336,
"@search.captions": [
{
"text": "test5 hello.",
"highlights": ""
}
],
"@search.documentDebugInfo": {
"vectors": {
"subscores": {
"documentBoost": 1,
"text": {
"searchScore": 6.500757217407227
},
"vectors": [
{}
]
}
}
},
"id": "5",
"name": "test",
"description": "test5 hello",
"category": "purple",
"ownerId": "sam"
},
{
"@search.score": 0.01666666753590107,
"@search.highlights": {
"category": [
"</em>purple</em>"
]
},
"@search.rerankerScore": 1.691611886024475,
"@search.rerankerBoostedScore": 1.691611886024475,
"@search.captions": [
{
"text": "test7 hello.",
"highlights": ""
}
],
"@search.documentDebugInfo": {
"vectors": {
"subscores": {
"documentBoost": 1,
"text": {
"searchScore": 39.23316955566406
},
"vectors": [
{}
]
}
}
},
"id": "7",
"name": "test",
"description": "test7 hello",
"category": "purple",
"ownerId": "sam"
},
{
"@search.score": 0.013698630034923553,
"@search.highlights": {
"category": [
"</em>purple</em>"
]
},
"@search.rerankerScore": 1.6781420707702637,
"@search.rerankerBoostedScore": 1.6781420707702637,
"@search.captions": [
{
"text": "test0 hello.",
"highlights": ""
}
],
"@search.documentDebugInfo": {
"vectors": {
"subscores": {
"documentBoost": 1,
"text": {
"searchScore": 21.559860229492188
},
"vectors": [
{}
]
}
}
},
"id": "0",
"name": "test",
"description": "test0 hello",
"category": "purple",
"ownerId": "benny"
},
{
"@search.score": 0.013888888992369175,
"@search.highlights": {
"category": [
"</em>purple</em>"
]
},
"@search.rerankerScore": 1.6653810739517212,
"@search.rerankerBoostedScore": 1.6653810739517212,
"@search.captions": [
{
"text": "test8 hello.",
"highlights": ""
}
],
"@search.documentDebugInfo": {
"vectors": {
"subscores": {
"documentBoost": 1,
"text": {
"searchScore": 23.01456642150879
},
"vectors": [
{}
]
}
}
},
"id": "8",
"name": "test",
"description": "test8 hello",
"category": "purple",
"ownerId": "sam"
},
{
"@search.score": 0.012658228166401386,
"@search.highlights": {
"category": [
"</em>purple</em>"
]
},
"@search.rerankerScore": 1.6618363857269287,
"@search.rerankerBoostedScore": 1.6618363857269287,
"@search.captions": [
{
"text": "test11 hello.",
"highlights": ""
}
],
"@search.documentDebugInfo": {
"vectors": {
"subscores": {
"documentBoost": 1,
"text": {
"searchScore": 6.500757217407227
},
"vectors": [
{}
]
}
}
},
"id": "11",
"name": "test",
"description": "test11 hello",
"category": "purple",
"ownerId": "sam"
}
]
}
SearchIndexSearchDocumentsSemanticPost
範例要求
POST https://myservice.search.windows.net/indexes('myindex')/docs/search.post.search?api-version=2025-11-01-preview
{
"count": true,
"highlightPostTag": "</em>",
"highlightPreTag": "<em>",
"queryType": "semantic",
"search": "how do clouds form",
"semanticConfiguration": "my-semantic-config",
"answers": "extractive|count-3",
"captions": "extractive|highlight-true",
"semanticErrorHandling": "partial",
"semanticMaxWaitInMilliseconds": 780
}
範例回覆
{
"@odata.count": 25,
"@search.answers": [
{
"key": "4123",
"text": "Sunlight heats the land all day, warming that moist air and causing it to rise high into the atmosphere until it cools and condenses into water droplets. Clouds generally form where air is ascending (over land in this case), but not where it is descending (over the river).",
"highlights": "Sunlight heats the land all day, warming that moist air and causing it to rise high into the atmosphere until it cools and condenses into water droplets. Clouds generally form<em> where air is ascending</em> (over land in this case), but not where it is<em> descending</em> (over the river).",
"score": 0.94639826
}
],
"@search.nextPageParameters": {
"count": true,
"highlightPostTag": "</em>",
"highlightPreTag": "<em>",
"queryType": "semantic",
"search": "how do clouds form",
"semanticConfiguration": "my-semantic-config",
"answers": "extractive|count-3",
"captions": "extractive|highlight-true",
"semanticErrorHandling": "partial",
"semanticMaxWaitInMilliseconds": 780,
"skip": 2,
"top": 8
},
"value": [
{
"@search.score": 0.5479723,
"@search.rerankerScore": 1.0321671911515296,
"@search.captions": [
{
"text": "Like all clouds, it forms when the air reaches its dew point—the temperature at which an air mass is cool enough for its water vapor to condense into liquid droplets. This false-color image shows valley fog, which is common in the Pacific Northwest of North America.",
"highlights": "Like all<em> clouds</em>, it<em> forms</em> when the air reaches its dew point—the temperature at which an air mass is cool enough for its water vapor to condense into liquid droplets. This false-color image shows valley<em> fog</em>, which is common in the Pacific Northwest of North America."
}
],
"id": "4123",
"title": "Earth Atmosphere",
"content": "Fog is essentially a cloud lying on the ground. Like all clouds, it forms when the air reaches its dew point—the temperature at \n\nwhich an air mass is cool enough for its water vapor to condense into liquid droplets.\n\nThis false-color image shows valley fog, which is common in the Pacific Northwest of North America. On clear winter nights, the \n\nground and overlying air cool off rapidly, especially at high elevations. Cold air is denser than warm air, and it sinks down into the \n\nvalleys. The moist air in the valleys gets chilled to its dew point, and fog forms. If undisturbed by winds, such fog may persist for \n\ndays. The Terra satellite captured this image of foggy valleys northeast of Vancouver in February 2010.\n\n\n",
"locations": [
"Pacific Northwest",
"North America",
"Vancouver"
]
}
],
"@odata.nextLink": "https://myservice.search.windows.net/indexes('myindex')/docs/search.post.search?api-version=2025-05-01-preview"
}
{
"@odata.count": 25,
"@search.answers": [
{
"key": "4123",
"text": "Sunlight heats the land all day, warming that moist air and causing it to rise high into the atmosphere until it cools and condenses into water droplets. Clouds generally form where air is ascending (over land in this case), but not where it is descending (over the river).",
"highlights": "Sunlight heats the land all day, warming that moist air and causing it to rise high into the atmosphere until it cools and condenses into water droplets. Clouds generally form<em> where air is ascending</em> (over land in this case), but not where it is<em> descending</em> (over the river).",
"score": 0.94639826
}
],
"@search.nextPageParameters": {
"count": true,
"highlightPostTag": "</em>",
"highlightPreTag": "<em>",
"queryType": "semantic",
"search": "how do clouds form",
"semanticConfiguration": "my-semantic-config",
"answers": "extractive|count-3",
"captions": "extractive|highlight-true",
"semanticErrorHandling": "partial",
"semanticMaxWaitInMilliseconds": 780,
"skip": 2,
"top": 8
},
"value": [
{
"@search.score": 0.5479723,
"@search.rerankerScore": 1.0321671911515296,
"@search.captions": [
{
"text": "Like all clouds, it forms when the air reaches its dew point—the temperature at which an air mass is cool enough for its water vapor to condense into liquid droplets. This false-color image shows valley fog, which is common in the Pacific Northwest of North America.",
"highlights": "Like all<em> clouds</em>, it<em> forms</em> when the air reaches its dew point—the temperature at which an air mass is cool enough for its water vapor to condense into liquid droplets. This false-color image shows valley<em> fog</em>, which is common in the Pacific Northwest of North America."
}
],
"id": "4123",
"title": "Earth Atmosphere",
"content": "Fog is essentially a cloud lying on the ground. Like all clouds, it forms when the air reaches its dew point—the temperature at \n\nwhich an air mass is cool enough for its water vapor to condense into liquid droplets.\n\nThis false-color image shows valley fog, which is common in the Pacific Northwest of North America. On clear winter nights, the \n\nground and overlying air cool off rapidly, especially at high elevations. Cold air is denser than warm air, and it sinks down into the \n\nvalleys. The moist air in the valleys gets chilled to its dew point, and fog forms. If undisturbed by winds, such fog may persist for \n\ndays. The Terra satellite captured this image of foggy valleys northeast of Vancouver in February 2010.\n\n\n",
"locations": [
"Pacific Northwest",
"North America",
"Vancouver"
]
}
],
"@odata.nextLink": "https://myservice.search.windows.net/indexes('myindex')/docs/search.post.search?api-version=2025-05-01-preview"
}
定義
| 名稱 | Description |
|---|---|
| Accept |
接受標頭。 |
|
Debug |
包含可用來進一步探索搜尋結果的偵錯資訊。 |
|
Document |
包含可用來進一步探索搜尋結果的偵錯資訊。 |
|
Error |
資源管理錯誤其他資訊。 |
|
Error |
錯誤詳細數據。 |
|
Error |
所有 Azure Resource Manager API 的常見錯誤回應,用於回傳失敗操作的錯誤細節。 (這也遵循 OData 錯誤回應格式。)。 |
|
Hybrid |
判斷計數和 Facet 是否應該包含符合搜尋查詢的所有檔,或只包含在 'maxTextRecallSize' 視窗內擷取的檔。 預設值為 'countAllResults'。 |
|
Hybrid |
設定混合式搜尋行為的查詢參數。 |
|
Query |
答案是從符合查詢的最相關文件內容擷取的文字段落。 答案會從頂端搜尋結果擷取。 答案候選項目會進行評分,並選取最上方的答案。 |
|
Query |
只有在查詢類型為 |
|
Query |
標題是文件相對於搜尋查詢最具代表性的段落。 它們通常用來做為文件摘要。 標題只會針對類型為 |
|
Query |
只有在查詢類型為 |
|
Query |
啟用偵錯工具,可用來進一步探索搜尋結果。 您可以同時啟用多個偵錯模式,方法是使用 |字元,例如:semantic|queryRewrites。 |
|
Query |
查詢的語言。 |
|
Query |
傳送至語意擴充程式的原始串連字串。 |
|
Query |
傳送至語意擴充程式的字段描述,以及其使用方式 |
|
Query |
本檔搜尋查詢的文字和向量查詢元件之間的子核心明細。 每個向量查詢都會以收到的相同順序顯示為個別物件。 |
|
Query |
包含查詢重寫的特定偵錯資訊。 |
|
Query |
只有在查詢類型為 |
|
Query |
包含查詢重寫的特定偵錯資訊。 |
|
Query |
藉由拼字更正個別搜尋查詢字詞來改善搜尋召回率。 |
|
Query |
指定搜尋查詢的語法。 預設值為 『simple』。 如果您的查詢使用 Lucene 查詢語法,請使用 'full',如果不需要查詢語法,請使用 'semantic'。 |
|
Scoring |
值,指定我們是要全域計算評分統計數據(例如文件頻率),以取得更一致的評分,還是針對較低的延遲在本機計算。 預設值為「本機」。 在評分之前,使用 'global' 來彙總全域評分統計資料。 使用全域評分統計資料可能會增加搜尋查詢的延遲。 |
|
Search |
包含索引搜尋結果的回應。 |
|
Search |
指定是否必須比對任何或所有搜尋字詞,才能將文件計為相符專案。 |
|
Search |
篩選、排序、Facet、分頁和其他搜尋查詢行為的參數。 |
|
Search |
包含搜尋查詢所找到的檔,加上相關聯的元數據。 |
|
Search |
向量查詢的結果會根據 ' 進行篩選 |
|
Semantic |
包含語意排名要求特定的偵錯資訊。 |
|
Semantic |
允許使用者選擇語意呼叫是否完全失敗,或回傳部分結果。 |
|
Semantic |
針對語意排名要求傳回部分回應的原因。 |
|
Semantic |
該場在語意豐富過程中的使用方式。 |
|
Semantic |
用於此要求的查詢重寫類型。 |
|
Semantic |
針對語意排名要求傳回的部分回應類型。 |
|
Single |
單一向量欄位結果。 Both |
|
Text |
查詢文字部分的 BM25 或傳統分數。 |
|
Vector |
決定是否在執行向量搜尋之前或之後套用篩選器。 |
|
Vectorizable |
提供需要向量化之影像的基底 64 編碼二進位檔時,要用於向量搜尋的查詢參數。 |
|
Vectorizable |
當提供代表需要向量化之影像值的 URL 時,要用於向量搜尋的查詢參數。 |
|
Vectorizable |
提供需要向量化的文字值時,要用於向量搜尋的查詢參數。 |
|
Vectorized |
提供原始向量值時,用於向量搜尋的查詢參數。 |
|
Vector |
正在執行的向量查詢類型。 |
|
Vectors |
「包含針對向量與混合搜尋的除錯資訊。」) |
|
Vector |
系統會根據向量相似度計量來篩選向量查詢的結果。 請注意,這是相似度計量的正式定義,而不是「距離」版本。 系統會根據欄位所使用的計量自動選擇臨界值方向(較大或更小)。 |
|
Vector |
那種用來篩選向量查詢的門檻。 |
Accept
接受標頭。
| 值 | Description |
|---|---|
| application/json;odata.metadata=none |
DebugInfo
包含可用來進一步探索搜尋結果的偵錯資訊。
| 名稱 | 類型 | Description |
|---|---|---|
| queryRewrites |
包含查詢重寫的特定偵錯資訊。 |
DocumentDebugInfo
包含可用來進一步探索搜尋結果的偵錯資訊。
| 名稱 | 類型 | Description |
|---|---|---|
| innerHits |
object |
包含複雜類型集合內相符之向量的特定偵錯資訊。 |
| semantic |
包含語意排名要求特定的偵錯資訊。 |
|
| vectors |
包含向量和混合式搜尋特有的偵錯資訊。 |
ErrorAdditionalInfo
資源管理錯誤其他資訊。
| 名稱 | 類型 | Description |
|---|---|---|
| info |
附加資訊。 |
|
| type |
string |
其他資訊類型。 |
ErrorDetail
錯誤詳細數據。
| 名稱 | 類型 | Description |
|---|---|---|
| additionalInfo |
錯誤附加資訊。 |
|
| code |
string |
錯誤碼。 |
| details |
錯誤詳細數據。 |
|
| message |
string |
錯誤訊息。 |
| target |
string |
錯誤目標。 |
ErrorResponse
所有 Azure Resource Manager API 的常見錯誤回應,用於回傳失敗操作的錯誤細節。 (這也遵循 OData 錯誤回應格式。)。
| 名稱 | 類型 | Description |
|---|---|---|
| error |
錯誤物件。 |
HybridCountAndFacetMode
判斷計數和 Facet 是否應該包含符合搜尋查詢的所有檔,或只包含在 'maxTextRecallSize' 視窗內擷取的檔。 預設值為 'countAllResults'。
| 值 | Description |
|---|---|
| countRetrievableResults |
在計算 'count' 和 'facets' 時,僅包含在 'maxTextRecallSize' 擷取視窗內相符的文件。 |
| countAllResults |
在計算 'count' 和 'facets' 時,包括搜尋查詢所比對的所有文件,無論這些文件是否在 'maxTextRecallSize' 擷取視窗內。 |
HybridSearch
設定混合式搜尋行為的查詢參數。
| 名稱 | 類型 | Description |
|---|---|---|
| countAndFacetMode |
判斷計數和 Facet 是否應該包含符合搜尋查詢的所有檔,或只包含在 'maxTextRecallSize' 視窗內擷取的檔。 |
|
| maxTextRecallSize |
integer (int32) |
決定混合式搜尋要求的文字查詢部分要擷取的文件數目上限。 這些文件將與符合向量查詢的文件結合,以產生單一最終結果清單。 選擇較大的 maxTextRecallSize 值將允許擷取和分頁更多文件 (使用 top 和 skip 參數),但代價是更高的資源使用率和更高的延遲。 值必須介於 1 到 10,000 之間。 預設值為 1000。 |
QueryAnswerResult
答案是從符合查詢的最相關文件內容擷取的文字段落。 答案會從頂端搜尋結果擷取。 答案候選項目會進行評分,並選取最上方的答案。
| 名稱 | 類型 | Description |
|---|---|---|
| highlights |
string |
與 Text 屬性中的文字段落相同,其中醒目提示的文字片語與查詢最相關。 |
| key |
string |
從中提取答案的文件的密鑰。 |
| score |
number (double) |
分數值表示答案相對於針對查詢傳回的其他答案,與查詢的相關性。 |
| text |
string |
從文檔內容中提取的文本段落作為答案。 |
QueryAnswerType
只有在查詢類型為 semantic時,此參數才有效。 如果設定,則查詢會傳回從最高排名文件中的重要段落擷取的答案。 傳回的答案數目可以藉由附加管道字元 | ,後面加上 count-<number of answers> answers 參數值後面的選項, extractive|count-3例如 。 預設計數為 1。 信賴度臨界值可以藉由附加管道字元 | ,後面加上 threshold-<confidence threshold> answers 參數值後面的選項, extractive|threshold-0.9例如 。 預設臨界值為 0.7。 答案的最大字元長度可以透過附加管道字元 '|' 後面跟著 '最大字元長度<的計數>數' (例如 'extractive|maxcharlength-600') 來配置。
| 值 | Description |
|---|---|
| none |
請勿傳回查詢的答案。 |
| extractive |
從傳回的文件內容擷取回應以自然語言表示為問題之查詢的回應候選專案。 |
QueryCaptionResult
標題是文件相對於搜尋查詢最具代表性的段落。 它們通常用來做為文件摘要。 標題只會針對類型為 semantic的查詢傳回。
| 名稱 | 類型 | Description |
|---|---|---|
| highlights |
string |
與 Text 屬性中的文字段落相同,其中醒目提示的片語與查詢最相關。 |
| text |
string |
從與搜尋查詢最相關的文件中擷取的代表性文字段落。 |
QueryCaptionType
只有在查詢類型為 semantic時,此參數才有效。 如果設定,查詢會傳回從最高排名檔中索引鍵段落擷取的標題。 當「標題」設定為 extractive時,預設會啟用反白顯示,並且可以透過附加管道字元 | 後面接著 highlight-<true/false> 選項來設定,例如 extractive|highlight-true。 預設為 None。 標題的最大字元長度可以透過附加管道字元 '|' 後面跟著 '最大字元長度<的計數>數' (例如 'extractive|maxcharlength-600') 來配置。
| 值 | Description |
|---|---|
| none |
請勿傳回查詢的標題。 |
| extractive |
從相符的檔擷取標題,其中包含與搜尋查詢相關的段落。 |
QueryDebugMode
啟用偵錯工具,可用來進一步探索搜尋結果。 您可以同時啟用多個偵錯模式,方法是使用 |字元,例如:semantic|queryRewrites。
| 值 | Description |
|---|---|
| disabled |
不會傳回任何查詢偵錯資訊。 |
| semantic |
允許用戶進一步探索其重新建立的結果。 |
| vector |
允許用戶進一步探索其混合式和向量查詢結果。 |
| queryRewrites |
允許使用者探索針對其搜尋要求所產生的查詢重寫清單。 |
| innerHits |
允許使用者擷取複雜類型集合中相符之向量的評分資訊。 |
| all |
開啟所有偵錯選項。 |
QueryLanguage
查詢的語言。
| 值 | Description |
|---|---|
| none |
未指定查詢語言。 |
| en-us |
查詢英語語言值(美國)。 |
| en-gb |
查詢英文 (英國) 的語言值。 |
| en-in |
查詢英文 (印度) 的語言值。 |
| en-ca |
查詢英文 (加拿大) 的語言值。 |
| en-au |
查詢英文 (澳洲) 的語言值。 |
| fr-fr |
查詢法文 (法國) 的語言值。 |
| fr-ca |
查詢法文 (加拿大) 的語言值。 |
| de-de |
查詢德文 (德國) 的語言值。 |
| es-es |
查詢西班牙文 (西班牙) 的語言值。 |
| es-mx |
查詢西班牙文 (墨西哥) 的語言值。 |
| zh-cn |
查詢中文 (中國) 的語言值。 |
| zh-tw |
查詢中文 (台灣) 的語言值。 |
| pt-br |
查詢葡萄牙語 (巴西) 的語言值。 |
| pt-pt |
查詢葡萄牙語 (葡萄牙) 的語言值。 |
| it-it |
查詢義大利語 (義大利) 的語言值。 |
| ja-jp |
查詢日文 (日本) 的語言值。 |
| ko-kr |
查詢韓語 (韓國) 的語言值。 |
| ru-ru |
查詢俄語 (俄羅斯) 的語言值。 |
| cs-cz |
捷克語 (捷克共和國) 的查詢語言值。 |
| nl-be |
查詢荷蘭語 (比利時) 的語言值。 |
| nl-nl |
查詢荷蘭語 (荷蘭) 的語言值。 |
| hu-hu |
查詢匈牙利文 (匈牙利) 的語言值。 |
| pl-pl |
查詢波蘭語 (波蘭) 的語言值。 |
| sv-se |
查詢瑞典語 (瑞典) 的語言值。 |
| tr-tr |
查詢土耳其語 (土耳其) 的語言值。 |
| hi-in |
查詢印地語 (印度) 的語言值。 |
| ar-sa |
查詢阿拉伯文 (沙烏地阿拉伯) 的語言值。 |
| ar-eg |
查詢阿拉伯文 (埃及) 的語言值。 |
| ar-ma |
查詢阿拉伯文 (摩洛哥) 的語言值。 |
| ar-kw |
查詢阿拉伯文 (科威特) 的語言值。 |
| ar-jo |
查詢阿拉伯文 (約旦) 的語言值。 |
| da-dk |
查詢丹麥文 (丹麥) 的語言值。 |
| no-no |
挪威文 (挪威) 的查詢語言值。 |
| bg-bg |
查詢保加利亞文 (保加利亞) 的語言值。 |
| hr-hr |
查詢克羅埃西亞文 (克羅埃西亞) 的語言值。 |
| hr-ba |
查詢克羅埃西亞文(波士尼亞和黑塞哥維那)的語言值。 |
| ms-my |
查詢馬來語 (馬來西亞) 的語言值。 |
| ms-bn |
查詢馬來語 (汶萊達魯薩蘭語) 的語言值。 |
| sl-sl |
查詢斯洛文尼亞語 (Slovenia) 的語言值。 |
| ta-in |
泰米爾語 (印度) 的查詢語言值。 |
| vi-vn |
查詢越南語(越南)的語言值。 |
| el-gr |
查詢希臘文 (希臘) 的語言值。 |
| ro-ro |
查詢羅馬尼亞語 (羅馬尼亞) 的語言值。 |
| is-is |
查詢冰島語 (Iceland) 的語言值。 |
| id-id |
印尼文 (印尼) 的查詢語言值。 |
| th-th |
查詢泰語 (泰國) 的語言值。 |
| lt-lt |
查詢立陶宛文 (Lithuania) 的語言值。 |
| uk-ua |
查詢烏克蘭文 (烏克蘭) 的語言值。 |
| lv-lv |
查詢拉脫維亞語 (Latvia) 的語言值。 |
| et-ee |
查詢愛沙尼亞語 (愛沙尼亞) 的語言值。 |
| ca-es |
查詢加泰隆隆語的語言值。 |
| fi-fi |
查詢芬蘭文 (芬蘭) 的語言值。 |
| sr-ba |
查詢塞爾維亞語 (波士尼亞和黑塞哥維那) 的語言值。 |
| sr-me |
查詢塞爾維亞語 (黑山) 的語言值。 |
| sr-rs |
查詢塞爾維亞文 (塞爾維亞) 的語言值。 |
| sk-sk |
查詢斯洛伐克語 (斯洛伐克) 的語言值。 |
| nb-no |
挪威文 (挪威) 的查詢語言值。 |
| hy-am |
查詢亞美尼亞語 (亞美尼亞) 的語言值。 |
| bn-in |
孟加拉文 (印度) 的查詢語言值。 |
| eu-es |
查詢巴斯克語的語言值。 |
| gl-es |
查詢加利西亞語的語言值。 |
| gu-in |
古吉拉特語 (印度) 的查詢語言值。 |
| he-il |
查詢希伯來文 (以色列) 的語言值。 |
| ga-ie |
查詢愛爾蘭語 (愛爾蘭) 的語言值。 |
| kn-in |
查詢卡納達語 (印度) 的語言值。 |
| ml-in |
查詢馬拉雅拉姆語 (印度) 的語言值。 |
| mr-in |
查詢馬拉地語 (印度) 的語言值。 |
| fa-ae |
查詢波斯語 (UAE) 的語言值。 |
| pa-in |
查詢旁遮普語 (印度) 的語言值。 |
| te-in |
泰盧固語 (印度) 的查詢語言值。 |
| ur-pk |
查詢烏爾都語 (巴基斯坦) 的語言值。 |
QueryResultDocumentRerankerInput
傳送至語意擴充程式的原始串連字串。
| 名稱 | 類型 | Description |
|---|---|---|
| content |
string |
用於語意擴充之內容欄位的原始串連字串。 |
| keywords |
string |
用於語意擴充之關鍵字欄位的原始串連字串。 |
| title |
string |
用於語意擴充的標題欄位的原始字串。 |
QueryResultDocumentSemanticField
傳送至語意擴充程式的字段描述,以及其使用方式
| 名稱 | 類型 | Description |
|---|---|---|
| name |
string |
傳送至語意擴充處理程序的欄位名稱 |
| state |
欄位用於語義豐富過程的方式(完全使用、部分使用或未使用) |
QueryResultDocumentSubscores
本檔搜尋查詢的文字和向量查詢元件之間的子核心明細。 每個向量查詢都會以收到的相同順序顯示為個別物件。
| 名稱 | 類型 | Description |
|---|---|---|
| documentBoost |
number (double) |
查詢文字部分的 BM25 或傳統分數。 |
| text |
查詢文字部分的 BM25 或傳統分數。 |
|
| vectors |
<string,
Single |
向量相似度與 |
QueryRewritesDebugInfo
包含查詢重寫的特定偵錯資訊。
| 名稱 | 類型 | Description |
|---|---|---|
| text |
針對文字查詢產生的查詢重寫清單。 |
|
| vectors |
針對可向量化文字查詢產生的查詢重寫清單。 |
QueryRewritesType
只有在查詢類型為 semantic時,此參數才有效。 當 QueryRewrites 設定為 generative時,查詢字詞會傳送至產生模型,該模型會產生 10 個 (預設) 重寫,以協助增加要求的召回率。 您可以藉由附加管道字元 | 後面接選項 count-<number of rewrites> (例如 generative|count-3) 來配置所要求的計數。 預設為 None。
| 值 | Description |
|---|---|
| none |
請勿為此查詢產生其他查詢重寫。 |
| generative |
產生替代查詢字詞,以增加搜尋要求的召回率。 |
QueryRewritesValuesDebugInfo
包含查詢重寫的特定偵錯資訊。
| 名稱 | 類型 | Description |
|---|---|---|
| inputQuery |
string |
生成式查詢重寫模型的輸入文字。 在某些情況下,使用者查詢和生成模型的輸入可能不相同。 |
| rewrites |
string[] |
查詢重寫清單。 |
QuerySpellerType
藉由拼字更正個別搜尋查詢字詞來改善搜尋召回率。
| 值 | Description |
|---|---|
| none |
拼字工具未啟用。 |
| lexicon |
拼字檢查程式會針對 queryLanguage 參數所指定的語言,使用靜態字典來更正個別查詢字詞。 |
QueryType
指定搜尋查詢的語法。 預設值為 『simple』。 如果您的查詢使用 Lucene 查詢語法,請使用 'full',如果不需要查詢語法,請使用 'semantic'。
| 值 | Description |
|---|---|
| simple |
使用簡單的查詢語法進行搜尋。 搜尋文字是使用簡單的查詢語言來解譯,該語言允許 +、* 和 “” 等符號。 依預設,除非指定 searchFields 參數,否則會在所有可搜尋欄位中評估查詢。 |
| full |
使用完整的 Lucene 查詢語法進行搜尋。 搜尋文字會使用 Lucene 查詢語言解譯,該語言允許欄位特定和加權搜尋,以及其他進階功能。 |
| semantic |
最適合以自然語言而不是關鍵字表達的查詢。 通過使用在 Web 語料庫上訓練的排名模型重新排名排名排名,提高搜索結果的準確性。 |
ScoringStatistics
值,指定我們是要全域計算評分統計數據(例如文件頻率),以取得更一致的評分,還是針對較低的延遲在本機計算。 預設值為「本機」。 在評分之前,使用 'global' 來彙總全域評分統計資料。 使用全域評分統計資料可能會增加搜尋查詢的延遲。
| 值 | Description |
|---|---|
| local |
評分統計資料將在本機計算,以降低延遲。 |
| global |
評分統計數據將全局計算,以獲得更一致的評分。 |
SearchDocumentsResult
包含索引搜尋結果的回應。
| 名稱 | 類型 | Description |
|---|---|---|
| @odata.count |
integer (int64) |
搜尋作業所找到的結果總數,如果未要求計數,則為 null。 如果存在,計數可能大於此回應中的結果數目。 如果您使用 $top 或 $skip 參數,或查詢無法在單一回應中傳回所有要求的文件,則可能會發生這種情況。 |
| @odata.nextLink |
string |
當查詢無法在單一回應中傳回所有要求的結果時,會傳回接續 URL。 您可以使用此 URL 來制定另一個 GET 或 POST 搜尋要求,以取得搜尋回應的下一部分。 請務必使用與產生此回應之要求相同的動詞 (GET 或 POST)。 |
| @search.answers |
搜尋作業的答案查詢結果;如果未指定答案查詢參數或設定為 『none』,則為 null。 |
|
| @search.coverage |
number (double) |
值,指出查詢中包含的索引百分比,如果要求中未指定 minimumCoverage,則為 null。 |
| @search.debug |
整體套用至搜尋結果的偵錯資訊。 |
|
| @search.facets |
object |
搜尋作業的 Facet 查詢結果,會組織為每個多面向字段的貯體集合;如果查詢未包含任何 Facet 運算式,則為 null。 |
| @search.nextPageParameters |
當查詢無法在單一回應中傳回所有要求的結果時,會傳回接續 JSON 承載。 你可以搭配使用 JSON |
|
| @search.semanticPartialResponseReason |
針對語意排名要求傳回部分回應的原因。 |
|
| @search.semanticPartialResponseType |
針對語意排名要求傳回的部分回應類型。 |
|
| @search.semanticQueryRewritesResultType |
用來擷取文件的查詢重寫類型。 |
|
| value |
查詢所傳回的結果序列。 |
SearchMode
指定是否必須比對任何或所有搜尋字詞,才能將文件計為相符專案。
| 值 | Description |
|---|---|
| any |
任何搜尋字詞都必須相符,才能將文件計為相符。 |
| all |
所有搜尋字詞都必須相符,才能將文件計為相符項。 |
SearchRequest
篩選、排序、Facet、分頁和其他搜尋查詢行為的參數。
| 名稱 | 類型 | Description |
|---|---|---|
| answers |
值,指定是否應該在搜尋回應中傳回答案。 |
|
| captions |
值,指定是否應該在搜尋回應中傳回標題。 |
|
| count |
boolean |
值,指定是否要擷取結果的總計數。 預設值為 false。 將此值設定為 true 可能會對效能造成影響。 請注意,傳回的計數是近似值。 |
| debug |
啟用偵錯工具,可用來進一步探索重新排名的結果。 |
|
| facets |
string[] |
要套用至搜尋查詢的Facet表達式清單。 每個 Facet 運算式都包含功能變數名稱,選擇性地後面接著以逗號分隔的名稱:值組清單。 |
| filter |
string |
要套用至搜尋查詢的 OData $filter表示式。 |
| highlight |
string |
要用於點擊醒目提示的功能變數名稱逗號分隔清單。 只有可搜尋的欄位可用於點擊醒目提示。 |
| highlightPostTag |
string |
附加至叫用醒目提示的字串標記。 必須使用 highlightPreTag 進行設定。 預設值為 </em>。 |
| highlightPreTag |
string |
在叫用醒目提示前面加上的字串標記。 必須使用 highlightPostTag 進行設定。 預設值為 em <>。 |
| hybridSearch |
設定混合式搜尋行為的查詢參數。 |
|
| minimumCoverage |
number (double) |
介於 0 到 100 之間的數位,表示搜尋查詢必須涵蓋的索引百分比,以便將查詢回報為成功。 此參數對於確保搜尋可用性,即使是只有一個複本的服務,也很有用。 預設值為 100。 |
| orderby |
string |
以逗號分隔的 OData 清單$orderby用來排序結果的運算式。 每個運算式可以是功能變數名稱或對 geo.distance() 或 search.score() 函式的呼叫。 每個表達式後面可以接著 asc 表示遞增,或 desc 表示遞減。 預設值為遞增順序。 系結會因檔比對分數而中斷。 如果未指定任何$orderby,則預設排序順序會依檔比對分數遞減。 最多可以有 32 個$orderby子句。 |
| queryLanguage |
指定搜尋查詢語言的值。 |
|
| queryRewrites |
指定是否應該產生查詢重寫以擴充搜尋查詢的值。 |
|
| queryType |
值,指定搜尋查詢的語法。 預設值為 『simple』。 如果您的查詢使用 Lucene 查詢語法,請使用 『full』。 |
|
| scoringParameters |
string[] |
要用於評分函式的參數值清單(例如 referencePointParameter),其格式為 name-values。 例如,如果評分配置檔使用名為 'mylocation' 的參數定義函式,則參數位符串會是 “mylocation--122.2,44.8” (不含引號)。 |
| scoringProfile |
string |
要評估比對檔比對分數的評分配置檔名稱,以便排序結果。 |
| scoringStatistics |
值,指定我們是要全域計算評分統計數據(例如文件頻率),以取得更一致的評分,還是針對較低的延遲在本機計算。 預設值為「本機」。 在評分之前,使用 'global' 來彙總全域評分統計資料。 使用全域評分統計資料可能會增加搜尋查詢的延遲。 |
|
| search |
string |
全文檢索搜尋查詢運算式;使用「*」或省略此參數來比對所有文件。 |
| searchFields |
string |
要設定全文搜索範圍之功能變數名稱的逗號分隔清單。 在完整的 Lucene 查詢中使用字段搜尋 (fieldName:searchExpression) 時,每個字段搜尋表達式的功能變數名稱優先於此參數中列出的任何功能變數名稱。 |
| searchMode |
值,指定是否必須比對任何或所有搜尋字詞,才能將檔計算為相符專案。 |
|
| select |
string |
要擷取的欄位以逗點分隔的清單。 如果未指定,則會包含標示為可在架構中擷取的所有欄位。 |
| semanticConfiguration |
string |
處理語意類型查詢文件時將使用的語意配置名稱。 |
| semanticErrorHandling |
允許使用者選擇語意呼叫是否應該完全失敗 (預設/目前行為) ,或傳回部分結果。 |
|
| semanticFields |
string |
用於語意排名的欄位名稱的逗點分隔清單。 |
| semanticMaxWaitInMilliseconds |
integer (int32) minimum: 700 |
容許使用者設定語意擴充在要求失敗之前完成處理所需的時間量上限。 |
| semanticQuery |
string |
允許設定單獨的搜尋查詢,該查詢將僅用於語意重新排序、語意標題和語意答案。 適用於需要在基底擷取和排名階段與 L2 語意階段之間使用不同查詢的案例。 |
| sessionId |
string |
用來建立黏性工作階段的值,有助於取得更一致的結果。 只要使用相同的 sessionId,就會嘗試以相同的副本集為目標。 請謹慎地重複重複使用相同的 sessionID 值,可能會干擾跨複本的要求負載平衡,並對搜尋服務的效能造成負面影響。 做為 sessionId 的值不能以 '_' 字元開頭。 |
| skip |
integer (int32) |
要略過的搜尋結果數目。 這個值不能大於 100,000。 如果您需要依序掃描檔,但因這項限制而無法使用skip,請考慮改為在完全排序的索引鍵上使用orderby,並改用範圍查詢篩選。 |
| speller |
一個用來指定拼字者用來修正個別搜尋詞的類型值。 |
|
| top |
integer (int32) |
要擷取的搜尋結果數目。 這可與$skip搭配使用,以實作搜尋結果的用戶端分頁。 如果結果因為伺服器端分頁而遭到截斷,回應會包含接續令牌,可用來針對下一頁的結果發出另一個搜尋要求。 |
| vectorFilterMode |
決定是否在執行向量搜尋之前或之後套用篩選器。 新索引的預設值為 'preFilter'。 |
|
| vectorQueries | VectorQuery[]: |
向量和混合式搜尋查詢的查詢參數。 |
SearchResult
包含搜尋查詢所找到的檔,加上相關聯的元數據。
| 名稱 | 類型 | Description |
|---|---|---|
| @search.captions |
標題是文件相對於搜尋查詢最具代表性的段落。 它們通常用來做為文件摘要。 只有針對「語意」類型的查詢才會傳回標題。 |
|
| @search.documentDebugInfo |
包含可用來進一步探索搜尋結果的偵錯資訊。 |
|
| @search.highlights |
object |
文件中的文字片段,指出相符的搜尋字詞,依每個適用的欄位組織;null 如果未啟用查詢的點擊醒目提示。 |
| @search.rerankerBoostedScore |
number (double) |
透過提升 Reranker 分數來計算的相關性分數。 搜尋結果依據語意設定中的 useSscoreingProfileBoostedRanking 排序,為 RerankerScore/RerankerBoostedScore。RerankerBoostedScore 僅在類型為「語意」的查詢時回傳。 |
| @search.rerankerScore |
number (double) |
語意排名器針對熱門搜尋結果計算的相關性分數。 搜尋結果會先依 RerankerScore 排序,然後依分數排序。 RerankerScore 只會針對類型為 'semantic' 的查詢傳回。 |
| @search.score |
number (double) |
文件與查詢傳回的其他文件相比的相關性分數。 |
SearchScoreThreshold
向量查詢的結果會根據 ' 進行篩選
| 名稱 | 類型 | Description |
|---|---|---|
| kind |
string:
search |
門檻的類型。 |
| value |
number (double) |
閾值會根據 |
SemanticDebugInfo
包含語意排名要求特定的偵錯資訊。
| 名稱 | 類型 | Description |
|---|---|---|
| contentFields |
傳送至語意擴充程式的內容欄位,以及它們的使用方式 |
|
| keywordFields |
傳送至語意擴充程式的關鍵字欄位,以及它們的使用方式 |
|
| rerankerInput |
傳送至語意擴充程式的原始串連字串。 |
|
| titleField |
傳送至語意擴充程式的標題欄位,以及它的使用方式 |
SemanticErrorMode
允許使用者選擇語意呼叫是否完全失敗,或回傳部分結果。
| 值 | Description |
|---|---|
| partial |
如果語意處理失敗,部分結果仍會傳回。 部分結果的定義取決於失敗的語義步驟以及失敗的原因。 |
| fail |
如果在語意處理步驟中出現異常,查詢將失敗,並根據錯誤傳回適當的HTTP程式碼。 |
SemanticErrorReason
針對語意排名要求傳回部分回應的原因。
| 值 | Description |
|---|---|
| maxWaitExceeded |
如果已設定,且 |
| capacityOverloaded |
請求受到節流。 只會傳回基本結果。 |
| transient |
語意過程的至少一個步驟失敗了。 |
SemanticFieldState
該場在語意豐富過程中的使用方式。
| 值 | Description |
|---|---|
| used |
該字段被充分利用用於語義豐富。 |
| unused |
該字段未用於語義豐富。 |
| partial |
該字段部分用於語義豐富。 |
SemanticQueryRewritesResultType
用於此要求的查詢重寫類型。
| 值 | Description |
|---|---|
| originalQueryOnly |
未成功針對此要求產生查詢重寫。 只有原始查詢可用來擷取結果。 |
SemanticSearchResultsType
針對語意排名要求傳回的部分回應類型。
| 值 | Description |
|---|---|
| baseResults |
沒有任何語義豐富或重新排名的結果。 |
| rerankedResults |
結果已使用重新排名模型重新排名,並將包含語義標題。 它們不會包含任何答案、答案突出顯示或標題突出顯示。 |
SingleVectorFieldResult
單一向量欄位結果。 Both
| 名稱 | 類型 | Description |
|---|---|---|
| searchScore |
number (double) |
此 |
| vectorSimilarity |
number (double) |
本文件的向量相似性分數。 請注意,這是相似度計量的正式定義,而不是「距離」版本。 例如,餘弦相似度而不是餘弦距離。 |
TextResult
查詢文字部分的 BM25 或傳統分數。
| 名稱 | 類型 | Description |
|---|---|---|
| searchScore |
number (double) |
查詢文字部分的 BM25 或傳統分數。 |
VectorFilterMode
決定是否在執行向量搜尋之前或之後套用篩選器。
| 值 | Description |
|---|---|
| postFilter |
篩選會在傳回候選的向量結果集之後套用。 視篩選選擇性而定,這可能會導致比參數 『k』 所要求的結果少。 |
| preFilter |
篩選會在搜尋查詢之前套用。 |
| strictPostFilter |
在傳回向量結果的全域前 k 個候選集之後,將套用篩選器。 這會導致比參數 'k' 所要求的結果少。 |
VectorizableImageBinaryQuery
提供需要向量化之影像的基底 64 編碼二進位檔時,要用於向量搜尋的查詢參數。
| 名稱 | 類型 | Description |
|---|---|---|
| base64Image |
string |
要向量化的圖像的基數 64 編碼二進制文件以執行向量搜索查詢。 |
| exhaustive |
boolean |
若為 true,則會在向量索引內的所有向量觸發詳盡的 K 近鄰搜尋。 適用於完全相符專案很重要的案例,例如判斷地面真相值。 |
| fields |
string |
要包含在搜尋向量中之 Collection(Edm.Single) 類型的向量字段。 |
| filterOverride |
string |
要套用至這個特定向量查詢的 OData 篩選表示式。 如果未在向量層級定義任何篩選表達式,則會改用最上層篩選參數中定義的表達式。 |
| k |
integer (int32) |
要作為熱門點擊傳回的近鄰數目。 |
| kind |
string:
image |
查詢類型。 |
| oversampling |
number (double) |
過度取樣因素。 最小值為 1。 它會覆寫索引定義中設定的 『defaultOversampling』 參數。 只有當 'rerankWithOriginalVectors' 為 true 時,才能進行設定。 只有在基礎向量欄位上使用壓縮方法時,才允許此參數。 |
| perDocumentVectorLimit |
integer (int32) |
控制向量搜尋查詢中每個文件可以比對的向量數量。 將其設定為 1 可確保每個文件最多匹配一個向量,從而確保結果來自不同的文件。 將其設定為 0(無限制)允許比對同一文件中的多個相關向量。 預設值為 0。 |
| threshold | VectorThreshold: |
用於向量查詢的臨界值。 請注意,只有在所有 'fields' 都使用相同的相似度計量時,才能設定這個值。 |
| weight |
number (float) |
相較於相同搜尋要求中的其他向量查詢和/或文字查詢,向量查詢的相對加權。 結合不同向量查詢所產生的多個排名清單結果,以及/或透過文字查詢擷取的結果時,會使用此值。 權數越高,符合該查詢的檔就會在最終排名中。 默認值為 1.0,值必須是大於零的正數。 |
VectorizableImageUrlQuery
當提供代表需要向量化之影像值的 URL 時,要用於向量搜尋的查詢參數。
| 名稱 | 類型 | Description |
|---|---|---|
| exhaustive |
boolean |
若為 true,則會在向量索引內的所有向量觸發詳盡的 K 近鄰搜尋。 適用於完全相符專案很重要的案例,例如判斷地面真相值。 |
| fields |
string |
要包含在搜尋向量中之 Collection(Edm.Single) 類型的向量字段。 |
| filterOverride |
string |
要套用至這個特定向量查詢的 OData 篩選表示式。 如果未在向量層級定義任何篩選表達式,則會改用最上層篩選參數中定義的表達式。 |
| k |
integer (int32) |
要作為熱門點擊傳回的近鄰數目。 |
| kind |
string:
image |
查詢類型。 |
| oversampling |
number (double) |
過度取樣因素。 最小值為 1。 它會覆寫索引定義中設定的 『defaultOversampling』 參數。 只有當 'rerankWithOriginalVectors' 為 true 時,才能進行設定。 只有在基礎向量欄位上使用壓縮方法時,才允許此參數。 |
| perDocumentVectorLimit |
integer (int32) |
控制向量搜尋查詢中每個文件可以比對的向量數量。 將其設定為 1 可確保每個文件最多匹配一個向量,從而確保結果來自不同的文件。 將其設定為 0(無限制)允許比對同一文件中的多個相關向量。 預設值為 0。 |
| threshold | VectorThreshold: |
用於向量查詢的臨界值。 請注意,只有在所有 'fields' 都使用相同的相似度計量時,才能設定這個值。 |
| url |
string |
要向量化以執行向量搜尋查詢之影像的URL。 |
| weight |
number (float) |
相較於相同搜尋要求中的其他向量查詢和/或文字查詢,向量查詢的相對加權。 結合不同向量查詢所產生的多個排名清單結果,以及/或透過文字查詢擷取的結果時,會使用此值。 權數越高,符合該查詢的檔就會在最終排名中。 默認值為 1.0,值必須是大於零的正數。 |
VectorizableTextQuery
提供需要向量化的文字值時,要用於向量搜尋的查詢參數。
| 名稱 | 類型 | Description |
|---|---|---|
| exhaustive |
boolean |
若為 true,則會在向量索引內的所有向量觸發詳盡的 K 近鄰搜尋。 適用於完全相符專案很重要的案例,例如判斷地面真相值。 |
| fields |
string |
要包含在搜尋向量中之 Collection(Edm.Single) 類型的向量字段。 |
| filterOverride |
string |
要套用至這個特定向量查詢的 OData 篩選表示式。 如果未在向量層級定義任何篩選表達式,則會改用最上層篩選參數中定義的表達式。 |
| k |
integer (int32) |
要作為熱門點擊傳回的近鄰數目。 |
| kind |
string:
text |
查詢類型。 |
| oversampling |
number (double) |
過度取樣因素。 最小值為 1。 它會覆寫索引定義中設定的 『defaultOversampling』 參數。 只有當 'rerankWithOriginalVectors' 為 true 時,才能進行設定。 只有在基礎向量欄位上使用壓縮方法時,才允許此參數。 |
| perDocumentVectorLimit |
integer (int32) |
控制向量搜尋查詢中每個文件可以比對的向量數量。 將其設定為 1 可確保每個文件最多匹配一個向量,從而確保結果來自不同的文件。 將其設定為 0(無限制)允許比對同一文件中的多個相關向量。 預設值為 0。 |
| queryRewrites |
您可以設定為讓產生模型重新撰寫查詢,再將它傳送至向量化。 |
|
| text |
string |
要向量化以執行向量搜尋查詢的文字。 |
| threshold | VectorThreshold: |
用於向量查詢的臨界值。 請注意,只有在所有 'fields' 都使用相同的相似度計量時,才能設定這個值。 |
| weight |
number (float) |
相較於相同搜尋要求中的其他向量查詢和/或文字查詢,向量查詢的相對加權。 結合不同向量查詢所產生的多個排名清單結果,以及/或透過文字查詢擷取的結果時,會使用此值。 權數越高,符合該查詢的檔就會在最終排名中。 默認值為 1.0,值必須是大於零的正數。 |
VectorizedQuery
提供原始向量值時,用於向量搜尋的查詢參數。
| 名稱 | 類型 | Description |
|---|---|---|
| exhaustive |
boolean |
若為 true,則會在向量索引內的所有向量觸發詳盡的 K 近鄰搜尋。 適用於完全相符專案很重要的案例,例如判斷地面真相值。 |
| fields |
string |
要包含在搜尋向量中之 Collection(Edm.Single) 類型的向量字段。 |
| filterOverride |
string |
要套用至這個特定向量查詢的 OData 篩選表示式。 如果未在向量層級定義任何篩選表達式,則會改用最上層篩選參數中定義的表達式。 |
| k |
integer (int32) |
要作為熱門點擊傳回的近鄰數目。 |
| kind |
string:
vector |
查詢類型。 |
| oversampling |
number (double) |
過度取樣因素。 最小值為 1。 它會覆寫索引定義中設定的 『defaultOversampling』 參數。 只有當 'rerankWithOriginalVectors' 為 true 時,才能進行設定。 只有在基礎向量欄位上使用壓縮方法時,才允許此參數。 |
| perDocumentVectorLimit |
integer (int32) |
控制向量搜尋查詢中每個文件可以比對的向量數量。 將其設定為 1 可確保每個文件最多匹配一個向量,從而確保結果來自不同的文件。 將其設定為 0(無限制)允許比對同一文件中的多個相關向量。 預設值為 0。 |
| threshold | VectorThreshold: |
用於向量查詢的臨界值。 請注意,只有在所有 'fields' 都使用相同的相似度計量時,才能設定這個值。 |
| vector |
number[] (float) |
搜尋查詢的向量表示。 |
| weight |
number (float) |
相較於相同搜尋要求中的其他向量查詢和/或文字查詢,向量查詢的相對加權。 結合不同向量查詢所產生的多個排名清單結果,以及/或透過文字查詢擷取的結果時,會使用此值。 權數越高,符合該查詢的檔就會在最終排名中。 默認值為 1.0,值必須是大於零的正數。 |
VectorQueryKind
正在執行的向量查詢類型。
| 值 | Description |
|---|---|
| vector |
向量查詢,其中提供原始向量值。 |
| text |
向量查詢,其中提供了需要向量化的文字值。 |
| imageUrl |
向量查詢,其中提供代表需要向量化的影像值的 URL。 |
| imageBinary |
向量查詢,其中提供了需要向量化的圖像的基數 64 編碼二進位檔。 |
VectorsDebugInfo
「包含針對向量與混合搜尋的除錯資訊。」)
| 名稱 | 類型 | Description |
|---|---|---|
| subscores |
在選擇結果集融合/組合方法(例如 RRF)之前文件的子分數細分。 |
VectorSimilarityThreshold
系統會根據向量相似度計量來篩選向量查詢的結果。 請注意,這是相似度計量的正式定義,而不是「距離」版本。 系統會根據欄位所使用的計量自動選擇臨界值方向(較大或更小)。
| 名稱 | 類型 | Description |
|---|---|---|
| kind |
string:
vector |
門檻的類型。 |
| value |
number (double) |
臨界值會根據相似性量度值進行篩選。 請注意,這是相似度計量的正式定義,而不是「距離」版本。 系統會根據欄位所使用的計量自動選擇臨界值方向(較大或更小)。 |
VectorThresholdKind
那種用來篩選向量查詢的門檻。
| 值 | Description |
|---|---|
| vectorSimilarity |
系統會根據向量相似度計量來篩選向量查詢的結果。 請注意,這是相似度計量的正式定義,而不是「距離」版本。 系統會根據欄位所使用的計量自動選擇臨界值方向(較大或更小)。 |
| searchScore |
向量查詢的結果會根據 『@search.score』 值進行篩選。 請注意,這是在搜尋回應中傳回 @search.score。 系統會針對較高的 @search.score選擇臨界值方向。 |