Join us Sept 17 at .local NYC! Use code WEB50 to save 50% on tickets. Learn more >
MongoDB Event
Docs Menu
Docs Home
/ / /
C#/.NET ドライバー

Atlas ベクトル検索クエリの実行

Atlas Vector Search を使用して、Atlas に保存されているデータに対してベクトル検索を実行できます。ベクトル検索を使用すると、キーワードの一致だけでなく、意味に基づいてデータをクエリできるため、より関連性の高い検索結果を検索できます。AI を搭載したアプリケーションは、セマンティック検索、ハイブリッド検索、生成検索、検索拡張生成(RAG)などのユースケースをサポートできます。

Atlas をベクトルデータベースとして使用することで、ベクトルデータを Atlas 内の他のデータと合わせてシームレスにインデックス化できます。これにより、コレクション内のフィールドをフィルターし、ベクトルデータに対してベクター検索クエリを実行することができます。また、ベクトル検索と全文検索クエリを組み合わせて、ユースケースに最も関連性の高い結果を返すことも可能です。Atlas Vector Search を一般的な AI フレームワークおよびサービスと統合することで、アプリケーションにベクトル検索を簡単に実装できます。

Atlas ベクトル検索の詳細については、 MongoDB Atlasドキュメントの Atlas ベクトル検索ガイドを参照してください。

このガイドの例では、sample_mflix データベースの sample_mflix.embedded_movies コレクションを使用します。このコレクションのサンプルデータセットを取得するには、.NET/C# ドライバーの始め方」を参照してください。

このガイドの例では、次のサンプルクラスを使用して、sample_mflix.embedded_movies コレクション内のドキュメントをモデル化します。

[BsonIgnoreExtraElements]
public class EmbeddedMovie
{
public ObjectId Id { get; set; }
public string Plot { get; set; }
public string Title { get; set; }
[BsonElement("plot_embedding")]
public float[] PlotEmbedding { get; set; }
}

ベクトル埋め込みは、データを表現するために使用するベクトルです。これらの埋め込みは、データ内の意味のある関係をキャプチャし、セマンティック検索や取得などのタスクを可能にします。

.NET/C# ドライバーは、さまざまなタイプのベクトル埋め込みをサポートしています。次のセクションでは、サポートされているベクトル埋め込みタイプについて説明いたします。

.NET/C# ドライバーは、ベクトル埋め込みにおける配列型の以下の表現をサポートしています。

  • BsonArray

  • Memory

  • ReadOnlyMemory

  • float[] および double[]

次の例では、前述の型のプロパティを持つクラスを示します。

public class BsonArrayVectors
{
public BsonArray BsonArrayVector { get; set; }
public Memory<float> MemoryVector { get; set; }
public ReadOnlyMemory<float> ReadOnlyMemoryVector { get; set; }
public float[] FloatArrayVector { get; set; }
}

Tip

Memory および ReadOnlyMemory 型の使用方法の詳細については、直列化ガイドの「配列の直列化パフォーマンスの向上」セクションを参照してください。

.NET/C# ドライバーは、ベクトル埋め込みで次のバイナリベクトル表現をサポートします。

  • BinaryVectorFloat32 (ビッグ エンディアン アーキテクチャではサポートされていません)

  • BinaryVectorInt8

  • BinaryVectorPackedBit

  • Memory<float>, Memory<byte>, Memory<sbyte>

  • ReadOnlyMemory<float>, ReadOnlyMemory<byte>, ReadOnlyMemory<sbyte>

  • float[], byte[], sbyte[]

注意

Memory<T>ReadOnlyMemory<T>、または配列型のバイナリベクトル表現を指定する際には、BinaryVector 属性を使用しなければなりません。

次の例では、前述の型のプロパティを持つクラスを示します。

public class BinaryVectors
{
public BinaryVectorInt8 ValuesInt8 { get; set; }
public BinaryVectorPackedBit ValuesPackedBit { get; set; }
public BinaryVectorFloat32 ValuesFloat { get; set; }
[BinaryVector(BinaryVectorDataType.Int8)]
public Memory<byte> ValuesByte { get; set; }
[BinaryVector(BinaryVectorDataType.Float32)]
public float[] ValuesFloat { get; set; }
}

Int8 バイナリ ベクトル型データをbyte または sbyte としてシリアル化できます。Float32 バイナリ ベクトル型データを float としてシリアル化することも可能です。次の例では、Int8 および Float32 のバイナリ ベクトル データをシリアル化します。

[BinaryVector(BinaryVectorDataType.Int8)]
public Memory<byte> ValuesByte { get; set; }
[BinaryVector(BinaryVectorDataType.Int8)]
public Memory<sbyte> ValuesSByte { get; set; }
[BinaryVector(BinaryVectorDataType.Float32)]
public float[] ValuesFloat { get; set; }

PackedBitベクトルデータを、 byteデータ型で表されるバイナリベクトルに逆直列化できるのは、ベクトルデータのパディング値が0 の場合のみです。ベクトルデータのパディング値が 0 と等しくない場合は、BsonVectorPackedBit にのみ逆直列化できます。

VectorSearch() メソッドを呼び出すことでベクトル検索クエリを実行することができます。コレクションでベクトル検索を実行するには、まず、ベクトルデータを含むフィールドと、そのフィールドをカバーするベクトル検索インデックスを持つコレクションが必要です。

ベクトル検索用のコレクション構成について詳しくは、MongoDB Atlas ドキュメントの Atlas Vector Search ガイドをご覧ください。

BinaryVectorFloat32BinaryVectorInt8、および BinaryVectorPackedBit のデータをBsonBinaryData 型に変換し、ベクトル検索クエリで使用するには、ToQueryVector() メソッドを使用します。次の例では、BinaryVectorInt8BsonBinaryData オブジェクトに変換します:

var binaryVector = new BinaryVectorInt8(new sbyte[] { 0, 1, 2, 3, 4 });
var queryVector = binaryVector.ToQueryVector();

配列で表現されたベクトルデータを QueryVector クラスのインスタンスとして指定し、ベクトル検索クエリで使用することができます。次の例では、ReadOnlyMemory<float> 値の配列を QueryVector オブジェクトとして作成し、ベクトル検索クエリで使用します。

QueryVector v = new QueryVector(new ReadOnlyMemory<float>([1.2f, 2.3f]));

sample_mflix データベースの embedded_movies コレクションを検討します。$vectorSearch ステージを使用して、コレクション内のドキュメントの plot_embedding フィールドでセマンティック検索を実行できます。次のセクションでは、このコレクションに対して Atlas Vector Search 操作を実行するためのさまざまな方法について説明します。

Tip

次の例で使用されるサンプルデータセットを取得するには、「.NET/C# ドライバーの始め方」を参照してください。次の例で使用されるサンプル Atlas Vector Search インデックスを作成するには、Atlas マニュアルの「Atlas Vector Search インデックスの作成」を参照してください。

この例では、PlotEmbedding フィールドにベクトルデータと Atlas Vector Search インデックスを含むコレクションに対して、Atlas Vector Search クエリを実行するために次の手順を実行します。

  1. 検索用の配列形式のベクトルデータを含む配列を作成します。

  2. インデックス名や最近傍数を含むVectorSearchOptionsオブジェクトを指定します。

  3. VectorSearch() ステージを使用してベクトル検索クエリを実行し、Project() ステージを使用して TitlePlot、および Score フィールドのみを表示する集計パイプラインを作成します

  4. クエリの結果を出力する

// Defines vector embeddings for the string "time travel"
var vector = new[] {-0.0016261312,-0.028070757,-0.011342932,-0.012775794,-0.0027440966,0.008683807,-0.02575152,-0.02020668,-0.010283281,-0.0041719596,0.021392956,0.028657231,-0.006634482,0.007490867,0.018593878,0.0038187427,0.029590257,-0.01451522,0.016061379,0.00008528442,-0.008943722,0.01627464,0.024311995,-0.025911469,0.00022596726,-0.008863748,0.008823762,-0.034921836,0.007910728,-0.01515501,0.035801545,-0.0035688248,-0.020299982,-0.03145631,-0.032256044,-0.028763862,-0.0071576433,-0.012769129,0.012322609,-0.006621153,0.010583182,0.024085402,-0.001623632,0.007864078,-0.021406285,0.002554159,0.012229307,-0.011762793,0.0051682983,0.0048484034,0.018087378,0.024325324,-0.037694257,-0.026537929,-0.008803768,-0.017767483,-0.012642504,-0.0062712682,0.0009771782,-0.010409906,0.017754154,-0.004671795,-0.030469967,0.008477209,-0.005218282,-0.0058480743,-0.020153364,-0.0032805866,0.004248601,0.0051449724,0.006791097,0.007650814,0.003458861,-0.0031223053,-0.01932697,-0.033615597,0.00745088,0.006321252,-0.0038154104,0.014555207,0.027697546,-0.02828402,0.0066711367,0.0077107945,0.01794076,0.011349596,-0.0052715978,0.014755142,-0.019753495,-0.011156326,0.011202978,0.022126047,0.00846388,0.030549942,-0.0041386373,0.018847128,-0.00033655585,0.024925126,-0.003555496,-0.019300312,0.010749794,0.0075308536,-0.018287312,-0.016567878,-0.012869096,-0.015528221,0.0078107617,-0.011156326,0.013522214,-0.020646535,-0.01211601,0.055928253,0.011596181,-0.017247654,0.0005939711,-0.026977783,-0.003942035,-0.009583511,-0.0055248477,-0.028737204,0.023179034,0.003995351,0.0219661,-0.008470545,0.023392297,0.010469886,-0.015874773,0.007890735,-0.009690142,-0.00024970944,0.012775794,0.0114762215,0.013422247,0.010429899,-0.03686786,-0.006717788,-0.027484283,0.011556195,-0.036068123,-0.013915418,-0.0016327957,0.0151016945,-0.020473259,0.004671795,-0.012555866,0.0209531,0.01982014,0.024485271,0.0105431955,-0.005178295,0.033162415,-0.013795458,0.007150979,0.010243294,0.005644808,0.017260984,-0.0045618312,0.0024725192,0.004305249,-0.008197301,0.0014203656,0.0018460588,0.005015015,-0.011142998,0.01439526,0.022965772,0.02552493,0.007757446,-0.0019726837,0.009503538,-0.032042783,0.008403899,-0.04609149,0.013808787,0.011749465,0.036388017,0.016314628,0.021939443,-0.0250051,-0.017354285,-0.012962398,0.00006107364,0.019113706,0.03081652,-0.018114036,-0.0084572155,0.009643491,-0.0034721901,0.0072642746,-0.0090636825,0.01642126,0.013428912,0.027724205,0.0071243206,-0.6858542,-0.031029783,-0.014595194,-0.011449563,0.017514233,0.01743426,0.009950057,0.0029706885,-0.015714826,-0.001806072,0.011856096,0.026444625,-0.0010663156,-0.006474535,0.0016161345,-0.020313311,0.0148351155,-0.0018393943,0.0057347785,0.018300641,-0.018647194,0.03345565,-0.008070676,0.0071443142,0.014301958,0.0044818576,0.003838736,-0.007350913,-0.024525259,-0.001142124,-0.018620536,0.017247654,0.007037683,0.010236629,0.06046009,0.0138887605,-0.012122675,0.037694257,0.0055081863,0.042492677,0.00021784494,-0.011656162,0.010276617,0.022325981,0.005984696,-0.009496873,0.013382261,-0.0010563189,0.0026507939,-0.041639622,0.008637156,0.026471283,-0.008403899,0.024858482,-0.00066686375,-0.0016252982,0.027590916,0.0051449724,0.0058647357,-0.008743787,-0.014968405,0.027724205,-0.011596181,0.0047650975,-0.015381602,0.0043718936,0.002159289,0.035908177,-0.008243952,-0.030443309,0.027564257,0.042625964,-0.0033688906,0.01843393,0.019087048,0.024578573,0.03268257,-0.015608194,-0.014128681,-0.0033538956,-0.0028757197,-0.004121976,-0.032389335,0.0034322033,0.058807302,0.010943064,-0.030523283,0.008903735,0.017500903,0.00871713,-0.0029406983,0.013995391,-0.03132302,-0.019660193,-0.00770413,-0.0038853872,0.0015894766,-0.0015294964,-0.006251275,-0.021099718,-0.010256623,-0.008863748,0.028550599,0.02020668,-0.0012962399,-0.003415542,-0.0022509254,0.0119360695,0.027590916,-0.046971202,-0.0015194997,-0.022405956,0.0016677842,-0.00018535563,-0.015421589,-0.031802863,0.03814744,0.0065411795,0.016567878,-0.015621523,0.022899127,-0.011076353,0.02841731,-0.002679118,-0.002342562,0.015341615,0.01804739,-0.020566562,-0.012989056,-0.002990682,0.01643459,0.00042527664,0.008243952,-0.013715484,-0.004835075,-0.009803439,0.03129636,-0.021432944,0.0012087687,-0.015741484,-0.0052016205,0.00080890034,-0.01755422,0.004811749,-0.017967418,-0.026684547,-0.014128681,0.0041386373,-0.013742141,-0.010056688,-0.013268964,-0.0110630235,-0.028337335,0.015981404,-0.00997005,-0.02424535,-0.013968734,-0.028310679,-0.027750863,-0.020699851,0.02235264,0.001057985,0.00081639783,-0.0099367285,0.013522214,-0.012016043,-0.00086471526,0.013568865,0.0019376953,-0.019020405,0.017460918,-0.023045745,0.008503866,0.0064678704,-0.011509543,0.018727167,-0.003372223,-0.0028690554,-0.0027024434,-0.011902748,-0.012182655,-0.015714826,-0.0098634185,0.00593138,0.018753825,0.0010146659,0.013029044,0.0003521757,-0.017620865,0.04102649,0.00552818,0.024485271,-0.009630162,-0.015608194,0.0006718621,-0.0008418062,0.012395918,0.0057980907,0.016221326,0.010616505,0.004838407,-0.012402583,0.019900113,-0.0034521967,0.000247002,-0.03153628,0.0011038032,-0.020819811,0.016234655,-0.00330058,-0.0032289368,0.00078973995,-0.021952773,-0.022459272,0.03118973,0.03673457,-0.021472929,0.0072109587,-0.015075036,0.004855068,-0.0008151483,0.0069643734,0.010023367,-0.010276617,-0.023019087,0.0068244194,-0.0012520878,-0.0015086699,0.022046074,-0.034148756,-0.0022192693,0.002427534,-0.0027124402,0.0060346797,0.015461575,0.0137554705,0.009230294,-0.009583511,0.032629255,0.015994733,-0.019167023,-0.009203636,0.03393549,-0.017274313,-0.012042701,-0.0009930064,0.026777849,-0.013582194,-0.0027590916,-0.017594207,-0.026804507,-0.0014236979,-0.022032745,0.0091236625,-0.0042419364,-0.00858384,-0.0033905501,-0.020739838,0.016821127,0.022539245,0.015381602,0.015141681,0.028817179,-0.019726837,-0.0051283115,-0.011489551,-0.013208984,-0.0047017853,-0.0072309524,0.01767418,0.0025658219,-0.010323267,0.012609182,-0.028097415,0.026871152,-0.010276617,0.021912785,0.0022542577,0.005124979,-0.0019710176,0.004518512,-0.040360045,0.010969722,-0.0031539614,-0.020366628,-0.025778178,-0.0110030435,-0.016221326,0.0036587953,0.016207997,0.003007343,-0.0032555948,0.0044052163,-0.022046074,-0.0008822095,-0.009363583,0.028230704,-0.024538586,0.0029840174,0.0016044717,-0.014181997,0.031349678,-0.014381931,-0.027750863,0.02613806,0.0004136138,-0.005748107,-0.01868718,-0.0010138329,0.0054348772,0.010703143,-0.003682121,0.0030856507,-0.004275259,-0.010403241,0.021113047,-0.022685863,-0.023032416,0.031429652,0.001792743,-0.005644808,-0.011842767,-0.04078657,-0.0026874484,0.06915057,-0.00056939584,-0.013995391,0.010703143,-0.013728813,-0.022939114,-0.015261642,-0.022485929,0.016807798,0.007964044,0.0144219175,0.016821127,0.0076241563,0.005461535,-0.013248971,0.015301628,0.0085171955,-0.004318578,0.011136333,-0.0059047225,-0.010249958,-0.018207338,0.024645219,0.021752838,0.0007614159,-0.013648839,0.01111634,-0.010503208,-0.0038487327,-0.008203966,-0.00397869,0.0029740208,0.008530525,0.005261601,0.01642126,-0.0038753906,-0.013222313,0.026537929,0.024671877,-0.043505676,0.014195326,0.024778508,0.0056914594,-0.025951454,0.017620865,-0.0021359634,0.008643821,0.021299653,0.0041686273,-0.009017031,0.04044002,0.024378639,-0.027777521,-0.014208655,0.0028623908,0.042119466,0.005801423,-0.028124074,-0.03129636,0.022139376,-0.022179363,-0.04067994,0.013688826,0.013328944,0.0046184794,-0.02828402,-0.0063412455,-0.0046184794,-0.011756129,-0.010383247,-0.0018543894,-0.0018593877,-0.00052024535,0.004815081,0.014781799,0.018007403,0.01306903,-0.020433271,0.009043689,0.033189073,-0.006844413,-0.019766824,-0.018767154,0.00533491,-0.0024575242,0.018727167,0.0058080875,-0.013835444,0.0040719924,0.004881726,0.012029372,0.005664801,0.03193615,0.0058047553,0.002695779,0.009290274,0.02361889,0.017834127,0.0049017193,-0.0036388019,0.010776452,-0.019793482,0.0067777685,-0.014208655,-0.024911797,0.002385881,0.0034988478,0.020899786,-0.0025858153,-0.011849431,0.033189073,-0.021312982,0.024965113,-0.014635181,0.014048708,-0.0035921505,-0.003347231,0.030869836,-0.0017161017,-0.0061346465,0.009203636,-0.025165047,0.0068510775,0.021499587,0.013782129,-0.0024475274,-0.0051149824,-0.024445284,0.006167969,0.0068844,-0.00076183246,0.030150073,-0.0055948244,-0.011162991,-0.02057989,-0.009703471,-0.020646535,0.008004031,0.0066378145,-0.019900113,-0.012169327,-0.01439526,0.0044252095,-0.004018677,0.014621852,-0.025085073,-0.013715484,-0.017980747,0.0071043274,0.011456228,-0.01010334,-0.0035321703,-0.03801415,-0.012036037,-0.0028990454,-0.05419549,-0.024058744,-0.024272008,0.015221654,0.027964126,0.03182952,-0.015354944,0.004855068,0.011522872,0.004771762,0.0027874154,0.023405626,0.0004242353,-0.03132302,0.007057676,0.008763781,-0.0027057757,0.023005757,-0.0071176565,-0.005238275,0.029110415,-0.010989714,0.013728813,-0.009630162,-0.029137073,-0.0049317093,-0.0008630492,-0.015248313,0.0043219104,-0.0055681667,-0.013175662,0.029723546,0.025098402,0.012849103,-0.0009996708,0.03118973,-0.0021709518,0.0260181,-0.020526575,0.028097415,-0.016141351,0.010509873,-0.022965772,0.002865723,0.0020493253,0.0020509914,-0.0041419696,-0.00039695262,0.017287642,0.0038987163,0.014795128,-0.014661839,-0.008950386,0.004431874,-0.009383577,0.0012604183,-0.023019087,0.0029273694,-0.033135757,0.009176978,-0.011023037,-0.002102641,0.02663123,-0.03849399,-0.0044152127,0.0004527676,-0.0026924468,0.02828402,0.017727496,0.035135098,0.02728435,-0.005348239,-0.001467017,-0.019766824,0.014715155,0.011982721,0.0045651635,0.023458943,-0.0010046692,-0.0031373003,-0.0006972704,0.0019043729,-0.018967088,-0.024311995,0.0011546199,0.007977373,-0.004755101,-0.010016702,-0.02780418,-0.004688456,0.013022379,-0.005484861,0.0017227661,-0.015394931,-0.028763862,-0.026684547,0.0030589928,-0.018513903,0.028363993,0.0044818576,-0.009270281,0.038920518,-0.016008062,0.0093902415,0.004815081,-0.021059733,0.01451522,-0.0051583014,0.023765508,-0.017874114,-0.016821127,-0.012522544,-0.0028390652,0.0040886537,0.020259995,-0.031216389,-0.014115352,-0.009176978,0.010303274,0.020313311,0.0064112223,-0.02235264,-0.022872468,0.0052449396,0.0005723116,0.0037321046,0.016807798,-0.018527232,-0.009303603,0.0024858483,-0.0012662497,-0.007110992,0.011976057,-0.007790768,-0.042999174,-0.006727785,-0.011829439,0.007024354,0.005278262,-0.017740825,-0.0041519664,0.0085905045,0.027750863,-0.038387362,0.024391968,0.00087721116,0.010509873,-0.00038508154,-0.006857742,0.0183273,-0.0037054466,0.015461575,0.0017394272,-0.0017944091,0.014181997,-0.0052682655,0.009023695,0.00719763,-0.013522214,0.0034422,0.014941746,-0.0016711164,-0.025298337,-0.017634194,0.0058714002,-0.005321581,0.017834127,0.0110630235,-0.03369557,0.029190388,-0.008943722,0.009363583,-0.0034222065,-0.026111402,-0.007037683,-0.006561173,0.02473852,-0.007084334,-0.010110005,-0.008577175,0.0030439978,-0.022712521,0.0054582027,-0.0012620845,-0.0011954397,-0.015741484,0.0129557345,-0.00042111133,0.00846388,0.008930393,0.016487904,0.010469886,-0.007917393,-0.011762793,-0.0214596,0.000917198,0.021672864,0.010269952,-0.007737452,-0.010243294,-0.0067244526,-0.015488233,-0.021552904,0.017127695,0.011109675,0.038067464,0.00871713,-0.0025591573,0.021312982,-0.006237946,0.034628596,-0.0045251767,0.008357248,0.020686522,0.0010696478,0.0076708077,0.03772091,-0.018700508,-0.0020676525,-0.008923728,-0.023298996,0.018233996,-0.010256623,0.0017860786,0.009796774,-0.00897038,-0.01269582,-0.018527232,0.009190307,-0.02372552,-0.042119466,0.008097334,-0.0066778013,-0.021046404,0.0019593548,0.011083017,-0.0016028056,0.012662497,-0.000059095124,0.0071043274,-0.014675168,0.024831824,-0.053582355,0.038387362,0.0005698124,0.015954746,0.021552904,0.031589597,-0.009230294,-0.0006147976,0.002625802,-0.011749465,-0.034362018,-0.0067844326,-0.018793812,0.011442899,-0.008743787,0.017474247,-0.021619547,0.01831397,-0.009037024,-0.0057247817,-0.02728435,0.010363255,0.034415334,-0.024032086,-0.0020126705,-0.0045518344,-0.019353628,-0.018340627,-0.03129636,-0.0034038792,-0.006321252,-0.0016161345,0.033642255,-0.000056075285,-0.005005019,0.004571828,-0.0024075406,-0.00010215386,0.0098634185,0.1980148,-0.003825407,-0.025191706,0.035161756,0.005358236,0.025111731,0.023485601,0.0023342315,-0.011882754,0.018287312,-0.0068910643,0.003912045,0.009243623,-0.001355387,-0.028603915,-0.012802451,-0.030150073,-0.014795128,-0.028630573,-0.0013487226,0.002667455,0.00985009,-0.0033972147,-0.021486258,0.009503538,-0.017847456,0.013062365,-0.014341944,0.005078328,0.025165047,-0.015594865,-0.025924796,-0.0018177348,0.010996379,-0.02993681,0.007324255,0.014475234,-0.028577257,0.005494857,0.00011725306,-0.013315615,0.015941417,0.009376912,0.0025158382,0.008743787,0.023832154,-0.008084005,-0.014195326,-0.008823762,0.0033455652,-0.032362677,-0.021552904,-0.0056081535,0.023298996,-0.025444955,0.0097301295,0.009736794,0.015274971,-0.0012937407,-0.018087378,-0.0039387033,0.008637156,-0.011189649,-0.00023846315,-0.011582852,0.0066411467,-0.018220667,0.0060846633,0.0376676,-0.002709108,0.0072776037,0.0034188742,-0.010249958,-0.0007747449,-0.00795738,-0.022192692,0.03910712,0.032122757,0.023898797,0.0076241563,-0.007397564,-0.003655463,0.011442899,-0.014115352,-0.00505167,-0.031163072,0.030336678,-0.006857742,-0.022259338,0.004048667,0.02072651,0.0030156737,-0.0042119464,0.00041861215,-0.005731446,0.011103011,0.013822115,0.021512916,0.009216965,-0.006537847,-0.027057758,-0.04054665,0.010403241,-0.0056281467,-0.005701456,-0.002709108,-0.00745088,-0.0024841821,0.009356919,-0.022659205,0.004061996,-0.013175662,0.017074378,-0.006141311,-0.014541878,0.02993681,-0.00028448965,-0.025271678,0.011689484,-0.014528549,0.004398552,-0.017274313,0.0045751603,0.012455898,0.004121976,-0.025458284,-0.006744446,0.011822774,-0.015035049,-0.03257594,0.014675168,-0.0039187097,0.019726837,-0.0047251107,0.0022825818,0.011829439,0.005391558,-0.016781142,-0.0058747325,0.010309938,-0.013049036,0.01186276,-0.0011246296,0.0062112883,0.0028190718,-0.021739509,0.009883412,-0.0073175905,-0.012715813,-0.017181009,-0.016607866,-0.042492677,-0.0014478565,-0.01794076,0.012302616,-0.015194997,-0.04433207,-0.020606548,0.009696807,0.010303274,-0.01694109,-0.004018677,0.019353628,-0.001991011,0.000058938927,0.010536531,-0.17274313,0.010143327,0.014235313,-0.024152048,0.025684876,-0.0012504216,0.036601283,-0.003698782,0.0007310093,0.004165295,-0.0029157067,0.017101036,-0.046891227,-0.017460918,0.022965772,0.020233337,-0.024072073,0.017220996,0.009370248,0.0010363255,0.0194336,-0.019606877,0.01818068,-0.020819811,0.007410893,0.0019326969,0.017887443,0.006651143,0.00067394477,-0.011889419,-0.025058415,-0.008543854,0.021579562,0.0047484366,0.014062037,0.0075508473,-0.009510202,-0.009143656,0.0046817916,0.013982063,-0.0027990784,0.011782787,0.014541878,-0.015701497,-0.029350337,0.021979429,0.01332228,-0.026244693,-0.0123492675,-0.003895384,0.0071576433,-0.035454992,-0.00046984528,0.0033522295,0.039347045,0.0005119148,0.00476843,-0.012995721,0.0024042083,-0.006931051,-0.014461905,-0.0127558,0.0034555288,-0.0074842023,-0.030256703,-0.007057676,-0.00807734,0.007804097,-0.006957709,0.017181009,-0.034575284,-0.008603834,-0.005008351,-0.015834786,0.02943031,0.016861115,-0.0050849924,0.014235313,0.0051449724,0.0025924798,-0.0025741523,0.04289254,-0.002104307,0.012969063,-0.008310596,0.00423194,0.0074975314,0.0018810473,-0.014248641,-0.024725191,0.0151016945,-0.017527562,0.0018727167,0.0002830318,0.015168339,0.0144219175,-0.004048667,-0.004358565,0.011836103,-0.010343261,-0.005911387,0.0022825818,0.0073175905,0.00403867,0.013188991,0.03334902,0.006111321,0.008597169,0.030123414,-0.015474904,0.0017877447,-0.024551915,0.013155668,0.023525586,-0.0255116,0.017220996,0.004358565,-0.00934359,0.0099967085,0.011162991,0.03092315,-0.021046404,-0.015514892,0.0011946067,-0.01816735,0.010876419,-0.10124666,-0.03550831,0.0056348112,0.013942076,0.005951374,0.020419942,-0.006857742,-0.020873128,-0.021259667,0.0137554705,0.0057880944,-0.029163731,-0.018767154,-0.021392956,0.030896494,-0.005494857,-0.0027307675,-0.006801094,-0.014821786,0.021392956,-0.0018110704,-0.0018843795,-0.012362596,-0.0072176233,-0.017194338,-0.018713837,-0.024272008,0.03801415,0.00015880188,0.0044951867,-0.028630573,-0.0014070367,-0.00916365,-0.026537929,-0.009576847,-0.013995391,-0.0077107945,0.0050016865,0.00578143,-0.04467862,0.008363913,0.010136662,-0.0006268769,-0.006591163,0.015341615,-0.027377652,-0.00093136,0.029243704,-0.020886457,-0.01041657,-0.02424535,0.005291591,-0.02980352,-0.009190307,0.019460259,-0.0041286405,0.004801752,0.0011787785,-0.001257086,-0.011216307,-0.013395589,0.00088137644,-0.0051616337,0.03876057,-0.0033455652,0.00075850025,-0.006951045,-0.0062112883,0.018140694,-0.006351242,-0.008263946,0.018154023,-0.012189319,0.0075508473,-0.044358727,-0.0040153447,0.0093302615,-0.010636497,0.032789204,-0.005264933,-0.014235313,-0.018393943,0.007297597,-0.016114693,0.015021721,0.020033404,0.0137688,0.0011046362,0.010616505,-0.0039453674,0.012109346,0.021099718,-0.0072842683,-0.019153694,-0.003768759,0.039320387,-0.006747778,-0.0016852784,0.018154023,0.0010963057,-0.015035049,-0.021033075,-0.04345236,0.017287642,0.016341286,-0.008610498,0.00236922,0.009290274,0.028950468,-0.014475234,-0.0035654926,0.015434918,-0.03372223,0.004501851,-0.012929076,-0.008483873,-0.0044685286,-0.0102233,0.01615468,0.0022792495,0.010876419,-0.0059647025,0.01895376,-0.0069976957,-0.0042952523,0.017207667,-0.00036133936,0.0085905045,0.008084005,0.03129636,-0.016994404,-0.014915089,0.020100048,-0.012009379,-0.006684466,0.01306903,0.00015765642,-0.00530492,0.0005277429,0.015421589,0.015528221,0.032202728,-0.003485519,-0.0014286962,0.033908837,0.001367883,0.010509873,0.025271678,-0.020993087,0.019846799,0.006897729,-0.010216636,-0.00725761,0.01818068,-0.028443968,-0.011242964,-0.014435247,-0.013688826,0.006101324,-0.0022509254,0.013848773,-0.0019077052,0.017181009,0.03422873,0.005324913,-0.0035188415,0.014128681,-0.004898387,0.005038341,0.0012320944,-0.005561502,-0.017847456,0.0008538855,-0.0047884234,0.011849431,0.015421589,-0.013942076,0.0029790192,-0.013702155,0.0001199605,-0.024431955,0.019926772,0.022179363,-0.016487904,-0.03964028,0.0050849924,0.017487574,0.022792496,0.0012504216,0.004048667,-0.00997005,0.0076041627,-0.014328616,-0.020259995,0.0005598157,-0.010469886,0.0016852784,0.01716768,-0.008990373,-0.001987679,0.026417969,0.023792166,0.0046917885,-0.0071909656,-0.00032051947,-0.023259008,-0.009170313,0.02071318,-0.03156294,-0.030869836,-0.006324584,0.013795458,-0.00047151142,0.016874444,0.00947688,0.00985009,-0.029883493,0.024205362,-0.013522214,-0.015075036,-0.030603256,0.029270362,0.010503208,0.021539574,0.01743426,-0.023898797,0.022019416,-0.0068777353,0.027857494,-0.021259667,0.0025758184,0.006197959,0.006447877,-0.00025200035,-0.004941706,-0.021246338,-0.005504854,-0.008390571,-0.0097301295,0.027244363,-0.04446536,0.05216949,0.010243294,-0.016008062,0.0122493,-0.0199401,0.009077012,0.019753495,0.006431216,-0.037960835,-0.027377652,0.016381273,-0.0038620618,0.022512587,-0.010996379,-0.0015211658,-0.0102233,0.007071005,0.008230623,-0.009490209,-0.010083347,0.024431955,0.002427534,0.02828402,0.0035721571,-0.022192692,-0.011882754,0.010056688,0.0011904413,-0.01426197,-0.017500903,-0.00010985966,0.005591492,-0.0077707744,-0.012049366,0.011869425,0.00858384,-0.024698535,-0.030283362,0.020140035,0.011949399,-0.013968734,0.042732596,-0.011649498,-0.011982721,-0.016967745,-0.0060913274,-0.007130985,-0.013109017,-0.009710136};
// Specifies that the vector search will consider the 150 nearest neighbors
// in the specified index
var options = new VectorSearchOptions<EmbeddedMovie>()
{
IndexName = "vector_index",
NumberOfCandidates = 150
};
// Builds aggregation pipeline and specifies that the $vectorSearch stage
// returns 10 results
var pipeline = new EmptyPipelineDefinition<EmbeddedMovie>()
.VectorSearch(m => m.PlotEmbedding, vector, 10, options)
.Project(Builders<EmbeddedMovie>.Projection
.Include(m => m.Title)
.Include(m => m.Plot)
.MetaVectorSearchScore(m => m.Score);

前の例の結果には、次のドキュメントが含まれています。

{ "_id" : { "$oid" : "573a13a0f29313caabd04a4f" }, "plot" : "A reporter, learning of time travelers visiting 20th century disasters, tries to change the history they know by averting upcoming disasters.", "title" : "Thrill Seekers", "score" : 0.926971435546875 }
{ "_id" : { "$oid" : "573a13d8f29313caabda6557" }, "plot" : "At the age of 21, Tim discovers he can travel in time and change what happens and has happened in his own life. His decision to make his world a better place by getting a girlfriend turns out not to be as easy as you might think.", "title" : "About Time", "score" : 0. 9267120361328125 }
{ "_id" : { "$oid" : "573a1399f29313caabceec0e" }, "plot" : "An officer for a security agency that regulates time travel, must fend for his life against a shady politician who has a tie to his past.", "title" : "Timecop", "score" : 0.9235687255859375 }
{ "_id" : { "$oid" : "573a13a5f29313caabd13b4b" }, "plot" : "Hoping to alter the events of the past, a 19th century inventor instead travels 800,000 years into the future, where he finds humankind divided into two warring races.", "title" : "The Time Machine", "score" : 0.9228668212890625 }
{ "_id" : { "$oid" : "573a13aef29313caabd2e2d7" }, "plot" : "After using his mother's newly built time machine, Dolf gets stuck involuntary in the year 1212. He ends up in a children's crusade where he confronts his new friends with modern techniques...", "title" : "Crusade in Jeans", "score" : 0.9228515625 }
{ "_id" : { "$oid" : "573a1399f29313caabcee36f" }, "plot" : "A time-travel experiment in which a robot probe is sent from the year 2073 to the year 1973 goes terribly wrong thrusting one of the project scientists, a man named Nicholas Sinclair into a...", "title" : "A.P.E.X.", "score" : 0.9199066162109375 }
{ "_id" : { "$oid" : "573a13c6f29313caabd715d3" }, "plot" : "Agent J travels in time to M.I.B.'s early days in 1969 to stop an alien from assassinating his friend Agent K and changing history.", "title" : "Men in Black 3", "score" : 0.919403076171875 }
{ "_id" : { "$oid" : "573a13d4f29313caabd98c13" }, "plot" : "Bound by a shared destiny, a teen bursting with scientific curiosity and a former boy-genius inventor embark on a mission to unearth the secrets of a place somewhere in time and space that exists in their collective memory.", "title" : "Tomorrowland", "score" : 0.9191131591796875 }
{ "_id" : { "$oid" : "573a13b6f29313caabd477fa" }, "plot" : "With the help of his uncle, a man travels to the future to try and bring his girlfriend back to life.", "title" : "Love Story 2050", "score" : 0. 917755126953125 }
{ "_id" : { "$oid" : "573a13b3f29313caabd3ebd4" }, "plot" : "A romantic drama about a Chicago librarian with a gene that causes him to involuntarily time travel, and the complications it creates for his marriage.", "title" : "The Time Traveler's Wife", "score" : 0.9172210693359375 }

次のコードサンプルは、前の例と同じベクトル検索クエリを実行しますが、集計パイプライン構文の代わりに LINQ 構文を使用します。

var results = collection.AsQueryable()
.VectorSearch(m => m.PlotEmbedding, vector, 10, options)
.Select(m => new { m.Title, m.Plot });

Atlas Vector Search の詳細については、MongoDB Atlas ドキュメントの Atlas Vector Search ガイドを参照してください。

このガイドで説明した関数や型の詳細については、次のAPIドキュメントを参照してください。

戻る

Atlas Search

項目一覧