or run

npx @tessl/cli init
Log in

Version

Tile

Overview

Evals

Files

Files

docs

batch.mddatasets.mdexperiments.mdfeature-store.mdgenerative-ai.mdindex.mdmodels.mdpipelines.mdtraining.mdvector-search.mdvision.md

vector-search.mddocs/

0

# Vector Search

1

2

High-performance vector similarity search with approximate nearest neighbor capabilities for embedding-based applications.

3

4

## Capabilities

5

6

### Matching Engine Index

7

8

Create and manage vector indices for similarity search with configurable algorithms and performance settings.

9

10

```python { .api }

11

class MatchingEngineIndex:

12

@classmethod

13

def create(

14

cls,

15

display_name: str,

16

contents_delta_uri: str,

17

config: Optional[Dict] = None,

18

labels: Optional[Dict[str, str]] = None,

19

description: Optional[str] = None,

20

**kwargs

21

) -> 'MatchingEngineIndex': ...

22

23

def update_embeddings(

24

self,

25

contents_delta_uri: str,

26

is_complete_overwrite: bool = False,

27

**kwargs

28

) -> None: ...

29

30

def upsert_datapoints(

31

self,

32

datapoints: List[Dict[str, Any]],

33

update_mask: Optional[str] = None,

34

**kwargs

35

) -> None: ...

36

37

def remove_datapoints(

38

self,

39

datapoint_ids: List[str],

40

**kwargs

41

) -> None: ...

42

```

43

44

### Index Endpoints

45

46

Deploy indices to endpoints for serving similarity queries with traffic management.

47

48

```python { .api }

49

class MatchingEngineIndexEndpoint:

50

@classmethod

51

def create(

52

cls,

53

display_name: str,

54

network: Optional[str] = None,

55

public_endpoint_enabled: bool = False,

56

labels: Optional[Dict[str, str]] = None,

57

description: Optional[str] = None,

58

**kwargs

59

) -> 'MatchingEngineIndexEndpoint': ...

60

61

def deploy_index(

62

self,

63

index: MatchingEngineIndex,

64

deployed_index_id: str,

65

display_name: Optional[str] = None,

66

machine_type: str = 'e2-standard-2',

67

min_replica_count: int = 1,

68

max_replica_count: int = 1,

69

enable_access_logging: bool = False,

70

**kwargs

71

) -> None: ...

72

73

def match(

74

self,

75

deployed_index_id: str,

76

queries: List[List[float]],

77

num_neighbors: int = 1,

78

filter: Optional[List[Dict[str, Any]]] = None,

79

**kwargs

80

) -> List[List[MatchNeighbor]]: ...

81

82

def batch_get_embeddings(

83

self,

84

requests: List[Dict[str, Any]],

85

**kwargs

86

) -> List[List[float]]: ...

87

```