embeddings

Documentation

Index

• func CosineDistance(v1, v2 []float64) float64
• func CreateEmbedding(ollamaUrl string, query llm.Query4Embedding, id string) (llm.VectorRecord, error)
• func GenerateContentFromSimilarities(similarities []llm.VectorRecord) string
• func GenerateContextFromSimilarities(similarities []llm.VectorRecord) string
• type BboltVectorStore
  • func (bvs *BboltVectorStore) Get(id string) (llm.VectorRecord, error)
  • func (bvs *BboltVectorStore) GetAll() ([]llm.VectorRecord, error)
  • func (bvs *BboltVectorStore) Initialize(dbPath string) error
  • func (bvs *BboltVectorStore) Save(vectorRecord llm.VectorRecord) (llm.VectorRecord, error)
  • func (bvs *BboltVectorStore) SearchMaxSimilarity(embeddingFromQuestion llm.VectorRecord) (llm.VectorRecord, error)
  • func (bvs *BboltVectorStore) SearchSimilarities(embeddingFromQuestion llm.VectorRecord, limit float64) ([]llm.VectorRecord, error)
  • func (bvs *BboltVectorStore) SearchTopNSimilarities(embeddingFromQuestion llm.VectorRecord, limit float64, max int) ([]llm.VectorRecord, error)
• type ElasticSearchStore
  • func (ess *ElasticSearchStore) Initialize(addresses []string, user, pwd string, cert []byte, indexName string) error
  • func (ess *ElasticSearchStore) Save(vectorRecord llm.VectorRecord) (llm.VectorRecord, error)
  • func (ess *ElasticSearchStore) SearchTopNSimilarities(embeddingFromQuestion llm.VectorRecord, size int) ([]llm.VectorRecord, error)
• type EmbeddingResponse
• type MemoryVectorStore
  • func (mvs *MemoryVectorStore) Get(id string) (llm.VectorRecord, error)
  • func (mvs *MemoryVectorStore) GetAll() ([]llm.VectorRecord, error)
  • func (mvs *MemoryVectorStore) Save(vectorRecord llm.VectorRecord) (llm.VectorRecord, error)
  • func (mvs *MemoryVectorStore) SearchMaxSimilarity(embeddingFromQuestion llm.VectorRecord) (llm.VectorRecord, error)
  • func (mvs *MemoryVectorStore) SearchSimilarities(embeddingFromQuestion llm.VectorRecord, limit float64) ([]llm.VectorRecord, error)
  • func (mvs *MemoryVectorStore) SearchTopNSimilarities(embeddingFromQuestion llm.VectorRecord, limit float64, max int) ([]llm.VectorRecord, error)
• type RedisVectorStore
  • func (rvs *RedisVectorStore) Get(id string) (llm.VectorRecord, error)
  • func (rvs *RedisVectorStore) GetAll() ([]llm.VectorRecord, error)
  • func (rvs *RedisVectorStore) Initialize(redisAddr string, redisPwd string, storeName string) error
  • func (rvs *RedisVectorStore) Save(vectorRecord llm.VectorRecord) (llm.VectorRecord, error)
  • func (rvs *RedisVectorStore) SearchMaxSimilarity(embeddingFromQuestion llm.VectorRecord) (llm.VectorRecord, error)
  • func (rvs *RedisVectorStore) SearchSimilarities(embeddingFromQuestion llm.VectorRecord, limit float64) ([]llm.VectorRecord, error)
  • func (rvs *RedisVectorStore) SearchTopNSimilarities(embeddingFromQuestion llm.VectorRecord, limit float64, max int) ([]llm.VectorRecord, error)
• type VectorStore

Functions

func CosineDistance

func CosineDistance(v1, v2 []float64) float64

func CreateEmbedding

func CreateEmbedding(ollamaUrl string, query llm.Query4Embedding, id string) (llm.VectorRecord, error)

func GenerateContentFromSimilarities

func GenerateContentFromSimilarities(similarities []llm.VectorRecord) string

func GenerateContextFromSimilarities

func GenerateContextFromSimilarities(similarities []llm.VectorRecord) string

GenerateContextFromSimilarities generates the context content from a slice of vector records.

Parameters: - similarities: a slice of llm.VectorRecord representing the similarities.

Returns: - string: the generated context content in XML format.

Types

type BboltVectorStore

type BboltVectorStore struct {
	// contains filtered or unexported fields
}

func (*BboltVectorStore) Get

func (bvs *BboltVectorStore) Get(id string) (llm.VectorRecord, error)

func (*BboltVectorStore) GetAll

func (bvs *BboltVectorStore) GetAll() ([]llm.VectorRecord, error)

func (*BboltVectorStore) Initialize

func (bvs *BboltVectorStore) Initialize(dbPath string) error

func (*BboltVectorStore) Save

func (bvs *BboltVectorStore) Save(vectorRecord llm.VectorRecord) (llm.VectorRecord, error)

func (*BboltVectorStore) SearchMaxSimilarity

func (bvs *BboltVectorStore) SearchMaxSimilarity(embeddingFromQuestion llm.VectorRecord) (llm.VectorRecord, error)

SearchMaxSimilarity searches for the vector record in the BboltVectorStore that has the maximum cosine distance similarity to the given embeddingFromQuestion.

Parameters: - embeddingFromQuestion: the vector record to compare similarities with.

Returns: - llm.VectorRecord: the vector record with the maximum cosine distance similarity. - error: an error if any occurred during the search.

func (*BboltVectorStore) SearchSimilarities

func (bvs *BboltVectorStore) SearchSimilarities(embeddingFromQuestion llm.VectorRecord, limit float64) ([]llm.VectorRecord, error)

SearchSimilarities searches for vector records in the BboltVectorStore that have a cosine distance similarity greater than or equal to the given limit.

Parameters:

• embeddingFromQuestion: the vector record to compare similarities with.
• limit: the minimum cosine distance similarity threshold.

Returns:

• []llm.VectorRecord: a slice of vector records that have a cosine distance similarity greater than or equal to the limit.
• error: an error if any occurred during the search.

func (*BboltVectorStore) SearchTopNSimilarities

func (bvs *BboltVectorStore) SearchTopNSimilarities(embeddingFromQuestion llm.VectorRecord, limit float64, max int) ([]llm.VectorRecord, error)

SearchTopNSimilarities searches for the top N similar vector records based on the given embedding from a question. It returns a slice of vector records and an error if any. The limit parameter specifies the minimum similarity score for a record to be considered similar. The max parameter specifies the maximum number of vector records to return.

type ElasticSearchStore

type ElasticSearchStore struct {
	// contains filtered or unexported fields
}

func (*ElasticSearchStore) Initialize

func (ess *ElasticSearchStore) Initialize(addresses []string, user, pwd string, cert []byte, indexName string) error

func (*ElasticSearchStore) Save

func (ess *ElasticSearchStore) Save(vectorRecord llm.VectorRecord) (llm.VectorRecord, error)

func (*ElasticSearchStore) SearchTopNSimilarities

func (ess *ElasticSearchStore) SearchTopNSimilarities(embeddingFromQuestion llm.VectorRecord, size int) ([]llm.VectorRecord, error)

type EmbeddingResponse

type EmbeddingResponse struct {
	Embedding []float64 `json:"embedding"`
}

type MemoryVectorStore

type MemoryVectorStore struct {
	Records map[string]llm.VectorRecord
}

func (*MemoryVectorStore) Get

func (mvs *MemoryVectorStore) Get(id string) (llm.VectorRecord, error)

func (*MemoryVectorStore) GetAll

func (mvs *MemoryVectorStore) GetAll() ([]llm.VectorRecord, error)

func (*MemoryVectorStore) Save

func (mvs *MemoryVectorStore) Save(vectorRecord llm.VectorRecord) (llm.VectorRecord, error)

func (*MemoryVectorStore) SearchMaxSimilarity

func (mvs *MemoryVectorStore) SearchMaxSimilarity(embeddingFromQuestion llm.VectorRecord) (llm.VectorRecord, error)

SearchMaxSimilarity finds the vector record in MemoryVectorStore with the maximum cosine distance similarity to the provided vector record.

Parameters:

• embeddingFromQuestion: llm.VectorRecord - the vector record to compare similarities with.

Returns:

• llm.VectorRecord: The vector record with the maximum similarity.
• error: Error if any.

func (*MemoryVectorStore) SearchSimilarities

func (mvs *MemoryVectorStore) SearchSimilarities(embeddingFromQuestion llm.VectorRecord, limit float64) ([]llm.VectorRecord, error)

SearchSimilarities searches for vector records in the MemoryVectorStore that have a cosine distance similarity greater than or equal to the given limit.

Parameters:

• embeddingFromQuestion: the vector record to compare similarities with.
• limit: the minimum cosine distance similarity threshold.

Returns:

• []llm.VectorRecord: a slice of vector records that have a cosine distance similarity greater than or equal to the limit.
• error: an error if any occurred during the search.

func (*MemoryVectorStore) SearchTopNSimilarities

func (mvs *MemoryVectorStore) SearchTopNSimilarities(embeddingFromQuestion llm.VectorRecord, limit float64, max int) ([]llm.VectorRecord, error)

SearchTopNSimilarities searches for the top N similar vector records based on the given embedding from a question. It returns a slice of vector records and an error if any. The limit parameter specifies the minimum similarity score for a record to be considered similar. The max parameter specifies the maximum number of vector records to return.

type RedisVectorStore

type RedisVectorStore struct {
	// contains filtered or unexported fields
}

func (*RedisVectorStore) Get

func (rvs *RedisVectorStore) Get(id string) (llm.VectorRecord, error)

func (*RedisVectorStore) GetAll

func (rvs *RedisVectorStore) GetAll() ([]llm.VectorRecord, error)

func (*RedisVectorStore) Initialize

func (rvs *RedisVectorStore) Initialize(redisAddr string, redisPwd string, storeName string) error

func (*RedisVectorStore) Save

func (rvs *RedisVectorStore) Save(vectorRecord llm.VectorRecord) (llm.VectorRecord, error)

func (*RedisVectorStore) SearchMaxSimilarity

func (rvs *RedisVectorStore) SearchMaxSimilarity(embeddingFromQuestion llm.VectorRecord) (llm.VectorRecord, error)

SearchMaxSimilarity searches for the vector record in the RedisVectorStore that has the maximum cosine distance similarity to the given embeddingFromQuestion.

Parameters: - embeddingFromQuestion: the vector record to compare similarities with.

Returns: - llm.VectorRecord: the vector record with the maximum cosine distance similarity. - error: an error if any occurred during the search.

func (*RedisVectorStore) SearchSimilarities

func (rvs *RedisVectorStore) SearchSimilarities(embeddingFromQuestion llm.VectorRecord, limit float64) ([]llm.VectorRecord, error)

SearchSimilarities searches for vector records in the RedisVectorStore that have a cosine distance similarity greater than or equal to the given limit.

Parameters:

• embeddingFromQuestion: the vector record to compare similarities with.
• limit: the minimum cosine distance similarity threshold.

Returns:

• []llm.VectorRecord: a slice of vector records that have a cosine distance similarity greater than or equal to the limit.
• error: an error if any occurred during the search.

func (*RedisVectorStore) SearchTopNSimilarities

func (rvs *RedisVectorStore) SearchTopNSimilarities(embeddingFromQuestion llm.VectorRecord, limit float64, max int) ([]llm.VectorRecord, error)

SearchTopNSimilarities searches for the top N similar vector records based on the given embedding from a question. It returns a slice of vector records and an error if any. The limit parameter specifies the minimum similarity score for a record to be considered similar. The max parameter specifies the maximum number of vector records to return.

type VectorStore

type VectorStore interface {
	Get(id string) (llm.VectorRecord, error)
	GetAll() ([]llm.VectorRecord, error)
	Save(vectorRecord llm.VectorRecord) (llm.VectorRecord, error)
	SearchMaxSimilarity(embeddingFromQuestion llm.VectorRecord) (llm.VectorRecord, error)
	SearchSimilarities(embeddingFromQuestion llm.VectorRecord, limit float64) ([]llm.VectorRecord, error)
	SearchTopNSimilarities(embeddingFromQuestion llm.VectorRecord, limit float64, max int) ([]llm.VectorRecord, error)
}