|
| 1 | +from chromadb.api.types import ( |
| 2 | + Embeddings, |
| 3 | + Documents, |
| 4 | + EmbeddingFunction, |
| 5 | + Space, |
| 6 | +) |
| 7 | +from chromadb.utils.embedding_functions.schemas import validate_config_schema |
| 8 | +from typing import List, Dict, Any, TypedDict, Optional |
| 9 | +import os |
| 10 | +import numpy as np |
| 11 | + |
| 12 | + |
| 13 | +class NomicQueryConfig(TypedDict): |
| 14 | + task_type: str |
| 15 | + |
| 16 | + |
| 17 | +class NomicEmbeddingFunction(EmbeddingFunction[Documents]): |
| 18 | + """ |
| 19 | + This class is used to get embeddings for a list of texts using the Nomic API. |
| 20 | + """ |
| 21 | + |
| 22 | + def __init__( |
| 23 | + self, |
| 24 | + model: str, |
| 25 | + task_type: str, |
| 26 | + query_config: Optional[NomicQueryConfig], |
| 27 | + api_key_env_var: str = "NOMIC_API_KEY", |
| 28 | + ): |
| 29 | + """ |
| 30 | + Initialize the NomicEmbeddingFunction. |
| 31 | +
|
| 32 | + Args: |
| 33 | + model (str): The name of the model to use for text embeddings. |
| 34 | + task_type (str): The type of task to embed with. See reference https://docs.nomic.ai/platform/embeddings-and-retrieval/text-embedding#embedding-task-types |
| 35 | + query_config (Optional[NomicQueryConfig]): The configuration for setting task type for queries |
| 36 | + api_key_env_var (str): The environment variable name for the Nomic API key. Defaults to "NOMIC_API_KEY". |
| 37 | +
|
| 38 | + Supported task types: search_document, search_query, classification, clustering |
| 39 | + """ |
| 40 | + try: |
| 41 | + from nomic import embed |
| 42 | + except ImportError: |
| 43 | + raise ValueError( |
| 44 | + "The nomic python package is not installed. Please install it with `pip install nomic`" |
| 45 | + ) |
| 46 | + |
| 47 | + self.model = model |
| 48 | + self.task_type = task_type |
| 49 | + self.api_key_env_var = api_key_env_var |
| 50 | + self.api_key = os.getenv(api_key_env_var) |
| 51 | + self.query_config = query_config |
| 52 | + if not self.api_key: |
| 53 | + raise ValueError(f"The {api_key_env_var} environment variable is not set.") |
| 54 | + self.embed = embed |
| 55 | + |
| 56 | + def __call__(self, input: Documents) -> Embeddings: |
| 57 | + if not all(isinstance(item, str) for item in input): |
| 58 | + raise ValueError("Nomic only supports text documents, not images") |
| 59 | + output = self.embed.text( |
| 60 | + model=self.model, |
| 61 | + texts=input, |
| 62 | + task_type=self.task_type, |
| 63 | + ) |
| 64 | + return [np.array(data.embedding) for data in output.data] |
| 65 | + |
| 66 | + def embed_query(self, input: Documents) -> Embeddings: |
| 67 | + if not all(isinstance(item, str) for item in input): |
| 68 | + raise ValueError("Nomic only supports text queries, not images") |
| 69 | + |
| 70 | + task_type = ( |
| 71 | + self.query_config.get("task_type") if self.query_config else self.task_type |
| 72 | + ) |
| 73 | + output = self.embed.text( |
| 74 | + model=self.model, |
| 75 | + texts=input, |
| 76 | + task_type=task_type, |
| 77 | + ) |
| 78 | + return [np.array(data.embedding) for data in output.data] |
| 79 | + |
| 80 | + @staticmethod |
| 81 | + def name() -> str: |
| 82 | + return "nomic" |
| 83 | + |
| 84 | + def default_space(self) -> Space: |
| 85 | + return "cosine" |
| 86 | + |
| 87 | + def supported_spaces(self) -> List[Space]: |
| 88 | + return ["cosine", "l2", "ip"] |
| 89 | + |
| 90 | + @staticmethod |
| 91 | + def build_from_config(config: Dict[str, Any]) -> "EmbeddingFunction[Documents]": |
| 92 | + model = config.get("model") |
| 93 | + api_key_env_var = config.get("api_key_env_var") |
| 94 | + task_type = config.get("task_type") |
| 95 | + query_config = config.get("query_config") |
| 96 | + if model is None or api_key_env_var is None or task_type is None: |
| 97 | + assert False, "This code should not be reached" # this is for type checking |
| 98 | + return NomicEmbeddingFunction( |
| 99 | + model=model, |
| 100 | + api_key_env_var=api_key_env_var, |
| 101 | + task_type=task_type, |
| 102 | + query_config=query_config, |
| 103 | + ) |
| 104 | + |
| 105 | + def get_config(self) -> Dict[str, Any]: |
| 106 | + return { |
| 107 | + "model": self.model, |
| 108 | + "api_key_env_var": self.api_key_env_var, |
| 109 | + "task_type": self.task_type, |
| 110 | + "query_config": self.query_config, |
| 111 | + } |
| 112 | + |
| 113 | + def validate_config_update( |
| 114 | + self, old_config: Dict[str, Any], new_config: Dict[str, Any] |
| 115 | + ) -> None: |
| 116 | + if "model" in new_config: |
| 117 | + raise ValueError( |
| 118 | + "The model cannot be changed after the embedding function has been initialized." |
| 119 | + ) |
| 120 | + |
| 121 | + @staticmethod |
| 122 | + def validate_config(config: Dict[str, Any]) -> None: |
| 123 | + """ |
| 124 | + Validate the configuration using the JSON schema. |
| 125 | +
|
| 126 | + Args: |
| 127 | + config: Configuration to validate |
| 128 | + """ |
| 129 | + validate_config_schema(config, "nomic") |
0 commit comments