microsoft/TypeAgent
Publicmirrored fromhttps://github.com/microsoft/TypeAgentAvailable
python/nprData/embedding.py
42lines · modecode
| 1 | # Copyright (c) Microsoft Corporation and Henry Lucco. |
| 2 | # Licensed under the MIT License. |
| 3 | |
| 4 | from dataclasses import dataclass |
| 5 | from typing import List |
| 6 | from openai import OpenAI |
| 7 | import os |
| 8 | |
| 9 | @dataclass |
| 10 | class Embedding: |
| 11 | values: List[float] |
| 12 | dimension: int |
| 13 | |
| 14 | @classmethod |
| 15 | def from_text(cls, text: str) -> "Embedding": |
| 16 | openai_api_key = os.environ.get("OPENAI_API_KEY") |
| 17 | if not openai_api_key: |
| 18 | raise ValueError("OPENAI_API_KEY environment variable is not set") |
| 19 | |
| 20 | openai_client = OpenAI( |
| 21 | api_key=openai_api_key |
| 22 | ) |
| 23 | text = text.strip().replace("\n", " ") |
| 24 | |
| 25 | embedding_model = os.environ.get("EMBEDDING_MODEL", "text-embedding-ada-002") |
| 26 | |
| 27 | embedding_value = openai_client.embeddings.create( |
| 28 | input=[text], |
| 29 | model=embedding_model |
| 30 | ).data[0].embedding |
| 31 | |
| 32 | return cls(embedding_value, len(embedding_value)) |
| 33 | |
| 34 | @classmethod |
| 35 | def from_dict(cls, embedding_dict: dict) -> "Embedding": |
| 36 | return cls(embedding_dict["values"], embedding_dict["dimension"]) |
| 37 | |
| 38 | def to_dict(self): |
| 39 | return { |
| 40 | "values": self.values, |
| 41 | "dimension": self.dimension |
| 42 | } |