RAG-Documind-AI / tests /unit /test_embeddings.py
AIenthusSP's picture
fix: fix linting issues
a6df84a
Raw History Blame Contribute Delete
2.53 kB
"""
Unit Tests for Embeddings
Tests for embedding model functionality.
"""
class TestEmbeddings:
"""Tests for embedding functionality."""
def test_embeddings_initialization(self):
"""Test that embeddings can be initialized."""
from src.rag.embeddings import get_embeddings
embeddings = get_embeddings()
assert embeddings is not None
def test_embed_single_text(self):
"""Test embedding a single text."""
from src.rag.embeddings import get_embeddings
embeddings = get_embeddings()
text = "This is a test sentence."
vector = embeddings.embed_query(text)
assert isinstance(vector, list)
assert len(vector) > 0
assert all(isinstance(v, float) for v in vector)
def test_embed_multiple_texts(self):
"""Test embedding multiple texts."""
from src.rag.embeddings import get_embeddings
embeddings = get_embeddings()
texts = ["First sentence.", "Second sentence.", "Third sentence."]
vectors = embeddings.embed_documents(texts)
assert len(vectors) == 3
assert all(len(v) == len(vectors[0]) for v in vectors)
def test_similar_texts_have_similar_embeddings(self):
"""Test that semantically similar texts have similar embeddings."""
from src.rag.embeddings import get_embeddings
import numpy as np
embeddings = get_embeddings()
# Similar sentences
text1 = "The cat sat on the mat."
text2 = "A cat is sitting on a mat."
# Different sentence
text3 = "The stock market crashed today."
vec1 = embeddings.embed_query(text1)
vec2 = embeddings.embed_query(text2)
vec3 = embeddings.embed_query(text3)
# Cosine similarity
def cosine_sim(a, b):
return np.dot(a, b) / (np.linalg.norm(a) * np.linalg.norm(b))
sim_12 = cosine_sim(vec1, vec2)
sim_13 = cosine_sim(vec1, vec3)
# Similar sentences should have higher similarity
assert sim_12 > sim_13
def test_empty_text_embedding(self):
"""Test embedding empty text."""
from src.rag.embeddings import get_embeddings
embeddings = get_embeddings()
# Should handle empty text gracefully
vector = embeddings.embed_query("")
assert isinstance(vector, list)