add test cases for embeddings (#1829)

tests/embeddings/test_azure_openai_embeddings.py

@@ -0,0 +1,46 @@
+import pytest
+from unittest.mock import Mock, patch
+from mem0.embeddings.azure_openai import AzureOpenAIEmbedding
+from mem0.configs.embeddings.base import BaseEmbedderConfig
+
+
+@pytest.fixture
+def mock_openai_client():
+    with patch("mem0.embeddings.azure_openai.AzureOpenAI") as mock_openai:
+        mock_client = Mock()
+        mock_openai.return_value = mock_client
+        yield mock_client
+
+
+def test_embed_text(mock_openai_client):
+    config = BaseEmbedderConfig(model="text-embedding-ada-002")
+    embedder = AzureOpenAIEmbedding(config)
+
+    mock_embedding_response = Mock()
+    mock_embedding_response.data = [Mock(embedding=[0.1, 0.2, 0.3])]
+    mock_openai_client.embeddings.create.return_value = mock_embedding_response
+
+    text = "Hello, this is a test."
+    embedding = embedder.embed(text)
+
+    mock_openai_client.embeddings.create.assert_called_once_with(
+        input=["Hello, this is a test."], model="text-embedding-ada-002"
+    )
+    assert embedding == [0.1, 0.2, 0.3]
+
+
+def test_embed_text_with_newlines(mock_openai_client):
+    config = BaseEmbedderConfig(model="text-embedding-ada-002")
+    embedder = AzureOpenAIEmbedding(config)
+
+    mock_embedding_response = Mock()
+    mock_embedding_response.data = [Mock(embedding=[0.4, 0.5, 0.6])]
+    mock_openai_client.embeddings.create.return_value = mock_embedding_response
+
+    text = "Hello,\nthis is a test\nwith newlines."
+    embedding = embedder.embed(text)
+
+    mock_openai_client.embeddings.create.assert_called_once_with(
+        input=["Hello, this is a test with newlines."], model="text-embedding-ada-002"
+    )
+    assert embedding == [0.4, 0.5, 0.6]

tests/embeddings/test_huggingface_embeddings.py

@@ -0,0 +1,74 @@
+import pytest
+from unittest.mock import Mock, patch
+from mem0.embeddings.huggingface import HuggingFaceEmbedding
+from mem0.configs.embeddings.base import BaseEmbedderConfig
+
+
+@pytest.fixture
+def mock_sentence_transformer():
+    with patch("mem0.embeddings.huggingface.SentenceTransformer") as mock_transformer:
+        mock_model = Mock()
+        mock_transformer.return_value = mock_model
+        yield mock_model
+
+
+def test_embed_default_model(mock_sentence_transformer):
+    config = BaseEmbedderConfig()
+    embedder = HuggingFaceEmbedding(config)
+
+    mock_sentence_transformer.encode.return_value = [0.1, 0.2, 0.3]
+    result = embedder.embed("Hello world")
+
+    mock_sentence_transformer.encode.assert_called_once_with("Hello world")
+
+    assert result == [0.1, 0.2, 0.3]
+
+
+def test_embed_custom_model(mock_sentence_transformer):
+    config = BaseEmbedderConfig(model="paraphrase-MiniLM-L6-v2")
+    embedder = HuggingFaceEmbedding(config)
+
+    mock_sentence_transformer.encode.return_value = [0.4, 0.5, 0.6]
+    result = embedder.embed("Custom model test")
+
+    mock_sentence_transformer.encode.assert_called_once_with("Custom model test")
+
+    assert result == [0.4, 0.5, 0.6]
+
+
+def test_embed_with_model_kwargs(mock_sentence_transformer):
+    config = BaseEmbedderConfig(
+        model="all-MiniLM-L6-v2", model_kwargs={"device": "cuda"}
+    )
+    embedder = HuggingFaceEmbedding(config)
+
+    mock_sentence_transformer.encode.return_value = [0.7, 0.8, 0.9]
+    result = embedder.embed("Test with device")
+
+    mock_sentence_transformer.encode.assert_called_once_with("Test with device")
+
+    assert result == [0.7, 0.8, 0.9]
+
+
+def test_embed_sets_embedding_dims(mock_sentence_transformer):
+    config = BaseEmbedderConfig()
+
+    mock_sentence_transformer.get_sentence_embedding_dimension.return_value = 384
+    embedder = HuggingFaceEmbedding(config)
+
+    assert embedder.config.embedding_dims == 384
+    mock_sentence_transformer.get_sentence_embedding_dimension.assert_called_once()
+
+
+def test_embed_with_custom_embedding_dims(mock_sentence_transformer):
+    config = BaseEmbedderConfig(model="all-mpnet-base-v2", embedding_dims=768)
+    embedder = HuggingFaceEmbedding(config)
+
+    mock_sentence_transformer.encode.return_value = [1.0, 1.1, 1.2]
+    result = embedder.embed("Custom embedding dims")
+
+    mock_sentence_transformer.encode.assert_called_once_with("Custom embedding dims")
+
+    assert embedder.config.embedding_dims == 768
+
+    assert result == [1.0, 1.1, 1.2]

tests/embeddings/test_ollama_embeddings.py

@@ -0,0 +1,43 @@
+import pytest
+from unittest.mock import Mock, patch
+from mem0.embeddings.ollama import OllamaEmbedding
+from mem0.configs.embeddings.base import BaseEmbedderConfig
+
+
+@pytest.fixture
+def mock_ollama_client():
+    with patch("mem0.embeddings.ollama.Client") as mock_ollama:
+        mock_client = Mock()
+        mock_client.list.return_value = {"models": [{"name": "nomic-embed-text"}]}
+        mock_ollama.return_value = mock_client
+        yield mock_client
+
+
+def test_embed_text(mock_ollama_client):
+    config = BaseEmbedderConfig(model="nomic-embed-text", embedding_dims=512)
+    embedder = OllamaEmbedding(config)
+
+    mock_response = {"embedding": [0.1, 0.2, 0.3, 0.4, 0.5]}
+    mock_ollama_client.embeddings.return_value = mock_response
+
+    text = "Sample text to embed."
+    embedding = embedder.embed(text)
+
+    mock_ollama_client.embeddings.assert_called_once_with(
+        model="nomic-embed-text", prompt=text
+    )
+
+    assert embedding == [0.1, 0.2, 0.3, 0.4, 0.5]
+
+
+def test_ensure_model_exists(mock_ollama_client):
+    config = BaseEmbedderConfig(model="nomic-embed-text", embedding_dims=512)
+    embedder = OllamaEmbedding(config)
+
+    mock_ollama_client.pull.assert_not_called()
+
+    mock_ollama_client.list.return_value = {"models": []}
+
+    embedder._ensure_model_exists()
+
+    mock_ollama_client.pull.assert_called_once_with("nomic-embed-text")

tests/embeddings/test_openai_embeddings.py

@@ -0,0 +1,88 @@
+import pytest
+from unittest.mock import Mock, patch
+from mem0.embeddings.openai import OpenAIEmbedding
+from mem0.configs.embeddings.base import BaseEmbedderConfig
+
+
+@pytest.fixture
+def mock_openai_client():
+    with patch("mem0.embeddings.openai.OpenAI") as mock_openai:
+        mock_client = Mock()
+        mock_openai.return_value = mock_client
+        yield mock_client
+
+
+def test_embed_default_model(mock_openai_client):
+    config = BaseEmbedderConfig()
+    embedder = OpenAIEmbedding(config)
+    mock_response = Mock()
+    mock_response.data = [Mock(embedding=[0.1, 0.2, 0.3])]
+    mock_openai_client.embeddings.create.return_value = mock_response
+
+    result = embedder.embed("Hello world")
+
+    mock_openai_client.embeddings.create.assert_called_once_with(
+        input=["Hello world"], model="text-embedding-3-small"
+    )
+    assert result == [0.1, 0.2, 0.3]
+
+
+def test_embed_custom_model(mock_openai_client):
+    config = BaseEmbedderConfig(model="text-embedding-2-medium", embedding_dims=1024)
+    embedder = OpenAIEmbedding(config)
+    mock_response = Mock()
+    mock_response.data = [Mock(embedding=[0.4, 0.5, 0.6])]
+    mock_openai_client.embeddings.create.return_value = mock_response
+
+    result = embedder.embed("Test embedding")
+
+    mock_openai_client.embeddings.create.assert_called_once_with(
+        input=["Test embedding"], model="text-embedding-2-medium"
+    )
+    assert result == [0.4, 0.5, 0.6]
+
+
+def test_embed_removes_newlines(mock_openai_client):
+    config = BaseEmbedderConfig()
+    embedder = OpenAIEmbedding(config)
+    mock_response = Mock()
+    mock_response.data = [Mock(embedding=[0.7, 0.8, 0.9])]
+    mock_openai_client.embeddings.create.return_value = mock_response
+
+    result = embedder.embed("Hello\nworld")
+
+    mock_openai_client.embeddings.create.assert_called_once_with(
+        input=["Hello world"], model="text-embedding-3-small"
+    )
+    assert result == [0.7, 0.8, 0.9]
+
+
+def test_embed_without_api_key_env_var(mock_openai_client):
+    config = BaseEmbedderConfig(api_key="test_key")
+    embedder = OpenAIEmbedding(config)
+    mock_response = Mock()
+    mock_response.data = [Mock(embedding=[1.0, 1.1, 1.2])]
+    mock_openai_client.embeddings.create.return_value = mock_response
+
+    result = embedder.embed("Testing API key")
+
+    mock_openai_client.embeddings.create.assert_called_once_with(
+        input=["Testing API key"], model="text-embedding-3-small"
+    )
+    assert result == [1.0, 1.1, 1.2]
+
+
+def test_embed_uses_environment_api_key(mock_openai_client, monkeypatch):
+    monkeypatch.setenv("OPENAI_API_KEY", "env_key")
+    config = BaseEmbedderConfig()
+    embedder = OpenAIEmbedding(config)
+    mock_response = Mock()
+    mock_response.data = [Mock(embedding=[1.3, 1.4, 1.5])]
+    mock_openai_client.embeddings.create.return_value = mock_response
+
+    result = embedder.embed("Environment key test")
+
+    mock_openai_client.embeddings.create.assert_called_once_with(
+        input=["Environment key test"], model="text-embedding-3-small"
+    )
+    assert result == [1.3, 1.4, 1.5]