Add Huggingface embeddings and Cohere embeddings (#63)

* Add huggingface embeddings and cohere embeddings
* Update openai interface and the mock for newer OpenAI SDK

---------

Co-authored-by: trducng <trungduc1992@gmail.com>

tests/test_embedding_models.py

@@ -2,6 +2,8 @@ import json
 from pathlib import Path
 from unittest.mock import patch
 
+from kotaemon.embeddings.cohere import CohereEmbdeddings
+from kotaemon.embeddings.huggingface import HuggingFaceEmbeddings
 from kotaemon.embeddings.openai import AzureOpenAIEmbeddings
 
 with open(Path(__file__).parent / "resources" / "embedding_openai_batch.json") as f:
@@ -12,7 +14,7 @@ with open(Path(__file__).parent / "resources" / "embedding_openai.json") as f:
 
 
 @patch(
-    "openai.api_resources.embedding.Embedding.create",
+    "openai.resources.embeddings.Embeddings.create",
     side_effect=lambda *args, **kwargs: openai_embedding,
 )
 def test_azureopenai_embeddings_raw(openai_embedding_call):
@@ -29,7 +31,7 @@ def test_azureopenai_embeddings_raw(openai_embedding_call):
 
 
 @patch(
-    "openai.api_resources.embedding.Embedding.create",
+    "openai.resources.embeddings.Embeddings.create",
     side_effect=lambda *args, **kwargs: openai_embedding_batch,
 )
 def test_azureopenai_embeddings_batch_raw(openai_embedding_call):
@@ -44,3 +46,42 @@ def test_azureopenai_embeddings_batch_raw(openai_embedding_call):
     assert isinstance(output[0], list)
     assert isinstance(output[0][0], float)
     openai_embedding_call.assert_called()
+
+
+@patch(
+    "sentence_transformers.SentenceTransformer",
+    side_effect=lambda *args, **kwargs: None,
+)
+@patch(
+    "langchain.embeddings.huggingface.HuggingFaceBgeEmbeddings.embed_query",
+    side_effect=lambda *args, **kwargs: [1.0, 2.1, 3.2],
+)
+def test_huggingface_embddings(
+    langchain_huggingface_embedding_call, sentence_transformers_init
+):
+    model = HuggingFaceEmbeddings(
+        model_name="intfloat/multilingual-e5-large",
+        model_kwargs={"device": "cpu"},
+        encode_kwargs={"normalize_embeddings": False},
+    )
+
+    output = model("Hello World")
+    assert isinstance(output, list)
+    assert isinstance(output[0], float)
+    sentence_transformers_init.assert_called()
+    langchain_huggingface_embedding_call.assert_called()
+
+
+@patch(
+    "langchain.embeddings.cohere.CohereEmbeddings.embed_query",
+    side_effect=lambda *args, **kwargs: [1.0, 2.1, 3.2],
+)
+def test_cohere_embddings(langchain_cohere_embedding_call):
+    model = CohereEmbdeddings(
+        model="embed-english-light-v2.0", cohere_api_key="my-api-key"
+    )
+
+    output = model("Hello World")
+    assert isinstance(output, list)
+    assert isinstance(output[0], float)
+    langchain_cohere_embedding_call.assert_called()