enhancement: reduced usage of numpy and substituted built-in libraries (

#8418)

* reduced usage of numpy and substituted built-in libraries

* added release note

* edited expit function to support both float as well as list (this case was giving error CI)

* revert code , numpy can't be removed here

* more cleaning

* fix relnote

---------

Co-authored-by: anakin87 <[email protected]>

haystack/components/embedders/backends/sentence_transformers_backend.py

@@ -2,9 +2,7 @@
 #
 # SPDX-License-Identifier: Apache-2.0
 
-from typing import Any, Dict, List, Optional, cast
-
-import numpy as np
+from typing import Any, Dict, List, Optional
 
 from haystack.lazy_imports import LazyImport
 from haystack.utils.auth import Secret
@@ -78,5 +76,5 @@ def __init__(
         )
 
     def embed(self, data: List[str], **kwargs) -> List[List[float]]:
-        embeddings = cast(np.ndarray, self.model.encode(data, **kwargs)).tolist()
+        embeddings = self.model.encode(data, **kwargs).tolist()
         return embeddings

haystack/testing/test_utils.py

@@ -5,8 +5,6 @@
 import os
 import random
 
-import numpy as np
-
 from haystack import logging
 
 logger = logging.getLogger(__name__)
@@ -23,7 +21,6 @@ def set_all_seeds(seed: int, deterministic_cudnn: bool = False) -> None:
     :param deterministic_cudnn: Enable for full reproducibility when using CUDA. Caution: might slow down training.
     """
     random.seed(seed)
-    np.random.seed(seed)
     os.environ["PYTHONHASHSEED"] = str(seed)
 
     try:

haystack/utils/expit.py

@@ -2,9 +2,13 @@
 #
 # SPDX-License-Identifier: Apache-2.0
 
-import numpy as np
+from numpy import exp
 
 
-def expit(x: float) -> float:
-    """Compute logistic sigmoid function. Maps input values to a range between 0 and 1"""
-    return 1 / (1 + np.exp(-x))
+def expit(x) -> float:
+    """
+    Compute logistic sigmoid function. Maps input values to a range between 0 and 1
+
+    :param x: input value. Can be a scalar or a numpy array.
+    """
+    return 1 / (1 + exp(-x))

releasenotes/notes/enhance-numpy-dependency-reduced-f8a8afa08be098d3.yaml

@@ -0,0 +1,4 @@
+---
+enhancements:
+  - |
+    Reduced numpy usage to speed up imports.

test/components/embedders/test_hugging_face_api_document_embedder.py

@@ -4,9 +4,9 @@
 import os
 from unittest.mock import MagicMock, patch
 
+import random
 import pytest
 from huggingface_hub.utils import RepositoryNotFoundError
-from numpy import array, random
 
 from haystack.components.embedders import HuggingFaceAPIDocumentEmbedder
 from haystack.dataclasses import Document
@@ -24,7 +24,7 @@ def mock_check_valid_model():
 
 
 def mock_embedding_generation(json, **kwargs):
-    response = str(array([random.rand(384) for i in range(len(json["inputs"]))]).tolist()).encode()
+    response = str([[random.random() for _ in range(384)] for _ in range(len(json["inputs"]))]).encode()
     return response
 
 

test/components/embedders/test_hugging_face_api_text_embedder.py

@@ -4,9 +4,9 @@
 import os
 from unittest.mock import MagicMock, patch
 
+import random
 import pytest
 from huggingface_hub.utils import RepositoryNotFoundError
-from numpy import array, random
 
 from haystack.components.embedders import HuggingFaceAPITextEmbedder
 from haystack.utils.auth import Secret
@@ -22,7 +22,7 @@ def mock_check_valid_model():
 
 
 def mock_embedding_generation(json, **kwargs):
-    response = str(array([random.rand(384) for i in range(len(json["inputs"]))]).tolist()).encode()
+    response = str([[random.random() for _ in range(384)] for _ in range(len(json["inputs"]))]).encode()
     return response
 
 

test/components/embedders/test_openai_document_embedder.py

@@ -5,7 +5,7 @@
 from typing import List
 from haystack.utils.auth import Secret
 
-import numpy as np
+import random
 import pytest
 
 from haystack import Document
@@ -16,7 +16,8 @@ def mock_openai_response(input: List[str], model: str = "text-embedding-ada-002"
     dict_response = {
         "object": "list",
         "data": [
-            {"object": "embedding", "index": i, "embedding": np.random.rand(1536).tolist()} for i in range(len(input))
+            {"object": "embedding", "index": i, "embedding": [random.random() for _ in range(1536)]}
+            for i in range(len(input))
         ],
         "model": model,
         "usage": {"prompt_tokens": 4, "total_tokens": 4},

test/components/embedders/test_sentence_transformers_document_embedder.py

@@ -3,7 +3,7 @@
 # SPDX-License-Identifier: Apache-2.0
 from unittest.mock import MagicMock, patch
 
-import numpy as np
+import random
 import pytest
 import torch
 
@@ -264,7 +264,9 @@ def test_warmup_doesnt_reload(self, mocked_factory):
     def test_run(self):
         embedder = SentenceTransformersDocumentEmbedder(model="model")
         embedder.embedding_backend = MagicMock()
-        embedder.embedding_backend.embed = lambda x, **kwargs: np.random.rand(len(x), 16).tolist()
+        embedder.embedding_backend.embed = lambda x, **kwargs: [
+            [random.random() for _ in range(16)] for _ in range(len(x))
+        ]
 
         documents = [Document(content=f"document number {i}") for i in range(5)]
 

test/components/embedders/test_sentence_transformers_text_embedder.py

@@ -4,7 +4,7 @@
 from unittest.mock import MagicMock, patch
 
 import torch
-import numpy as np
+import random
 import pytest
 
 from haystack.components.embedders.sentence_transformers_text_embedder import SentenceTransformersTextEmbedder
@@ -239,7 +239,9 @@ def test_warmup_doesnt_reload(self, mocked_factory):
     def test_run(self):
         embedder = SentenceTransformersTextEmbedder(model="model")
         embedder.embedding_backend = MagicMock()
-        embedder.embedding_backend.embed = lambda x, **kwargs: np.random.rand(len(x), 16).tolist()
+        embedder.embedding_backend.embed = lambda x, **kwargs: [
+            [random.random() for _ in range(16)] for _ in range(len(x))
+        ]
 
         text = "a nice text to embed"
 

test/components/evaluators/test_faithfulness_evaluator.py

@@ -5,7 +5,6 @@
 import math
 from typing import List
 
-import numpy as np
 import pytest
 
 from haystack import Pipeline

test/components/retrievers/test_in_memory_embedding_retriever.py

@@ -4,7 +4,6 @@
 from typing import Dict, Any
 
 import pytest
-import numpy as np
 
 from haystack import Pipeline, DeserializationError
 from haystack.document_stores.types import FilterPolicy
@@ -135,7 +134,7 @@ def test_valid_run(self):
 
         assert "documents" in result
         assert len(result["documents"]) == top_k
-        assert np.array_equal(result["documents"][0].embedding, [1.0, 1.0, 1.0, 1.0])
+        assert result["documents"][0].embedding == [1.0, 1.0, 1.0, 1.0]
 
     def test_invalid_run_wrong_store_type(self):
         SomeOtherDocumentStore = document_store_class("SomeOtherDocumentStore")
@@ -165,4 +164,4 @@ def test_run_with_pipeline(self):
         results_docs = result["retriever"]["documents"]
         assert results_docs
         assert len(results_docs) == top_k
-        assert np.array_equal(results_docs[0].embedding, [1.0, 1.0, 1.0, 1.0])
+        assert results_docs[0].embedding == [1.0, 1.0, 1.0, 1.0]