[OpenSearch] Add chunks specific to an app_id if present (#765)
This commit is contained in:
@@ -17,14 +17,15 @@ class TestImageChunker(unittest.TestCase):
|
||||
chunker.set_data_type(DataType.IMAGES)
|
||||
|
||||
image_path = "./tmp/image.jpeg"
|
||||
result = chunker.create_chunks(MockLoader(), image_path)
|
||||
app_id = "app1"
|
||||
result = chunker.create_chunks(MockLoader(), image_path, app_id=app_id)
|
||||
|
||||
expected_chunks = {
|
||||
"doc_id": "123",
|
||||
"doc_id": f"{app_id}--123",
|
||||
"documents": [image_path],
|
||||
"embeddings": ["embedding"],
|
||||
"ids": ["140bedbf9c3f6d56a9846d2ba7088798683f4da0c248231336e6a05679e4fdfe"],
|
||||
"metadatas": [{"data_type": "images", "doc_id": "123", "url": "none"}],
|
||||
"metadatas": [{"data_type": "images", "doc_id": f"{app_id}--123", "url": "none"}],
|
||||
}
|
||||
self.assertEqual(expected_chunks, result)
|
||||
|
||||
@@ -37,14 +38,15 @@ class TestImageChunker(unittest.TestCase):
|
||||
chunker.set_data_type(DataType.IMAGES)
|
||||
|
||||
image_path = "./tmp/image.jpeg"
|
||||
result = chunker.create_chunks(MockLoader(), image_path)
|
||||
app_id = "app1"
|
||||
result = chunker.create_chunks(MockLoader(), image_path, app_id=app_id)
|
||||
|
||||
expected_chunks = {
|
||||
"doc_id": "123",
|
||||
"doc_id": f"{app_id}--123",
|
||||
"documents": [image_path],
|
||||
"embeddings": ["embedding"],
|
||||
"ids": ["140bedbf9c3f6d56a9846d2ba7088798683f4da0c248231336e6a05679e4fdfe"],
|
||||
"metadatas": [{"data_type": "images", "doc_id": "123", "url": "none"}],
|
||||
"metadatas": [{"data_type": "images", "doc_id": f"{app_id}--123", "url": "none"}],
|
||||
}
|
||||
self.assertEqual(expected_chunks, result)
|
||||
|
||||
|
||||
@@ -1,31 +1,35 @@
|
||||
# ruff: noqa: E501
|
||||
|
||||
import unittest
|
||||
|
||||
from embedchain.chunkers.text import TextChunker
|
||||
from embedchain.config import ChunkerConfig
|
||||
from embedchain.models.data_type import DataType
|
||||
|
||||
|
||||
class TestTextChunker(unittest.TestCase):
|
||||
def test_chunks(self):
|
||||
class TestTextChunker:
|
||||
def test_chunks_without_app_id(self):
|
||||
"""
|
||||
Test the chunks generated by TextChunker.
|
||||
# TODO: Not a very precise test.
|
||||
"""
|
||||
chunker_config = ChunkerConfig(chunk_size=10, chunk_overlap=0, length_function=len)
|
||||
chunker = TextChunker(config=chunker_config)
|
||||
text = "Lorem ipsum dolor sit amet, consectetur adipiscing elit."
|
||||
# Data type must be set manually in the test
|
||||
chunker.set_data_type(DataType.TEXT)
|
||||
|
||||
result = chunker.create_chunks(MockLoader(), text)
|
||||
|
||||
documents = result["documents"]
|
||||
assert len(documents) > 5
|
||||
|
||||
self.assertGreaterEqual(len(documents), 5)
|
||||
|
||||
# Additional test cases can be added to cover different scenarios
|
||||
def test_chunks_with_app_id(self):
|
||||
"""
|
||||
Test the chunks generated by TextChunker with app_id
|
||||
"""
|
||||
chunker_config = ChunkerConfig(chunk_size=10, chunk_overlap=0, length_function=len)
|
||||
chunker = TextChunker(config=chunker_config)
|
||||
text = "Lorem ipsum dolor sit amet, consectetur adipiscing elit."
|
||||
chunker.set_data_type(DataType.TEXT)
|
||||
result = chunker.create_chunks(MockLoader(), text)
|
||||
documents = result["documents"]
|
||||
assert len(documents) > 5
|
||||
|
||||
def test_big_chunksize(self):
|
||||
"""
|
||||
@@ -36,12 +40,9 @@ class TestTextChunker(unittest.TestCase):
|
||||
text = "Lorem ipsum dolor sit amet, consectetur adipiscing elit."
|
||||
# Data type must be set manually in the test
|
||||
chunker.set_data_type(DataType.TEXT)
|
||||
|
||||
result = chunker.create_chunks(MockLoader(), text)
|
||||
|
||||
documents = result["documents"]
|
||||
|
||||
self.assertEqual(len(documents), 1)
|
||||
assert len(documents) == 1
|
||||
|
||||
def test_small_chunksize(self):
|
||||
"""
|
||||
@@ -53,14 +54,9 @@ class TestTextChunker(unittest.TestCase):
|
||||
text = """0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ!"#$%&\'()*+,-./:;<=>?@[\\]^_`{|}~ \t\n\r\x0b\x0c"""
|
||||
# Data type must be set manually in the test
|
||||
chunker.set_data_type(DataType.TEXT)
|
||||
|
||||
result = chunker.create_chunks(MockLoader(), text)
|
||||
|
||||
documents = result["documents"]
|
||||
|
||||
print(documents)
|
||||
|
||||
self.assertEqual(len(documents), len(text))
|
||||
assert len(documents) == len(text)
|
||||
|
||||
def test_word_count(self):
|
||||
chunker_config = ChunkerConfig(chunk_size=1, chunk_overlap=0, length_function=len)
|
||||
@@ -69,7 +65,7 @@ class TestTextChunker(unittest.TestCase):
|
||||
|
||||
document = ["ab cd", "ef gh"]
|
||||
result = chunker.get_word_count(document)
|
||||
self.assertEqual(result, 4)
|
||||
assert result == 4
|
||||
|
||||
|
||||
class MockLoader:
|
||||
|
||||
Reference in New Issue
Block a user