diff --git a/efemel/cli.py b/efemel/cli.py
index 385bcbe..4cbf191 100644
--- a/efemel/cli.py
+++ b/efemel/cli.py
@@ -1,7 +1,8 @@
+from concurrent.futures import ThreadPoolExecutor
+from concurrent.futures import as_completed
 import os
-import shutil
-from concurrent.futures import ThreadPoolExecutor, as_completed
 from pathlib import Path
+import shutil
 
 import click
 
diff --git a/efemel/hooks_manager.py b/efemel/hooks_manager.py
index 3006a6f..bd7adfd 100644
--- a/efemel/hooks_manager.py
+++ b/efemel/hooks_manager.py
@@ -1,8 +1,8 @@
 # hooks_manager.py
+from collections.abc import Callable
 import importlib.util
 import os
 import sys
-from collections.abc import Callable
 from typing import Any
 
 
diff --git a/efemel/pipeline.py b/efemel/pipeline.py
index cb22bd2..34492d5 100644
--- a/efemel/pipeline.py
+++ b/efemel/pipeline.py
@@ -7,14 +7,20 @@
 """
 
 from collections import deque
-from collections.abc import Callable, Generator, Iterable
-from concurrent.futures import FIRST_COMPLETED, ThreadPoolExecutor, wait
+from collections.abc import Callable
+from collections.abc import Generator
+from collections.abc import Iterable
+from concurrent.futures import FIRST_COMPLETED
+from concurrent.futures import ThreadPoolExecutor
+from concurrent.futures import wait
 from itertools import chain
-from typing import Any, Self, TypeVar
+from typing import Any
+from typing import Self
+from typing import TypeVar
+from typing import overload
 
 T = TypeVar("T")  # Type variable for the elements in the pipeline
 U = TypeVar("U")  # Type variable for transformed elements
-V = TypeVar("V")  # Type variable for additional transformations
 
 
 class Pipeline[T]:
@@ -49,6 +55,8 @@ def __init__(self, source: Iterable[T], chunk_size: int = 1000) -> None:
     else:
       self.generator = self._chunked(source, chunk_size)
 
+    self.chunk_size = chunk_size
+
   @staticmethod
   def _chunked(iterable: Iterable[T], size: int) -> Generator[list[T], None, None]:
     """Break an iterable into chunks of specified size."""
@@ -62,10 +70,11 @@ def _chunked(iterable: Iterable[T], size: int) -> Generator[list[T], None, None]
       yield chunk
 
   @classmethod
-  def _from_chunks(cls, chunks: Iterable[list[T]]) -> "Pipeline[T]":
+  def _from_chunks(cls, chunks: Iterable[list[T]], chunk_size: int = 1000) -> "Pipeline[T]":
     """Create a pipeline directly from an iterable of chunks."""
     p = cls([])
     p.generator = (chunk for chunk in chunks)
+    p.chunk_size = chunk_size
     return p
 
   @classmethod
@@ -111,7 +120,7 @@ def filter(self, predicate: Callable[[T], bool]) -> "Pipeline[T]":
     def filter_chunk(chunk: list[T]) -> list[T]:
       return [x for x in chunk if predicate(x)]
 
-    return Pipeline._from_chunks(filter_chunk(chunk) for chunk in self.generator)
+    return Pipeline._from_chunks((filter_chunk(chunk) for chunk in self.generator), self.chunk_size)
 
   def map(self, function: Callable[[T], U]) -> "Pipeline[U]":
     """Transform elements using a function, applied per chunk."""
@@ -119,7 +128,7 @@ def map(self, function: Callable[[T], U]) -> "Pipeline[U]":
     def map_chunk(chunk: list[T]) -> list[U]:
       return [function(x) for x in chunk]
 
-    return Pipeline._from_chunks(map_chunk(chunk) for chunk in self.generator)
+    return Pipeline._from_chunks((map_chunk(chunk) for chunk in self.generator), self.chunk_size)
 
   def reduce(self, function: Callable[[U, T], U], initial: U) -> "Pipeline[U]":
     """Reduce elements to a single value using the given function."""
@@ -137,7 +146,7 @@ def tap_chunk(chunk: list[T]) -> list[T]:
         function(item)
       return chunk
 
-    return Pipeline._from_chunks(tap_chunk(chunk) for chunk in self.generator)
+    return Pipeline._from_chunks((tap_chunk(chunk) for chunk in self.generator), self.chunk_size)
 
   def each(self, function: Callable[[T], Any]) -> None:
     """Apply function to each element (terminal operation)."""
@@ -162,13 +171,42 @@ def apply(self, *functions: Callable[[Self], "Pipeline[U]"]) -> "Pipeline[U]":
       result = function(result)
     return result
 
-  def flatten(self: "Pipeline[Iterable[U]]") -> "Pipeline[U]":
-    """Flatten pipeline of iterables into single pipeline."""
+  @overload
+  def flatten(self: "Pipeline[list[U]]") -> "Pipeline[U]": ...
+
+  @overload
+  def flatten(self: "Pipeline[tuple[U, ...]]") -> "Pipeline[U]": ...
+
+  @overload
+  def flatten(self: "Pipeline[set[U]]") -> "Pipeline[U]": ...
+
+  def flatten(
+    self: "Pipeline[list[U]] | Pipeline[tuple[U, ...]] | Pipeline[set[U]]",
+  ) -> "Pipeline[Any]":
+    """Flatten iterable chunks into a single pipeline of elements.
+
+    This method flattens each chunk of iterables and maintains the chunked
+    structure to avoid memory issues with large datasets. After flattening,
+    the data is re-chunked to maintain the original chunk_size.
+
+    Example:
+        [[1, 2], [3, 4]] -> [1, 2, 3, 4]
+        [(1, 2), (3, 4)] -> [1, 2, 3, 4]
+
+    Note:
+        We need to overload this method to handle different iterable types because
+        using Iterable[U] does not preserve the type information for the flattened elements.
+        It returns Pipeline[Any] instead of Pipeline[U], which is incorrect.
+    """
 
-    def flatten_chunk(chunk: list[Iterable[U]]) -> list[U]:
-      return [item for iterable in chunk for item in iterable]
+    def flatten_generator() -> Generator[Any, None, None]:
+      """Generator that yields individual flattened items."""
+      for chunk in self.generator:
+        for iterable in chunk:
+          yield from iterable
 
-    return Pipeline._from_chunks(flatten_chunk(chunk) for chunk in self.generator)
+    # Re-chunk the flattened stream to maintain consistent chunk size
+    return Pipeline._from_chunks(self._chunked(flatten_generator(), self.chunk_size), self.chunk_size)
 
   def concurrent(
     self,
@@ -256,7 +294,7 @@ def unordered_generator() -> Generator[list[U], None, None]:
               continue
 
     gen = ordered_generator() if ordered else unordered_generator()
-    return Pipeline._from_chunks(gen)
+    return Pipeline._from_chunks(gen, self.chunk_size)
 
   @classmethod
   def chain(cls, *pipelines: "Pipeline[T]") -> "Pipeline[T]":
@@ -274,4 +312,6 @@ def chain_generator():
       for pipeline in pipelines:
         yield from pipeline.generator
 
-    return cls._from_chunks(chain_generator())
+    # Use chunk_size from the first pipeline, or default if no pipelines
+    chunk_size = pipelines[0].chunk_size if pipelines else 1000
+    return cls._from_chunks(chain_generator(), chunk_size)
diff --git a/efemel/process.py b/efemel/process.py
index aa26d74..ce6b20f 100644
--- a/efemel/process.py
+++ b/efemel/process.py
@@ -1,7 +1,7 @@
-import importlib.util
-import sys
 from importlib.abc import MetaPathFinder
+import importlib.util
 from pathlib import Path
+import sys
 
 
 class EnvironmentModuleFinder(MetaPathFinder):
diff --git a/efemel/readers/local.py b/efemel/readers/local.py
index 3f29b4d..eef4cc7 100644
--- a/efemel/readers/local.py
+++ b/efemel/readers/local.py
@@ -1,5 +1,5 @@
-import os
 from glob import glob
+import os
 from pathlib import Path
 
 
diff --git a/pyproject.toml b/pyproject.toml
index 6a0017a..ee38b38 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -52,6 +52,15 @@ indent-style = "space"
 skip-magic-trailing-comma = false
 line-ending = "auto"
 
+[tool.ruff.lint.isort]
+force-single-line = true
+order-by-type = true
+known-first-party = ["efemel"]
+combine-as-imports = false
+force-sort-within-sections = true
+case-sensitive = false
+split-on-trailing-comma = false
+
 [tool.pytest.ini_options]
 testpaths = ["tests"]
 python_files = ["test_*.py", "*_test.py"]
diff --git a/tests/test_cli.py b/tests/test_cli.py
index f4e348e..909e2df 100644
--- a/tests/test_cli.py
+++ b/tests/test_cli.py
@@ -1,9 +1,9 @@
 import json
-import shutil
 from pathlib import Path
+import shutil
 
-import pytest
 from click.testing import CliRunner
+import pytest
 
 from efemel.cli import cli
 
diff --git a/tests/test_pipeline.py b/tests/test_pipeline.py
index 7e85607..cb6baf9 100644
--- a/tests/test_pipeline.py
+++ b/tests/test_pipeline.py
@@ -402,13 +402,6 @@ def test_flatten_basic(self):
     result = pipeline.flatten()
     assert result.to_list() == [1, 2, 3, 4, 5]
 
-  def test_flatten_with_strings(self):
-    """Test flatten with string iterables."""
-    pipeline = Pipeline(["hello", "world"])
-
-    result = pipeline.flatten()
-    assert result.to_list() == ["h", "e", "l", "l", "o", "w", "o", "r", "l", "d"]
-
   def test_flatten_empty_lists(self):
     """Test flatten with empty lists."""
     pipeline = Pipeline([[], [1, 2], [], [3]])
@@ -423,6 +416,72 @@ def test_flatten_nested_tuples(self):
     result = pipeline.flatten()
     assert result.to_list() == [1, 2, 3, 4, 5, 6]
 
+  def test_flatten_chunked_processing(self):
+    """Test that flatten maintains chunked processing and doesn't consume entire pipeline."""
+
+    # Create a large dataset that would cause OOM if consumed all at once
+    def generate_nested_data():
+      for i in range(1000):
+        yield [i * 2, i * 2 + 1]
+
+    # Use small chunk size to verify chunked processing
+    pipeline = Pipeline(generate_nested_data(), chunk_size=10)
+    result = pipeline.flatten()
+
+    # Verify first few elements without consuming the entire pipeline
+    first_10 = []
+    count = 0
+    for item in result:
+      first_10.append(item)
+      count += 1
+      if count >= 10:
+        break
+
+    assert first_10 == [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
+
+  def test_flatten_preserves_chunk_structure(self):
+    """Test that flatten preserves chunked structure."""
+    # Create pipeline with known chunk structure
+    data = [[1, 2], [3, 4], [5, 6], [7, 8]]
+    pipeline = Pipeline(data, chunk_size=2)  # 2 sublists per chunk
+
+    result = pipeline.flatten()
+
+    # Verify the result is correct
+    assert result.to_list() == [1, 2, 3, 4, 5, 6, 7, 8]
+
+  def test_flatten_maintains_chunk_size(self):
+    """Test that flatten maintains the original chunk_size setting."""
+    # Create test data with varying sizes of sublists
+    data = [[1, 2], [3, 4, 5], [6], [7, 8, 9, 10]]
+    chunk_size = 3
+    pipeline = Pipeline(data, chunk_size=chunk_size)
+
+    # Flatten the pipeline
+    flattened = pipeline.flatten()
+
+    # Verify chunk_size is preserved
+    assert flattened.chunk_size == chunk_size
+
+    # Collect chunks to verify structure
+    chunks = []
+    for chunk in flattened.generator:
+      chunks.append(chunk)
+
+    # Verify all chunks except the last have the expected size
+    for i, chunk in enumerate(chunks[:-1]):
+      assert len(chunk) == chunk_size, f"Chunk {i} has size {len(chunk)}, expected {chunk_size}"
+
+    # Last chunk can be smaller than chunk_size
+    if chunks:
+      assert len(chunks[-1]) <= chunk_size, f"Last chunk has size {len(chunks[-1])}, should be <= {chunk_size}"
+
+    # Verify the final result is correct
+    result = []
+    for chunk in chunks:
+      result.extend(chunk)
+    assert result == [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
+
 
 class TestPipelineEdgeCases:
   """Test Pipeline edge cases and error conditions."""