chore: Remove sortedcollections and sortedcontainers packages

pyproject.toml

@@ -42,8 +42,6 @@ dependencies = [
     "pydantic-settings>=2.2.0,!=2.7.0,!=2.7.1,!=2.8.0",
     "pydantic>=2.8.0,!=2.10.0,!=2.10.1,!=2.10.2",
     "pyee>=9.0.0",
-    "sortedcollections>=2.1.0",
-    "sortedcontainers>=2.4.0",
     "tldextract>=5.1.0",
     "typing-extensions>=4.1.0",
     "yarl>=1.18.0",
@@ -103,7 +101,6 @@ dev = [
     "pytest~=8.4.0",
     "ruff~=0.12.0",
     "setuptools", # setuptools are used by pytest, but not explicitly required
-    "sortedcontainers-stubs~=2.4.0",
     "types-beautifulsoup4~=4.12.0.20240229",
     "types-cachetools~=6.1.0.20250717",
     "types-colorama~=0.4.15.20240106",
@@ -247,7 +244,6 @@ module = [
     "litestar",                     # Example code shows deploy on Google Cloud Run.
     "loguru",                       # Example code shows integration of loguru and crawlee for JSON logging.
     "sklearn.linear_model",         # Untyped and stubs not available
-    "sortedcollections",            # Untyped and stubs not available
     "cookiecutter.*",               # Untyped and stubs not available
     "inquirer.*",                   # Untyped and stubs not available
     "warcio.*",                     # Example code shows WARC files creation.

src/crawlee/_autoscaling/snapshotter.py

@@ -2,13 +2,11 @@
 
 from __future__ import annotations
 
+import bisect
 from datetime import datetime, timedelta, timezone
 from logging import getLogger
-from operator import attrgetter
 from typing import TYPE_CHECKING, TypeVar, cast
 
-from sortedcontainers import SortedList
-
 from crawlee import service_locator
 from crawlee._autoscaling._types import ClientSnapshot, CpuSnapshot, EventLoopSnapshot, MemorySnapshot, Snapshot
 from crawlee._utils.byte_size import ByteSize
@@ -25,7 +23,15 @@
 
 logger = getLogger(__name__)
 
-T = TypeVar('T')
+T = TypeVar('T', bound=Snapshot)
+
+
+class SortedSnapshotList(list[T]):
+    """A list that maintains sorted order by `created_at` attribute for snapshot objects."""
+
+    def add(self, item: T) -> None:
+        """Add an item to the list maintaining sorted order by `created_at` using binary search."""
+        bisect.insort(self, item, key=lambda item: item.created_at)
 
 
 @docs_group('Autoscaling')
@@ -127,8 +133,14 @@ def from_config(cls, config: Configuration | None = None) -> Snapshotter:
         )
 
     @staticmethod
-    def _get_sorted_list_by_created_at(input_list: list[T]) -> SortedList[T]:
-        return SortedList(input_list, key=attrgetter('created_at'))
+    def _get_sorted_list_by_created_at(input_list: list[T]) -> SortedSnapshotList[T]:
+        """Create a sorted list from the input list.
+
+        Returns a custom list that maintains sorted order by created_at when items are added.
+        """
+        result = SortedSnapshotList[T]()
+        result.extend(input_list)
+        return result
 
     @property
     def active(self) -> bool:

tests/unit/_autoscaling/test_snapshotter.py

@@ -10,6 +10,7 @@
 from crawlee import service_locator
 from crawlee._autoscaling import Snapshotter
 from crawlee._autoscaling._types import ClientSnapshot, CpuSnapshot, EventLoopSnapshot, Snapshot
+from crawlee._autoscaling.snapshotter import SortedSnapshotList
 from crawlee._utils.byte_size import ByteSize
 from crawlee._utils.system import CpuInfo, MemoryInfo
 from crawlee.configuration import Configuration
@@ -299,3 +300,34 @@ def create_event_data(creation_time: datetime) -> EventSystemInfoData:
         assert cpu_samples[0].created_at == time_old
         assert memory_samples[1].created_at == time_new
         assert cpu_samples[1].created_at == time_new
+
+
+def test_sorted_snapshot_list_add_maintains_order() -> None:
+    """Test that SortedSnapshotList.add method maintains sorted order by created_at with multiple items."""
+    sorted_list = SortedSnapshotList[CpuSnapshot]()
+
+    # Create snapshots with different timestamps (more items to test binary search better)
+    now = datetime.now(timezone.utc)
+    snapshots = [
+        CpuSnapshot(used_ratio=0.1, max_used_ratio=0.95, created_at=now - timedelta(seconds=50)),  # oldest
+        CpuSnapshot(used_ratio=0.2, max_used_ratio=0.95, created_at=now - timedelta(seconds=40)),
+        CpuSnapshot(used_ratio=0.3, max_used_ratio=0.95, created_at=now - timedelta(seconds=30)),
+        CpuSnapshot(used_ratio=0.4, max_used_ratio=0.95, created_at=now - timedelta(seconds=20)),
+        CpuSnapshot(used_ratio=0.5, max_used_ratio=0.95, created_at=now - timedelta(seconds=10)),
+        CpuSnapshot(used_ratio=0.6, max_used_ratio=0.95, created_at=now - timedelta(seconds=5)),
+        CpuSnapshot(used_ratio=0.7, max_used_ratio=0.95, created_at=now),  # newest
+    ]
+
+    # Add snapshots in random order to test binary search insertion
+    add_order = [3, 0, 5, 1, 6, 2, 4]  # indices in random order
+    for i in add_order:
+        sorted_list.add(snapshots[i])
+
+    # Verify the list is sorted by created_at (should be in original order)
+    assert len(sorted_list) == 7
+    for i, snapshot in enumerate(sorted_list):
+        assert snapshot == snapshots[i], f'Item at index {i} is not correctly sorted'
+        if i > 0:
+            prev_time = sorted_list[i - 1].created_at
+            curr_time = snapshot.created_at
+            assert prev_time <= curr_time, f'Items at indices {i - 1} and {i} are not in chronological order'