Merge pull request #10 from SaridakisStamatisChristos/codex/fix-coverage-report-upload-issue

Fix CI pipeline to publish coverage

Author: SaridakisStamatisChristos

.github/workflows/ci.yml

@@ -12,8 +12,12 @@ jobs:
       - uses: actions/setup-python@v5
         with:
           python-version: "3.11"
-      - run: python -m pip install --upgrade pip pytest
-      - run: python -m pytest -q
+      - name: Install dependencies
+        run: |
+          python -m pip install --upgrade pip
+          python -m pip install -r kll_sketch/requirements-test.txt
+      - name: Run tests with coverage
+        run: python -m pytest --cov=kll_sketch --cov-report=term-missing --cov-report=xml kll_sketch/tests
       - name: Upload coverage reports to Codecov
         uses: codecov/codecov-action@v5
         with:

kll_sketch/kll_sketch.py

@@ -12,7 +12,7 @@
 import struct
 import random
 import heapq
-from bisect import bisect_left
+from bisect import bisect_left, bisect_right
 from typing import Iterable, List, Tuple, Optional
 
 
@@ -205,6 +205,13 @@ def _capacity_exceeded(self) -> bool:
         return self._total_items() > int(self._k * self._SOFT_CAP_FACTOR)
 
     def _level_capacity(self, level: int) -> int:
+        # Preserve exactness while the total population fits inside ``k`` by
+        # deferring compaction.  This mirrors the streaming ingestion behaviour
+        # where no compression is triggered before exceeding ``k`` items, which
+        # keeps small merges numerically identical to simple extension.
+        if self._n <= self._k:
+            return self._k
+
         # Geometric schedule; sum over levels ≈ O(k)
         base = max(self._LEVEL_BASE_MIN, self._k // 8)
         return base * (1 << max(0, level))
@@ -399,7 +406,7 @@ def _batched_quantiles(self, qs: List[float]) -> List[float]:
         search_lo = 0
         for idx, q in ordered:
             target = q * (self._n - 1)
-            pos = bisect_left(prefix, target - 1e-12, lo=search_lo)
+            pos = bisect_right(prefix, target, lo=search_lo)
             if pos >= len(vals):
                 pos = len(vals) - 1
             out[idx] = vals[pos]

kll_sketch/pyproject.toml

@@ -33,7 +33,7 @@ test = [
 ]
 
 [tool.pytest.ini_options]
-addopts = "--strict-config --strict-markers --cov=kll_sketch --cov-report=term-missing"
+addopts = "--strict-config --strict-markers --cov=kll_sketch --cov-report=term-missing --cov-report=xml"
 testpaths = ["kll_sketch/tests"]
 filterwarnings = [
   "error",