From be4c34c9acddb631eb145822b1b2ab77a7a3bd21 Mon Sep 17 00:00:00 2001
From: Izzet Yildirim <yildirim2@llnl.gov>
Date: Thu, 4 Dec 2025 14:14:17 +0300
Subject: [PATCH] chore: Remove custom Dask aggregation utilities by deleting
 `dask_agg.py`.

---
 dfanalyzer/utils/dask_agg.py | 71 ------------------------------------
 1 file changed, 71 deletions(-)
 delete mode 100644 dfanalyzer/utils/dask_agg.py

diff --git a/dfanalyzer/utils/dask_agg.py b/dfanalyzer/utils/dask_agg.py
deleted file mode 100644
index 34601e8..0000000
--- a/dfanalyzer/utils/dask_agg.py
+++ /dev/null
@@ -1,71 +0,0 @@
-import dask.dataframe as dd
-import numpy as np
-import portion as P
-from betterset import BetterSet as S
-
-
-def nunique():
-    return dd.Aggregation(
-        name="nunique",
-        chunk=lambda s: s.apply(lambda x: list(set(x))),
-        agg=lambda s0: s0.obj.groupby(level=list(range(s0.obj.index.nlevels))).sum(),
-        finalize=lambda s1: s1.apply(lambda final: len(set(final))),
-    )
-
-
-def quantile_stats(min, max):
-    def quantile_stats_finalize(values_list):
-        if len(values_list) == 0:
-            return [np.nan, np.nan, np.nan]
-        values_array = np.array(values_list)
-        q_min, q_max = np.quantile(values_array, [min, max])
-        filtered_mask = (values_array >= q_min) & (values_array <= q_max)
-        filtered_values = values_array[filtered_mask]
-        if len(filtered_values) == 0:
-            return [np.nan, np.nan, np.nan]
-        return [np.mean(filtered_values), np.std(filtered_values), len(filtered_values)]
-
-    return dd.Aggregation(
-        f"q{min * 100:.0f}_q{max * 100:.0f}_stats",
-        lambda s: s.apply(lambda x: x.replace(0, np.nan).dropna().tolist()),
-        lambda s0: s0.obj.groupby(level=0).sum(),
-        lambda s1: s1.apply(quantile_stats_finalize),
-    )
-
-
-def unique_set():
-    return dd.Aggregation(
-        'unique',
-        lambda s: s.apply(lambda x: S(x.dropna().unique().tolist())),
-        lambda s0: s0.apply(S.flatten),
-    )
-
-
-def unique_set_flatten():
-    return dd.Aggregation(
-        'unique',
-        lambda s: s.apply(S.flatten),
-        lambda s0: s0.agg(S.flatten),
-    )
-
-
-def union_portions():
-    def union_s(s):
-        emp = P.empty()
-        for x in s:
-            emp = emp | x
-        return emp
-
-    def fin(s):
-        val = 0.0
-        for i in s:
-            if not i.is_empty():
-                val += i.upper - i.lower
-        return val
-
-    return dd.Aggregation(
-        'portion',
-        union_s,
-        union_s,
-        fin,
-    )