toolbox-utils 5.4.0__tar.gz → 5.5.5__tar.gz

{toolbox_utils-5.4.0 → toolbox_utils-5.5.5}/CHANGELOG.md

@@ -1,3 +1,39 @@
+## v5.5.5 (2026-06-11)
+
+### Fix
+
+- support numpy 1.* by casting ndiff to int64
+
+## v5.5.4 (2026-06-10)
+
+### Fix
+
+- support for pandas 3
+
+## v5.5.3 (2026-06-10)
+
+### Fix
+
+- added interval check for monthly, quarterly, and yearly frequencies in asbestfreq
+
+## v5.5.2 (2026-06-10)
+
+### Fix
+
+- rearranged frequency guessing in asbestfreq since an sub-daily time series all on XXXX-01-01 was thought of as a YS frequency instead of hourly
+
+## v5.5.1 (2026-06-10)
+
+### Fix
+
+- made asbestfreq more robust
+
+## v5.5.0 (2026-06-09)
+
+### Feat
+
+- add "rename_index" keyword "common_kwds" function
+
 ## v5.4.0 (2026-06-09)
 
 ### Feat

{toolbox_utils-5.4.0 → toolbox_utils-5.5.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: toolbox_utils
-Version: 5.4.0
+Version: 5.5.5
 Summary: Python toolbox for common utilities
 Author-email: Tim Cera <tim@cerazone.net>
 License: BSD-3-Clause

toolbox_utils-5.5.5/VERSION

@@ -0,0 +1 @@
+5.5.5

{toolbox_utils-5.4.0 → toolbox_utils-5.5.5}/pyproject.toml

@@ -61,7 +61,7 @@ github = "https://github.com/timcera/toolbox_utils"
 name = "cz_conventional_commits"
 tag_format = "v$version"
 update_changelog_on_bump = true
-version = "5.4.0"
+version = "5.5.5"
 version_files = ["VERSION"]
 
 [tool.isort]

{toolbox_utils-5.4.0 → toolbox_utils-5.5.5}/src/toolbox_utils/readers/wdm.py

@@ -143,10 +143,12 @@ def wdm_extract(wdmfile, *idsn):
         tgroup = dattr["TGROUP"]
         tstep = dattr["TSSTEP"]
         tcode = dattr["TCODE"]
-        cindex = pd.date_range(start=start, periods=len(records) + 1, freq=freq[tgroup])
+        cindex = pd.date_range(
+            start=start, periods=len(records) + 1, freq=freq[tgroup]
+        ).astype("datetime64[ns]")
         tindex = pd.date_range(
             start=start, end=cindex[-1], freq=str(tstep) + freq[tcode]
-        )
+        ).astype("datetime64[ns]")
         counts = np.diff(np.searchsorted(tindex, cindex))
 
         # Get timeseries data

{toolbox_utils-5.4.0 → toolbox_utils-5.5.5}/src/toolbox_utils/tsutils.py

@@ -1402,6 +1402,7 @@ def common_kwds(
     names: Optional[Union[List[str], str]] = None,
     usecols: Optional[List[Union[int, str]]] = None,
     por: bool = False,
+    rename_index: bool = True,
 ):
     """
     Process all common_kwds across sub-commands into single function.
@@ -1476,10 +1477,18 @@ def common_kwds(
 
     ntsd = _date_slice(ntsd, start_date=start_date, end_date=end_date, por=por)
 
-    if (
-        not ntsd.index.name or "Datetime" not in ntsd.index.name
-    ) and ntsd.index.inferred_type == "datetime64":
-        ntsd.index.name = "Datetime"
+    if rename_index and ((not ntsd.index.name) or ("Datetime" not in ntsd.index.name)):
+        ntsd.index.name = "UniqueID"
+
+        if isinstance(ntsd.index, pd.DatetimeIndex):
+            timezone = ntsd.index.tz
+            if timezone:
+                ntsd.index.name = f"Datetime:{timezone}"
+            else:
+                ntsd.index.name = "Datetime"
+
+        if isinstance(ntsd.index, pd.PeriodIndex):
+            ntsd.index.name = "Period"
 
     if dropna in ("any", "all"):
         ntsd = ntsd.dropna(axis="index", how=dropna)
@@ -1721,19 +1730,21 @@ def asbestfreq(data: DataFrame, force_freq: Optional[str] = None) -> DataFrame:
     if force_freq:
         return _replace_nan_with_na(data, freq=force_freq)
 
+    ndata = data.dropna(how="all")
+
     # Since pandas doesn't set data.index.freq and data.index.freqstr when
     # using .asfreq, this function returns that PANDAS time offset alias code
     # also.  Not ideal at all.
     #
     # This gets most of the frequencies...
-    if data.index.inferred_freq:
+    if ndata.index.inferred_freq:
         with suppress(ValueError):
             return _replace_nan_with_na(data, freq=data.index.inferred_freq)
 
     # pd.infer_freq would fail if given a large dataset
-    slic = slice(None, 999) if len(data.index) > 1000 else slice(None, None)
+    slic = slice(None, 999) if len(ndata.index) > 1000 else slice(None, None)
     try:
-        infer_freq = pd.infer_freq(data.index[slic])
+        infer_freq = pd.infer_freq(ndata.index[slic])
     except ValueError:
         infer_freq = None
     if infer_freq:
@@ -1741,8 +1752,9 @@ def asbestfreq(data: DataFrame, force_freq: Optional[str] = None) -> DataFrame:
 
     data.index = data.index.astype("datetime64[ns]")
     ndiff = (
-        data.index.astype("int64").values[1:] - data.index.astype("int64").values[:-1]
-    )
+        ndata.index.astype("datetime64[ns]").values[1:]
+        - ndata.index.astype("datetime64[ns]").values[:-1]
+    ).astype("int64")
 
     if np.any(ndiff <= 0):
         raise ValueError(
@@ -1761,57 +1773,53 @@ def asbestfreq(data: DataFrame, force_freq: Optional[str] = None) -> DataFrame:
     # Use the minimum of the intervals to test a new interval.
     # Should work for fixed intervals.
     ndiff = sorted(set(ndiff))
-    ngcd = (
-        ndiff[0].astype("int64")
-        if len(ndiff) == 1
-        else reduce(np.gcd, [i.astype("int64") for i in ndiff])
-    )
-    if ngcd < 1000:
-        infer_freq = f"{ngcd}{pandas_offset_by_version('ns')}"
-    elif ngcd < 1000000:
-        infer_freq = f"{ngcd // 1000}{pandas_offset_by_version('us')}"
-    elif ngcd < 1000000000:
-        infer_freq = f"{ngcd // 1000000}{pandas_offset_by_version('ms')}"
-    elif ngcd < 60000000000:
-        infer_freq = f"{ngcd // 1000000000}{pandas_offset_by_version('s')}"
-    elif ngcd < 3600000000000:
-        infer_freq = f"{ngcd // 60000000000}{pandas_offset_by_version('min')}"
-    elif ngcd < 86400000000000:
-        infer_freq = f"{ngcd // 3600000000000}{pandas_offset_by_version('h')}"
-    elif ngcd < 604800000000000:
-        infer_freq = f"{ngcd // 86400000000000}D"
-    elif ngcd < 2419200000000000:
-        infer_freq = f"{ngcd // 604800000000000}W"
-        if np.all(data.index.dayofweek == data.index[0].dayofweek):
-            infer_freq = f"{infer_freq}-{_WEEKLIES[data.index[0].dayofweek]}"
-        else:
-            infer_freq = "D"
-    if infer_freq:
-        return _replace_nan_with_na(data, freq=infer_freq)
-
-    # At this point pd.infer_freq failed probably because of missing values.
-    # The following algorithm would not capture things like BQ, BQS
-    # ...etc.
-    if np.all(data.index.is_month_start):
-        if np.all(data.index.month == data.index[0].month):
-            # Actually yearly with different start
-            infer_freq = f"YE-{_ANNUALS[data.index[0].month - 1]}"
-        else:
-            infer_freq = "MS"
-    elif np.all(data.index.is_month_end):
+    ngcd = ndiff[0] if len(ndiff) == 1 else reduce(np.gcd, ndiff)
+    ndiff = np.array(ndiff)
+    # ngcd is the greatest common divisor of the intervals in nanoseconds.
+    infer_freq = f"{ngcd}{pandas_offset_by_version('ns')}"
+    if data.index.is_month_end.all() and all(ndiff >= 2_419_200_000_000_000):
+        infer_freq = pandas_offset_by_version("ME")
         if np.all(data.index.month == data.index[0].month):
             # Actually yearly with different ends
-            infer_freq = f"YE-{_ANNUALS[data.index[0].month]}"
-        else:
-            infer_freq = "ME"
-    elif np.all(data.index.is_quarter_end):
+            infer_freq = (
+                f"{pandas_offset_by_version('YE')}-{_ANNUALS[data.index[0].month]}"
+            )
+    elif data.index.is_month_start.all() and all(ndiff >= 2_419_200_000_000_000):
+        infer_freq = pandas_offset_by_version("MS")
+        if np.all(data.index.month == data.index[0].month):
+            # Actually yearly with different start
+            infer_freq = (
+                f"{pandas_offset_by_version('YS')}-{_ANNUALS[data.index[0].month - 1]}"
+            )
+    elif data.index.is_quarter_end.all() and all(ndiff >= 7_257_600_000_000_000):
         infer_freq = pandas_offset_by_version("QE")
-    elif np.all(data.index.is_quarter_start):
+    elif data.index.is_quarter_start.all() and all(ndiff >= 7_257_600_000_000_000):
         infer_freq = pandas_offset_by_version("QS")
-    elif np.all(data.index.is_year_end):
+    elif data.index.is_year_end.all() and all(ndiff >= 31_536_000_000_000_000):
         infer_freq = pandas_offset_by_version("YE")
-    elif np.all(data.index.is_year_start):
+    elif data.index.is_year_start.all() and all(ndiff >= 31_536_000_000_000_000):
         infer_freq = pandas_offset_by_version("YS")
+    elif ngcd < 1_000:
+        infer_freq = f"{ngcd}{pandas_offset_by_version('ns')}"
+    elif ngcd % 604_800_000_000_000 == 0:
+        infer_freq = f"{ngcd // 604_800_000_000_000}W"
+        if np.all(data.index.dayofweek == data.index[0].dayofweek):
+            infer_freq = f"{infer_freq}-{_WEEKLIES[data.index[0].dayofweek]}"
+    elif ngcd % 86_400_000_000_000 == 0:
+        infer_freq = f"{ngcd // 86_400_000_000_000}D"
+    elif ngcd % 3_600_000_000_000 == 0:
+        infer_freq = f"{ngcd // 3_600_000_000_000}{pandas_offset_by_version('h')}"
+    elif ngcd % 60_000_000_000 == 0:
+        infer_freq = f"{ngcd // 60_000_000_000}{pandas_offset_by_version('min')}"
+    elif ngcd % 1_000_000_000 == 0:
+        infer_freq = f"{ngcd // 1_000_000_000}{pandas_offset_by_version('s')}"
+    elif ngcd % 1_000_000 == 0:
+        infer_freq = f"{ngcd // 1_000_000}{pandas_offset_by_version('ms')}"
+    elif ngcd % 1_000 == 0:
+        infer_freq = f"{ngcd // 1_000}{pandas_offset_by_version('us')}"
+
+    if infer_freq:
+        return _replace_nan_with_na(data, freq=infer_freq)
 
     return _replace_nan_with_na(data, freq=infer_freq)
 

{toolbox_utils-5.4.0 → toolbox_utils-5.5.5}/src/toolbox_utils/utils.py

@@ -41,7 +41,7 @@ def pandas_period_by_version(new_offset: str) -> str:
     new_to_old_period = {}
     major, minor = pd.__version__.split(".")[:2]
     if (int(major) + int(minor) / 10) < 2.2:
-        new_to_old_freq = {
+        new_to_old_period = {
             "Y": "A",
             "h": "H",
             "min": "T",

{toolbox_utils-5.4.0 → toolbox_utils-5.5.5}/src/toolbox_utils.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: toolbox_utils
-Version: 5.4.0
+Version: 5.5.5
 Summary: Python toolbox for common utilities
 Author-email: Tim Cera <tim@cerazone.net>
 License: BSD-3-Clause

{toolbox_utils-5.4.0 → toolbox_utils-5.5.5}/tests/test_read.py

@@ -11,7 +11,6 @@ class TestRead(TestCase):
         dr = pandas.date_range("2000-01-01", periods=2, freq="D")
 
         ts = pandas.Series([4.5, 4.6], index=dr)
-        print(ts.index.dtype)
 
         self.read_direct = pandas.DataFrame(ts, columns=["Value"])
         self.read_direct.index.name = "Datetime"

{toolbox_utils-5.4.0 → toolbox_utils-5.5.5}/tests/test_wdm.py

@@ -21,7 +21,7 @@ class TestWDM(TestCase):
         ret2 = tsutils.asbestfreq(
             pd.read_csv("tests/data_wdm_1.csv", index_col=0, parse_dates=True)
         )
-        assert_frame_equal(ret1, ret2, check_dtype=False)
+        assert_frame_equal(ret1, ret2, check_dtype=False, check_index_type=False)
 
     def test_extract2(self):
         ret1 = tsutils.common_kwds("tests/data.wdm,2")
@@ -34,14 +34,13 @@ class TestWDM(TestCase):
     def test_extract_range(self):
         ret1 = tsutils.common_kwds("tests/data.wdm,1:2")
         ret1.columns = ["data.wdm_1", "data.wdm_2"]
+
+        ret2 = pd.read_csv("tests/data_wdm_1.csv", index_col=0, parse_dates=True)
         ret2 = tsutils.asbestfreq(
-            pd.read_csv("tests/data_wdm_1.csv", index_col=0, parse_dates=True)
-        )
-        ret2 = ret2.join(
-            tsutils.asbestfreq(
-                pd.read_csv("tests/data_wdm_2.csv", index_col=0, parse_dates=True)
-            ),
-            how="outer",
+            ret2.join(
+                pd.read_csv("tests/data_wdm_2.csv", index_col=0, parse_dates=True),
+                how="outer",
+            )
         )
         try:
             ret1 = ret1.replace(np.nan, pd.NA)
@@ -51,20 +50,20 @@ class TestWDM(TestCase):
             ret2 = ret2.replace(np.nan, pd.NA)
         except RecursionError:
             pass
-        assert_frame_equal(ret1, ret2, check_dtype=False, rtol=1e-7)
+        assert_frame_equal(
+            ret1, ret2, check_dtype=False, rtol=1e-7, check_index_type=False
+        )
 
     def test_extract_range_plus(self):
         ret1 = tsutils.common_kwds("tests/data.wdm,1+2")
         ret1.columns = ["data.wdm_1", "data.wdm_2"]
         ret1[ret1.isnull()] = np.nan
+        ret2 = pd.read_csv("tests/data_wdm_1.csv", index_col=0, parse_dates=True)
         ret2 = tsutils.asbestfreq(
-            pd.read_csv("tests/data_wdm_1.csv", index_col=0, parse_dates=True)
-        )
-        ret2 = ret2.join(
-            tsutils.asbestfreq(
-                pd.read_csv("tests/data_wdm_2.csv", index_col=0, parse_dates=True)
-            ),
-            how="outer",
+            ret2.join(
+                pd.read_csv("tests/data_wdm_2.csv", index_col=0, parse_dates=True),
+                how="outer",
+            )
         )
         try:
             ret1 = ret1.replace(np.nan, pd.NA)
@@ -74,4 +73,6 @@ class TestWDM(TestCase):
             ret2 = ret2.replace(np.nan, pd.NA)
         except RecursionError:
             pass
-        assert_frame_equal(ret1, ret2, check_dtype=False, rtol=1e-7)
+        assert_frame_equal(
+            ret1, ret2, check_dtype=False, rtol=1e-7, check_index_type=False
+        )

toolbox_utils-5.4.0/VERSION

@@ -1 +0,0 @@
-5.4.0