From e3d0e2ada5dfd3fd05cbbebc5098806e16555778 Mon Sep 17 00:00:00 2001
From: veenstrajelmer <60435591+veenstrajelmer@users.noreply.github.com>
Date: Wed, 23 Oct 2024 20:56:31 +0200
Subject: [PATCH] updated pandas version and applied diff on timedeltas (#161)

* updated minimal pandas version

* updated code to use diff for timedeltas

* updated minimal numpy version

* updated whatsnew
---
 docs/whats-new.md                     |  8 ++++----
 kenmerkendewaarden/data_analysis.py   |  4 +---
 kenmerkendewaarden/tidalindicators.py |  4 +---
 pyproject.toml                        | 22 +++++++++++-----------
 4 files changed, 17 insertions(+), 21 deletions(-)

diff --git a/docs/whats-new.md b/docs/whats-new.md
index 5633466..9eb03ef 100644
--- a/docs/whats-new.md
+++ b/docs/whats-new.md
@@ -3,10 +3,10 @@
 ## UNRELEASED
 
 ### Feat
-- expanded physical_break_dict in [#151](https://github.com/Deltares-research/kenmerkendewaarden/pull/151)
-- linear fit for slotgemiddelden (no nodal) in [#157](https://github.com/Deltares-research/kenmerkendewaarden/pull/157)
-- exposed yearly tidal components with `kw.calc_getijcomponenten()` in [#159](https://github.com/Deltares-research/kenmerkendewaarden/pull/159)
-
+- expanded physical_break_dict in #151
+- linear fit for slotgemiddelden (no nodal) in #157
+- exposed yearly tidal components with `kw.calc_getijcomponenten()` in #159
+- support for timedelta `diff()` because of update to `pandas>=2.1.4` in #161
 
 ## 0.3.0 (2024-10-01)
 
diff --git a/kenmerkendewaarden/data_analysis.py b/kenmerkendewaarden/data_analysis.py
index 91c0dad..8068286 100644
--- a/kenmerkendewaarden/data_analysis.py
+++ b/kenmerkendewaarden/data_analysis.py
@@ -221,9 +221,7 @@ def get_flat_meta_from_dataset(ds):
 def get_stats_from_dataframe(df):
     df_times = df.index
     ts_dupltimes = df_times.duplicated()
-    ts_timediff = (
-        df_times[1:] - df_times[:-1]
-    )  # TODO: from pandas 2.1.4 the following also works: df_times.diff()[1:]
+    ts_timediff = df_times.diff()[1:]
 
     ds_stats = {}
     ds_stats["tstart"] = df_times.min()
diff --git a/kenmerkendewaarden/tidalindicators.py b/kenmerkendewaarden/tidalindicators.py
index 45f5ab0..04e77e4 100644
--- a/kenmerkendewaarden/tidalindicators.py
+++ b/kenmerkendewaarden/tidalindicators.py
@@ -221,9 +221,7 @@ def compute_expected_counts(ser_meas, freq):
     """
     # TODO: beware of series with e.g. only first and last value of month/year, this will result in freq=30days and then expected count of 2, it will pass even if there is almost no data
     df_meas = pd.DataFrame(ser_meas)
-    df_meas["timediff"] = pd.TimedeltaIndex([pd.NaT]).append(
-        df_meas.index[1:] - df_meas.index[:-1]
-    )  # TODO: from pandas>=2.1.4 the following also works: df_times.diff() (which results in a timedeltaindex of the correct length)
+    df_meas["timediff"] = df_meas.index.diff()
     period_index = pd.PeriodIndex(df_meas.index, freq=freq)
     # compute median freq, the mean could be skewed in case of large gaps
     median_freq = df_meas.groupby(period_index)["timediff"].median()
diff --git a/pyproject.toml b/pyproject.toml
index c2c65d2..393937e 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -12,25 +12,25 @@ keywords = ["kenmerkendewaarden"]
 license = { text = "GPLv3" }
 requires-python = ">=3.9"
 dependencies = [
-	#numpy 1.21 is EOL since june 2023
-	"numpy>=1.22.0",
-	#pandas<2.0.0 supports non-nanosecond timestep and therefore larger min/max range
-	"pandas>=2.0.0",
-	#matplotlib<3.5.2 raises "ValueError: Multi-dimensional indexing" in hatyan.plot_timeseries()
+	#numpy>=1.25.0 supports wide range of python versions
+	"numpy>=1.25.0",
+	#pandas>=2.1.4 supports diff on timedeltas
+	"pandas>=2.1.4",
+	#matplotlib>=3.5.2 resolved "ValueError: Multi-dimensional indexing" in hatyan.plot_timeseries()
 	"matplotlib>=3.5.2",
-	#netcdf4<1.5.4 pip install fails in py39
+	#netcdf4>=1.5.4 pip install succeeds in py39
 	"netcdf4>=1.5.4",
-	#xarray<2023.4.0 conflicts with pandas<2.0.0 for resampling, only available for py39
+	#xarray>=2023.4.0 works with pandas<2.0.0 for resampling, only available for py39
 	"xarray>=2023.4.0",
 	#rws-ddlpy>=0.6.0 returns correct measurements_amount dataframe
 	"rws-ddlpy>=0.6.0",
-	#hatyan<2.9.0 has different datetime/tzone handling in astrog
+	#hatyan>=2.9.0 has different datetime/tzone handling in astrog
 	"hatyan>=2.9.0",
-	#statsmodels<0.13.2 has no support for pandas 2.0.0
+	#statsmodels>=0.13.2 has support for pandas 2.0.0
 	"statsmodels>=0.13.2",
-	#pyproj<3.1.0 not pip installable in py38
+	#pyproj>=3.1.0 pip installable in py38
 	"pyproj>=3.1.0",
-	#pooch<1.1.0 do not have attribute retrieve
+	#pooch>=1.1.0 has attribute retrieve
 	"pooch>=1.1.0",
 ]
 classifiers = [