From 5d8968fc5fa33ec62712d0df2f92891b47ce24c7 Mon Sep 17 00:00:00 2001
From: Sander Roosendaal <sanderroosendaal@proton.me>
Date: Sat, 13 Apr 2024 17:19:47 +0200
Subject: [PATCH] boxchart pandas to polars

---
 rowers/interactiveplots.py            |  13 +++++------
 rowers/tests/testdata/testdata.tcx.gz | Bin 4000 -> 4001 bytes
 rowers/views/analysisviews.py         |  30 +++++++++-----------------
 3 files changed, 15 insertions(+), 28 deletions(-)

diff --git a/rowers/interactiveplots.py b/rowers/interactiveplots.py
index 1f8a073f..6f2b42e0 100644
--- a/rowers/interactiveplots.py
+++ b/rowers/interactiveplots.py
@@ -328,7 +328,7 @@ def interactive_workouttype_piechart(workouts):
 def interactive_boxchart(datadf, fieldname, extratitle='',
                          spmmin=0, spmmax=0, workmin=0, workmax=0):
 
-    if datadf.empty:  # pragma: no cover
+    if datadf.is_empty():  # pragma: no cover
         return '', 'It looks like there are no data matching your filter'
 
     columns = datadf.columns
@@ -339,20 +339,17 @@ def interactive_boxchart(datadf, fieldname, extratitle='',
     if 'date' not in columns:  # pragma: no cover
         return '', 'Not enough data'
 
-    try:
-        datadf.date = datadf.date.apply(lambda x:x.strftime("%Y-%m-%d"))
-    except AttributeError:
-        datadf.date = "2000-01-01"
+    datadf = datadf.with_columns((pl.col("date").apply(lambda x:x.strftime("%Y-%m-%d"))).alias("date"))
+    datadf = datadf.with_columns((pl.col(fieldname)).alias("value"))
     
-    datadf['value'] = datadf[fieldname]
-    data_dict = datadf.to_dict("records")
+    data_dict = datadf.to_dicts()
     boxplot_data = {
         "metric": metricsdicts[fieldname]["verbose_name"],
         "data": data_dict
     }
 
     
-    script, div = get_chart("/boxplot", boxplot_data)
+    script, div = get_chart("/boxplot", boxplot_data, debug=False)
     return script, div
 
 
diff --git a/rowers/tests/testdata/testdata.tcx.gz b/rowers/tests/testdata/testdata.tcx.gz
index 96809caca2971e112783a06ee9e4fc7fc5cc2180..0873ae7aafdc85f8930c098d869ebd25fb92fe0e 100644
GIT binary patch
delta 257
zcmV+c0sj7=AE6%xABzYGMpv;0s}6q`?D6m5@C3d(?Y}u(><)OaJJTP(F@L`PwAy`t
z`1mSuytw+X?~5+IJy#cRH{JTAJ3HGxt@LrVK0WFBE}ggAy<LgQ$Kwtkz3yLKT-|t0
z9$j9Zt&Y1s{q*Y><NDlRoGg9!{p;1)$@S0uCB2X7&>Iiu!<OgWf6^thv*2;&=P!3#
zt<z>NANSjCxPG*so&WG<y6K~zmS@XuvrHWFrw@<r#(f<=>(-}FpB(-0BD2~4x!e9@
zdW&DYKL2f*?hN1|!0!LU#f>j^Pk#HDjvj1Luh!`~JWeQ=S5MN@_$^)f=)wO0Dd@J=
H#DD<+oo177

delta 256
zcmV+b0ssD?AD|xwABzYGuvD=Is}6rS;SintZsDub{+q+a?tlloGyU-!^XKbNtKIj9
zkFOHPi>nX&zUb21b9M1{)2&asv$O5fN*`D2)03|6(s{ex+m)z%Jnrz(>;Bcn)s5HW
z(dFgY>bUFEPrrUKuFw6&$<lY<zh0f4T>so(()*YWz4359Y<b@OCtWf-3vO_J{&Kg~
zI&Jpyalh?`>qq<9`43;Fn?Cw!dA95}%fumn`tay(+}Gi=ZhiXn$<YrlGMnw6yX`-w
zxA?{D^WT=~&Hx?)?EXJo-1uVm<hPIM=)o5CYMq|L<Aic~^&~xw-_oUz9{dl?Nt05<
GfB^t2mY#9|

diff --git a/rowers/views/analysisviews.py b/rowers/views/analysisviews.py
index 391d92ad..414430c0 100644
--- a/rowers/views/analysisviews.py
+++ b/rowers/views/analysisviews.py
@@ -927,28 +927,18 @@ def boxplotdata(workouts, options):
     ids = [w.id for w in workouts]
 
     # prepare data frame
-    datadf, extracols = dataprep.read_cols_df_sql(ids, fieldlist)
+    datadf = getsmallrowdata_pl(fieldlist, ids)
 
-    datadf = dataprep.clean_df_stats(datadf, workstrokesonly=workstrokesonly)
+    datadf = dataprep.clean_df_stats_pl(datadf, workstrokesonly=workstrokesonly)
 
-    datadf = dataprep.filter_df(datadf, 'spm', spmmin,
-                                largerthan=True)
-    datadf = dataprep.filter_df(datadf, 'spm', spmmax,
-                                largerthan=False)
-    datadf = dataprep.filter_df(datadf, 'driveenergy', workmin,
-                                largerthan=True)
-    datadf = dataprep.filter_df(datadf, 'driveneergy', workmax,
-                                largerthan=False)
+    datadf = datadf.filter(
+        pl.col("spm")>spmmin,
+        pl.col("spm")<spmmax,
+        pl.col("driveenergy")>workmin,
+        pl.col("driveenergy")<workmax,
+    )
 
-    datadf.dropna(axis=0, how='any', inplace=True)
-    datadf = datadf[datadf['workoutid'].isin(ids) == True]
-
-    datadf['workoutid'].replace(datemapping, inplace=True)
-    datadf.rename(columns={"workoutid": "date"}, inplace=True)
-
-    datadf['date'] = pd.to_datetime(datadf['date'], errors='coerce')
-    datadf = datadf.dropna(subset=['date'])
-    datadf = datadf.sort_values(['date'])
+    datadf = datadf.with_columns((pl.col("workoutid").apply(lambda x: datemapping[x])).alias("date"))
 
     if userid == 0:  # pragma: no cover
         extratitle = ''
@@ -958,7 +948,7 @@ def boxplotdata(workouts, options):
 
     savedata = options.get('savedata',False)
     if savedata: # pragma: no cover
-        return datadf 
+        return datadf.to_pandas() 
 
     script, div = interactive_boxchart(datadf, plotfield,
                                        extratitle=extratitle,