In [1]:
import numpy as np
import pandas as pd
import mikeio
from mikeio import ItemInfo, EUMType, EUMUnit
Write a dfs0¶
A mikeio.Dataset contains the information needed to write a dfs file. A Dataset consists of one or more mikeio.DataArrays each corresponding to an "item" in a dfs file.
In [2]:
nt = 10
time = pd.date_range("2000-1-1", periods=nt, freq='H')
d1 = np.zeros(nt)
item = ItemInfo("Zeros", EUMType.Water_Level)
da1 = mikeio.DataArray(d1, time=time, item=item)
d2 = np.ones(nt)
item = ItemInfo("Ones", EUMType.Discharge, EUMUnit.meter_pow_3_per_sec)
da2 = mikeio.DataArray(d2, time=time, item=item)
ds = mikeio.Dataset([da1, da2])
ds
Out[2]:
<mikeio.Dataset> dims: (time:10) time: 2000-01-01 00:00:00 - 2000-01-01 09:00:00 (10 records) geometry: GeometryUndefined() items: 0: Zeros <Water Level> (meter) 1: Ones <Discharge> (meter pow 3 per sec)
In [3]:
ds.is_equidistant
Out[3]:
True
In [4]:
ds.to_dfs("test.dfs0", title="Zeros and ones")
Read a timeseries¶
A dfs file is easily read with mikeio.read which returns a Dataset.
In [5]:
ds = mikeio.read("test.dfs0")
ds
Out[5]:
<mikeio.Dataset> dims: (time:10) time: 2000-01-01 00:00:00 - 2000-01-01 09:00:00 (10 records) geometry: GeometryUndefined() items: 0: Zeros <Water Level> (meter) 1: Ones <Discharge> (meter pow 3 per sec)
From comma separated file¶
In [6]:
df = pd.read_csv("../tests/testdata/co2-mm-mlo.csv", parse_dates=True, index_col='Date', na_values=-99.99)
df.head()
Out[6]:
| Decimal Date | Average | Interpolated | Trend | Number of Days | |
|---|---|---|---|---|---|
| Date | |||||
| 1958-03-01 | 1958.208 | 315.71 | 315.71 | 314.62 | -1 |
| 1958-04-01 | 1958.292 | 317.45 | 317.45 | 315.29 | -1 |
| 1958-05-01 | 1958.375 | 317.50 | 317.50 | 314.71 | -1 |
| 1958-06-01 | 1958.458 | NaN | 317.10 | 314.85 | -1 |
| 1958-07-01 | 1958.542 | 315.86 | 315.86 | 314.98 | -1 |
Remove missing values
In [7]:
df = df.dropna()
df = df[["Average","Trend"]]
df.plot()
Out[7]:
<Axes: xlabel='Date'>
A dataframe with a datetimeindex can be used to create a dfs0 with a non-equidistant time axis by first converting it to a mikeio.Dataset.
In [8]:
ds = mikeio.from_pandas(df)
ds
Out[8]:
<mikeio.Dataset> dims: (time:720) time: 1958-03-01 00:00:00 - 2018-09-01 00:00:00 (720 non-equidistant records) geometry: GeometryUndefined() items: 0: Average <Undefined> (undefined) 1: Trend <Undefined> (undefined)
And then write to a dfs0 file:
In [9]:
ds.to_dfs("mauna_loa_co2.dfs0")
To get a equidistant time axis first interpolate to regularly spaced values, in this case daily.
The code for this can be written in many ways, below is an example, where we avoid temporary variables.
In [10]:
(
df.resample("D") # resample to daily
.interpolate() # interpolate linearly
.pipe(mikeio.from_pandas) # convert to mikeio.Dataset
.to_dfs("mauna_loa_co2_daily.dfs0") # save to dfs0
)
Read a timeseries¶
In [11]:
res = mikeio.read("test.dfs0")
res
Out[11]:
<mikeio.Dataset> dims: (time:10) time: 2000-01-01 00:00:00 - 2000-01-01 09:00:00 (10 records) geometry: GeometryUndefined() items: 0: Zeros <Water Level> (meter) 1: Ones <Discharge> (meter pow 3 per sec)
In [12]:
res.time
Out[12]:
DatetimeIndex(['2000-01-01 00:00:00', '2000-01-01 01:00:00',
'2000-01-01 02:00:00', '2000-01-01 03:00:00',
'2000-01-01 04:00:00', '2000-01-01 05:00:00',
'2000-01-01 06:00:00', '2000-01-01 07:00:00',
'2000-01-01 08:00:00', '2000-01-01 09:00:00'],
dtype='datetime64[ns]', freq=None)
In [13]:
res.to_numpy()
Out[13]:
array([[0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],
[1., 1., 1., 1., 1., 1., 1., 1., 1., 1.]])
Or as a Pandas dataframe¶
A mikeio.Dataset ds is converted to a pandas dataframe with ds.to_dataframe()
In [14]:
dfs0file = "../tests/testdata/da_diagnostic.dfs0"
df = mikeio.read(dfs0file).to_dataframe()
df.head()
Out[14]:
| State 1Sign. Wave Height | State 2Sign. Wave Height | Mean StateSign. Wave Height | MeasurementSign. Wave Height | |
|---|---|---|---|---|
| 2017-10-27 00:00:00 | 1.749465 | 1.749465 | 1.749465 | 1.72 |
| 2017-10-27 00:10:00 | 1.811340 | 1.796895 | 1.807738 | NaN |
| 2017-10-27 00:20:00 | 1.863424 | 1.842759 | 1.853422 | NaN |
| 2017-10-27 00:30:00 | 1.922261 | 1.889839 | 1.897670 | NaN |
| 2017-10-27 00:40:00 | 1.972455 | 1.934886 | 1.935281 | NaN |
In [15]:
dfs0file = "../tests/testdata/random.dfs0"
df = mikeio.read(dfs0file).to_dataframe()
df.head()
Out[15]:
| VarFun01 | NotFun | |
|---|---|---|
| 2017-01-01 00:00:00 | 0.843547 | 0.640486 |
| 2017-01-01 05:00:00 | 0.093729 | 0.653257 |
| 2017-01-01 10:00:00 | NaN | NaN |
| 2017-01-01 15:00:00 | 0.305065 | 0.214208 |
| 2017-01-01 20:00:00 | 0.900190 | 0.999157 |
Create a timeseries with non-equidistant data¶
In [16]:
d1 = np.random.uniform(low=0.0, high=5.0, size=5)
time = pd.DatetimeIndex(["2000-1-1", "2000-1-8", "2000-1-10", "2000-2-22", "2000-11-29"])
da = mikeio.DataArray(d1, time=time, item=ItemInfo("Random"))
da
Out[16]:
<mikeio.DataArray> name: Random dims: (time:5) time: 2000-01-01 00:00:00 - 2000-11-29 00:00:00 (5 non-equidistant records) geometry: GeometryUndefined() values: [3.927, 0.526, ..., 2.795]
In [17]:
da.is_equidistant
Out[17]:
False
In [18]:
da.to_dfs("neq.dfs0", title="Non equidistant")
Create a timeseries with accumulated timestep¶
Find correct eum units¶
In [19]:
EUMType.search("prec")
Out[19]:
[Correction of precipitation, Precipitation correction, Precipitation, Specific Precipitation, Precipitation Rate]
In [20]:
EUMType.Precipitation_Rate.units
Out[20]:
[mm per day, mm per hour, cm per hour, meter per sec, meter per day, feet per day, inch per hour, inch per min, inch per day, mm per year]
In [21]:
from mikecore.DfsFile import DataValueType
n= 1000
time = pd.date_range("2017-01-01 00:00", freq='H', periods=n)
# use default name and unit based on type
item = ItemInfo(EUMType.Water_Level, data_value_type=DataValueType.Instantaneous)
da1 = mikeio.DataArray(data=np.random.random([n]), time=time, item=item)
# use a custom name
item = ItemInfo("Nedbør", EUMType.Precipitation_Rate, data_value_type=DataValueType.Accumulated)
da2 = mikeio.DataArray(data=np.random.random([n]), time=time, item=item)
ds = mikeio.Dataset([da1, da2])
ds.to_dfs('accumulated.dfs0')
In [22]:
ds = mikeio.read("accumulated.dfs0")
ds
Out[22]:
<mikeio.Dataset> dims: (time:1000) time: 2017-01-01 00:00:00 - 2017-02-11 15:00:00 (1000 records) geometry: GeometryUndefined() items: 0: Water Level <Water Level> (meter) 1: Nedbør <Precipitation Rate> (mm per day) - 1
Modify an existing timeseries¶
In [23]:
ds = mikeio.read("test.dfs0")
ds
Out[23]:
<mikeio.Dataset> dims: (time:10) time: 2000-01-01 00:00:00 - 2000-01-01 09:00:00 (10 records) geometry: GeometryUndefined() items: 0: Zeros <Water Level> (meter) 1: Ones <Discharge> (meter pow 3 per sec)
In [24]:
ds['Ones']
Out[24]:
<mikeio.DataArray> name: Ones dims: (time:10) time: 2000-01-01 00:00:00 - 2000-01-01 09:00:00 (10 records) geometry: GeometryUndefined() values: [1, 1, ..., 1]
Modify the data in some way...
In [25]:
ds['Ones'] = ds['Ones']*np.pi
ds['Ones'].values
Out[25]:
array([3.14159265, 3.14159265, 3.14159265, 3.14159265, 3.14159265,
3.14159265, 3.14159265, 3.14159265, 3.14159265, 3.14159265])
In [26]:
ds.to_dfs("modified.dfs0")
In [27]:
res = mikeio.read("modified.dfs0")
res['Ones']
Out[27]:
<mikeio.DataArray> name: Ones dims: (time:10) time: 2000-01-01 00:00:00 - 2000-01-01 09:00:00 (10 records) geometry: GeometryUndefined() values: [3.142, 3.142, ..., 3.142]
The second item is not modified.
In [28]:
res['Zeros']
Out[28]:
<mikeio.DataArray> name: Zeros dims: (time:10) time: 2000-01-01 00:00:00 - 2000-01-01 09:00:00 (10 records) geometry: GeometryUndefined() values: [0, 0, ..., 0]
Convert units¶
Read a file with waterlevel i meters.
In [29]:
filename = "../tests/testdata/waterlevel_viken.dfs0"
# filename = r"C:\Program Files (x86)\DHI\MIKE Zero\2021\Examples\MIKE_21\FlowModel_FM\HD\Oresund\Data\1993\Boundary_Conditions\waterlevel_viken.dfs0"
ds = mikeio.read(filename)
ds
Out[29]:
<mikeio.Dataset> dims: (time:577) time: 1993-12-02 00:00:00 - 1993-12-14 00:00:00 (577 records) geometry: GeometryUndefined() items: 0: ST 2: WL (m) <Water Level> (meter)
In [30]:
ds.plot()
Out[30]:
<Axes: >
The aim is to convert this timeseries to feet (1m = 3.3 ft)
In [31]:
ds[0] = ds[0]*3.3
Which units are acceptable?
In [32]:
ds.items[0].type.units
Out[32]:
[meter, kilometer, centimeter, millimeter, feet, feet US, inch, inch US, mile, mile US, yard, yard US]
In [33]:
ds[0].item = ItemInfo("Viken", ds[0].item.type, EUMUnit.feet)
In [34]:
ds.to_dfs("wl_feet.dfs0")
Extrapolation¶
In [35]:
# filename = r"C:\Program Files (x86)\DHI\MIKE Zero\2021\Examples\MIKE_21\FlowModel_FM\HD\Oresund\Data\1993\Boundary_Conditions\waterlevel_viken.dfs0"
filename = "../tests/testdata/waterlevel_viken.dfs0"
ds = mikeio.read(filename)
df = ds.to_dataframe()
df.plot()
Out[35]:
<Axes: >
In [36]:
rng = pd.date_range("1993-12-1","1994-1-1",freq='30t')
ix = pd.DatetimeIndex(rng)
dfr = df.reindex(ix)
dfr.plot()
Out[36]:
<Axes: >
Replace NaN with constant extrapolation (forward fill + back fill).
In [37]:
dfr = dfr.ffill().bfill()
dfr.plot()
Out[37]:
<Axes: >
In [38]:
dfr.to_dfs0("Viken_extrapolated.dfs0",
items=ds.items,
title="Caution extrapolated data!"
)
Clean up¶
In [39]:
import os
os.remove("test.dfs0")
os.remove("modified.dfs0")
os.remove("neq.dfs0")
os.remove("accumulated.dfs0")
os.remove("wl_feet.dfs0")
os.remove("mauna_loa_co2_daily.dfs0")
os.remove("mauna_loa_co2.dfs0")
os.remove("Viken_extrapolated.dfs0")