Add read_bsrn function

docs/sphinx/source/api.rst

@@ -482,6 +482,7 @@ relevant to solar energy modeling.
    iotools.parse_psm3
    iotools.get_pvgis_tmy
    iotools.read_pvgis_tmy
+   iotools.read_bsrn
 
 A :py:class:`~pvlib.location.Location` object may be created from metadata
 in some files.

docs/sphinx/source/whatsnew/v0.9.0.rst

@@ -58,6 +58,8 @@ Deprecations
 
 Enhancements
 ~~~~~~~~~~~~
+* Add :func:`~pvlib.iotools.read_bsrn` for reading BSRN solar radiation data
+  files. (:pull:`1145`, :issue:`1015`)
 * In :py:class:`~pvlib.modelchain.ModelChain`, attributes which contain
   output of models are now collected into ``ModelChain.results``.
   (:pull:`1076`, :issue:`1067`)
@@ -113,3 +115,4 @@ Contributors
 * Mark Mikofski (:ghuser:`mikofski`)
 * Nate Croft (:ghuser:`ncroft-b4`)
 * Kevin Anderson (:ghuser:`kanderso-nrel`)
+* Adam R. Jensen (:ghuser:`AdamRJensen`)

pvlib/iotools/__init__.py

@@ -13,3 +13,4 @@
 from pvlib.iotools.psm3 import read_psm3  # noqa: F401
 from pvlib.iotools.psm3 import parse_psm3  # noqa: F401
 from pvlib.iotools.pvgis import get_pvgis_tmy, read_pvgis_tmy  # noqa: F401
+from pvlib.iotools.bsrn import read_bsrn  # noqa: F401

pvlib/iotools/bsrn.py

@@ -0,0 +1,137 @@
+"""Functions to read data from the Baseline Surface Radiation Network (BSRN).
+.. codeauthor:: Adam R. Jensen<adam-r-j@hotmail.com>
+"""
+
+import pandas as pd
+import gzip
+
+COL_SPECS = [(0, 3), (4, 9), (10, 16), (16, 22), (22, 27), (27, 32), (32, 39),
+             (39, 45), (45, 50), (50, 55), (55, 64), (64, 70), (70, 75)]
+
+BSRN_COLUMNS = ['day', 'minute',
+                'ghi', 'ghi_std', 'ghi_min', 'ghi_max',
+                'dni', 'dni_std', 'dni_min', 'dni_max',
+                'empty', 'empty', 'empty', 'empty', 'empty',
+                'dhi', 'dhi_std', 'dhi_min', 'dhi_max',
+                'lwd', 'lwd_std', 'lwd_min', 'lwd_max',
+                'temp_air', 'relative_humidity', 'pressure']
+
+
+def read_bsrn(filename):
+    """
+    Read a BSRN station-to-archive file into a DataFrame.
+
+    The BSRN (Baseline Surface Radiation Network) is a world wide network
+    of high-quality solar radiation monitoring stations as described in [1]_.
+    The function only parses the basic measurements (LR0100), which include
+    global, diffuse, direct and downwelling long-wave radiation [2]_. Future
+    updates may include parsing of additional data and meta-data.
+
+    BSRN files are freely avaiable and can be accessed via FTP [3]_. Required
+    username and password are easily obtainable as described in the BSRN's
+    Data Release Guidelines [4]_.
+
+
+
+    Parameters
+    ----------
+    filename: str
+        A relative or absolute file path.
+
+    Returns
+    -------
+    data: Dataframe
+        A Dataframe with the columns as described below. For more extensive
+        description of the variables, consult [2]_.
+
+    Notes
+    -----
+    The data Dataframe includes the following fields:
+
+    =======================  ======  ==========================================
+    Key                      Format  Description
+    =======================  ======  ==========================================
+    day                      int     Day of the month 1-31
+    minute                   int     Minute of the day 0-1439
+    ghi                      float   Mean global horizontal irradiance [W/m^2]
+    ghi_std                  float   Std. global horizontal irradiance [W/m^2]
+    ghi_min                  float   Min. global horizontal irradiance [W/m^2]
+    ghi_max                  float   Max. global horizontal irradiance [W/m^2]
+    dni                      float   Mean direct normal irradiance [W/m^2]
+    dni_std                  float   Std. direct normal irradiance [W/m^2]
+    dni_min                  float   Min. direct normal irradiance [W/m^2]
+    dni_max                  float   Max. direct normal irradiance [W/m^2]
+    dhi                      float   Mean diffuse horizontal irradiance [W/m^2]
+    dhi_std                  float   Std. diffuse horizontal irradiance [W/m^2]
+    dhi_min                  float   Min. diffuse horizontal irradiance [W/m^2]
+    dhi_max                  float   Max. diffuse horizontal irradiance [W/m^2]
+    lwd                      float   Mean. downward long-wave radiation [W/m^2]
+    lwd_std                  float   Std. downward long-wave radiation [W/m^2]
+    lwd_min                  float   Min. downward long-wave radiation [W/m^2]
+    lwd_max                  float   Max. downward long-wave radiation [W/m^2]
+    temp_air                 float   Air temperature [°C]
+    relative_humidity        float   Relative humidity [%]
+    pressure                 float   Atmospheric pressure [hPa]
+    =======================  ======  ==========================================
+
+    References
+    ----------
+    .. [1] `World Radiation Monitoring Center - Baseline Surface Radiation
+        Network (BSRN) <https:/https://bsrn.awi.de/>`_
+    .. [2] `Update of the Technical Plan for BSRN Data Management, 2013,
+       Global Climate Observing System (GCOS) GCOS-172.
+       <https://bsrn.awi.de/fileadmin/user_upload/bsrn.awi.de/Publications/gcos-174.pdf>`_
+    .. [3] `BSRN Data Retrieval via FTP
+       <https://bsrn.awi.de/data/data-retrieval-via-ftp/>`_
+    .. [4] `BSRN Data Release Guidelines
+       <https://bsrn.awi.de/data/conditions-of-data-release/>`_
+    """
+
+    # Read file and store the starting line number for each logical record (LR)
+    line_no_dict = {}
+    if str(filename).endswith('.gz'):  # check if file is a gzipped (.gz) file
+        open_func, mode = gzip.open, 'rt'
+    else:
+        open_func, mode = open, 'r'
+    with open_func(filename, mode) as f:
+        for num, line in enumerate(f):
+            if num == 1:  # Get month and year from the 2nd line
+                start_date = pd.Timestamp(year=int(line[7:11]),
+                                          month=int(line[3:6]), day=1,
+                                          tz='UTC')  # BSRN timestamps are UTC
+            if line.startswith('*'):  # Find start of all logical records
+                line_no_dict[line[2:6]] = num  # key is 4 digit LR number
+
+    # Determine start and end line of logical record LR0100 to be parsed
+    start_row = line_no_dict['0100'] + 1  # Start line number
+    # If LR0100 is the last logical record, then read rest of file
+    if start_row-1 == max(line_no_dict.values()):
+        end_row = num  # then parse rest of the file
+    else:  # otherwise parse until the beginning of the next logical record
+        end_row = min([i for i in line_no_dict.values() if i > start_row])
+    nrows = end_row-start_row
+
+    # Read file as a fixed width file (fwf)
+    data = pd.read_fwf(filename, skiprows=start_row, nrows=nrows, header=None,
+                       colspecs=COL_SPECS, na_values=[-999.0, -99.9])
+
+    # Create multi-index and unstack, resulting in one column for each variable
+    data = data.set_index([data.index // 2, data.index % 2])
+    data = data.unstack(level=1).swaplevel(i=0, j=1, axis='columns')
+
+    # Sort columns to match original order and assign column names
+    data = data.reindex(sorted(data.columns), axis='columns')
+    data.columns = BSRN_COLUMNS
+    # Drop empty columns
+    data = data.drop('empty', axis='columns')
+
+    # Change day and minute type to integer
+    data['day'] = data['day'].astype('Int64')
+    data['minute'] = data['minute'].astype('Int64')
+
+    # Set datetime index
+    data.index = (start_date
+                  + pd.to_timedelta(data['day']-1, unit='d')
+                  + pd.to_timedelta(data['minute'], unit='min'))
+
+    return data

pvlib/tests/iotools/test_bsrn.py

@@ -0,0 +1,34 @@
+"""
+tests for :mod:`pvlib.iotools.bsrn`
+"""
+
+
+import pandas as pd
+import pytest
+
+from pvlib.iotools import bsrn
+from conftest import DATA_DIR
+
+
+# Awaiting permission from BSRN to upload test file
+testfile = DATA_DIR / 'bsrn-pay0616.dat.gz'
+
+
+def test_read_bsrn_columns():
+    data = bsrn.read_bsrn(testfile)
+    assert 'ghi' in data.columns
+    assert 'dni_std' in data.columns
+    assert 'dhi_min' in data.columns
+    assert 'lwd_max' in data.columns
+    assert 'relative_humidity' in data.columns
+
+
+@pytest.fixture
+def expected_index():
+    start = pd.Timestamp(2016, 6, 1, 0, 0)
+    return pd.date_range(start=start, periods=43200, freq='1min', tz='UTC')
+
+
+def test_format_index():
+    actual = bsrn.read_bsrn(testfile)
+    assert actual.index.equals(expected_index())