Remove networking code from iotools.read_tmy3, tkinter from read_tmy2/3 (#1004)

* remove networking code from tmy.read_tmy3

* whatsnew

* whats even newer

* improvements from review

* forgot open(pathlib.Path) doesn't work on python 3.5

* move line outside with block

* whatsnew entry for tkinter removal

Author: kandersolar

docs/sphinx/source/whatsnew/v0.8.0.rst

@@ -16,7 +16,15 @@ API Changes with Deprecations
 API Changes
 ~~~~~~~~~~~
 * Removed ``run_parallel_calculations`` and ``n_workers_for_parallel_calcs``
-  from :py:func:`pvlib.bifacial.pvfactors_timeseries` inputs (:issue:`902`)(:pull:`934`)
+  from :py:func:`pvlib.bifacial.pvfactors_timeseries` inputs (:issue:`902`) (:pull:`934`)
+* :py:func:`pvlib.iotools.read_tmy3` can now only read local data files because
+  the NREL RREDC server hosting the TMY3 dataset has been retired. For
+  fetching TMY data from NREL servers, :py:func:`pvlib.iotools.get_psm3` is
+  now recommended to retrieve newer PSM3 data over the older TMY3 data.
+  (:issue:`996`) (:pull:`1004`)
+* The tkinter-based file selection dialog has been removed from
+  :py:func:`pvlib.iotools.read_tmy2` and :py:func:`pvlib.iotools.read_tmy3`;
+  the filepath is now a required parameter. (:pull:`1004`)
 
 Enhancements
 ~~~~~~~~~~~~

pvlib/iotools/tmy.py

@@ -3,29 +3,26 @@
 """
 
 import datetime
-import io
 import re
-from urllib.request import urlopen, Request
 import pandas as pd
 
 
-def read_tmy3(filename=None, coerce_year=None, recolumn=True):
+def read_tmy3(filename, coerce_year=None, recolumn=True):
     '''
     Read a TMY3 file in to a pandas dataframe.
 
     Note that values contained in the metadata dictionary are unchanged
     from the TMY3 file (i.e. units are retained). In the case of any
-    discrepencies between this documentation and the TMY3 User's Manual
+    discrepancies between this documentation and the TMY3 User's Manual
     [1]_, the TMY3 User's Manual takes precedence.
 
     The TMY3 files were updated in Jan. 2015. This function requires the
     use of the updated files.
 
     Parameters
     ----------
-    filename : None or string, default None
-        If None, attempts to use a Tkinter file browser. A string can be
-        a relative file path, absolute file path, or url.
+    filename : str
+        A relative file path or absolute file path.
 
     coerce_year : None or int, default None
         If supplied, the year of the index will be set to `coerce_year`, except
@@ -161,46 +158,25 @@ def read_tmy3(filename=None, coerce_year=None, recolumn=True):
        Update: Users Manual. 472 pp.; NREL Report No. TP-581-41364.
     '''
 
-    if filename is None:
-        try:
-            filename = _interactive_load()
-        except ImportError:
-            raise ImportError('Interactive load failed. tkinter not supported '
-                              'on this system. Try installing X-Quartz and '
-                              'reloading')
-
     head = ['USAF', 'Name', 'State', 'TZ', 'latitude', 'longitude', 'altitude']
 
-    if str(filename).startswith('http'):
-        request = Request(filename, headers={'User-Agent': (
-            'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_13_5) '
-            'AppleWebKit/537.36 (KHTML, like Gecko) Chrome/67.0.3396.87 '
-            'Safari/537.36')})
-        response = urlopen(request)
-        csvdata = io.StringIO(response.read().decode(errors='ignore'))
-    else:
-        # assume it's accessible via the file system
-        csvdata = open(str(filename), 'r')
-
-    # read in file metadata, advance buffer to second line
-    firstline = csvdata.readline()
-    if 'Request Rejected' in firstline:
-        raise IOError('Remote server rejected TMY file request')
+    with open(str(filename), 'r') as csvdata:
+        # read in file metadata, advance buffer to second line
+        firstline = csvdata.readline()
+        # use pandas to read the csv file buffer
+        # header is actually the second line, but tell pandas to look for
+        # header information on the 1st line (0 indexing) because we've already
+        # advanced past the true first line with the readline call above.
+        data = pd.read_csv(csvdata, header=0)
 
     meta = dict(zip(head, firstline.rstrip('\n').split(",")))
-
     # convert metadata strings to numeric types
     meta['altitude'] = float(meta['altitude'])
     meta['latitude'] = float(meta['latitude'])
     meta['longitude'] = float(meta['longitude'])
     meta['TZ'] = float(meta['TZ'])
     meta['USAF'] = int(meta['USAF'])
 
-    # use pandas to read the csv file/stringio buffer
-    # header is actually the second line in file, but tell pandas to look for
-    # header information on the 1st line (0 indexing) because we've already
-    # advanced past the true first line with the readline call above.
-    data = pd.read_csv(csvdata, header=0)
     # get the date column as a pd.Series of numpy datetime64
     data_ymd = pd.to_datetime(data['Date (MM/DD/YYYY)'], format='%m/%d/%Y')
     # shift the time column so that midnite is 00:00 instead of 24:00
@@ -231,13 +207,6 @@ def read_tmy3(filename=None, coerce_year=None, recolumn=True):
     return data, meta
 
 
-def _interactive_load():
-    import tkinter
-    from tkinter.filedialog import askopenfilename
-    tkinter.Tk().withdraw()  # Start interactive file input
-    return askopenfilename()
-
-
 def _recolumn(tmy3_dataframe):
     """
     Rename the columns of the TMY3 DataFrame.
@@ -295,9 +264,8 @@ def read_tmy2(filename):
 
     Parameters
     ----------
-    filename : None or string
-        If None, attempts to use a Tkinter file browser. A string can be
-        a relative file path, absolute file path, or url.
+    filename : str
+        A relative or absolute file path.
 
     Returns
     -------
@@ -412,14 +380,6 @@ def read_tmy2(filename):
        for TMY2s". NREL 1995.
     '''
 
-    if filename is None:
-        try:
-            filename = _interactive_load()
-        except ImportError:
-            raise ImportError('Interactive load failed. tkinter not supported '
-                              'on this system. Try installing X-Quartz and '
-                              'reloading')
-
     # paste in the column info as one long line
     string = '%2d%2d%2d%2d%4d%4d%4d%1s%1d%4d%1s%1d%4d%1s%1d%4d%1s%1d%4d%1s%1d%4d%1s%1d%4d%1s%1d%2d%1s%1d%2d%1s%1d%4d%1s%1d%4d%1s%1d%3d%1s%1d%4d%1s%1d%3d%1s%1d%3d%1s%1d%4d%1s%1d%5d%1s%1d%10d%3d%1s%1d%3d%1s%1d%3d%1s%1d%2d%1s%1d'  # noqa: E501
     columns = 'year,month,day,hour,ETR,ETRN,GHI,GHISource,GHIUncertainty,DNI,DNISource,DNIUncertainty,DHI,DHISource,DHIUncertainty,GHillum,GHillumSource,GHillumUncertainty,DNillum,DNillumSource,DNillumUncertainty,DHillum,DHillumSource,DHillumUncertainty,Zenithlum,ZenithlumSource,ZenithlumUncertainty,TotCld,TotCldSource,TotCldUncertainty,OpqCld,OpqCldSource,OpqCldUncertainty,DryBulb,DryBulbSource,DryBulbUncertainty,DewPoint,DewPointSource,DewPointUncertainty,RHum,RHumSource,RHumUncertainty,Pressure,PressureSource,PressureUncertainty,Wdir,WdirSource,WdirUncertainty,Wspd,WspdSource,WspdUncertainty,Hvis,HvisSource,HvisUncertainty,CeilHgt,CeilHgtSource,CeilHgtUncertainty,PresentWeather,Pwat,PwatSource,PwatUncertainty,AOD,AODSource,AODUncertainty,SnowDepth,SnowDepthSource,SnowDepthUncertainty,LastSnowfall,LastSnowfallSource,LastSnowfallUncertaint'  # noqa: E501

pvlib/tests/iotools/test_tmy.py

@@ -1,8 +1,7 @@
 import numpy as np
 import pandas as pd
-import pytest
 from pvlib.iotools import tmy
-from conftest import DATA_DIR, RERUNS, RERUNS_DELAY
+from conftest import DATA_DIR
 
 # test the API works
 from pvlib.iotools import read_tmy3
@@ -16,13 +15,6 @@ def test_read_tmy3():
     tmy.read_tmy3(TMY3_TESTFILE)
 
 
-@pytest.mark.remote_data
-@pytest.mark.flaky(reruns=RERUNS, reruns_delay=RERUNS_DELAY)
-def test_read_tmy3_remote():
-    url = 'http://rredc.nrel.gov/solar/old_data/nsrdb/1991-2005/data/tmy3/703165TYA.CSV'
-    tmy.read_tmy3(url)
-
-
 def test_read_tmy3_recolumn():
     data, meta = tmy.read_tmy3(TMY3_TESTFILE)
     assert 'GHISource' in data.columns
@@ -47,6 +39,7 @@ def test_read_tmy3_no_coerce_year():
     assert data.index[-2] == pd.Timestamp('1998-12-31 23:00:00-09:00')
     assert data.index[-1] == pd.Timestamp('1999-01-01 00:00:00-09:00')
 
+
 def test_read_tmy2():
     tmy.read_tmy2(TMY2_TESTFILE)