Merge pull request #67 from j-andrews7/fix/64-current-year

Fix for calculation of current season/year (fixes #64)

kenpompy/misc.py

@@ -7,6 +7,23 @@
 import pandas as pd
 from bs4 import BeautifulSoup
 from io import StringIO
+import re
+
+def get_current_season(browser):
+	"""
+	Scrapes the KenPom homepage to get the latest season year that has data published
+
+	Args:
+		browser (mechanicalsoup StatefulBrowser): Authenticated browser with full access to kenpom.com generated
+            by the `login` function.
+
+	Returns:
+		current_season (int): Number corresponding to the last season year that has data published
+	"""
+	browser.open('https://kenpom.com/index.php')
+	page_title = browser.page.select_one('#content-header h2').text
+	YEAR_PATTERN = r'^(\d{4})'
+	return int(re.match(YEAR_PATTERN, page_title).group(0))
 
 def get_pomeroy_ratings(browser, season=None):
     """

kenpompy/team.py

@@ -4,8 +4,8 @@
 """
 
 import pandas as pd
-import datetime
 from io import StringIO
+from .misc import get_current_season
 
 def get_valid_teams(browser, season=None):
 	"""
@@ -39,9 +39,6 @@ def get_valid_teams(browser, season=None):
 
 	return team_list
 
-
-
-
 def get_schedule(browser, team=None, season=None):
 	"""
 	Scrapes a team's schedule from (https://kenpom.com/team.php) into a dataframe.
@@ -62,19 +59,17 @@ def get_schedule(browser, team=None, season=None):
 	"""
 
 	url = 'https://kenpom.com/team.php'
-
-	date = datetime.date.today()
-	currentYear = date.strftime("%Y")
+	current_season = get_current_season(browser)
 
 	if season:
 		if int(season) < 2002:
 			raise ValueError(
 				'season cannot be less than 2002, as data only goes back that far.')
-		if int(season) > int(currentYear):
+		if int(season) > int(current_season):
 			raise ValueError(
 				'season cannot be greater than the current year.')
 	else:
-		season = int(currentYear)
+		season = current_season
 
 	if team==None or team not in get_valid_teams(browser, season):
 			raise ValueError(

kenpompy/utils.py

@@ -47,4 +47,4 @@ def login(email, password):
 		raise Exception(
 			'Logging in to kenpom.com failed - account subscription is expired')
 
-	return browser
+	return browser

tests/test_misc.py

@@ -3,6 +3,15 @@
 from kenpompy.FanMatch import FanMatch
 import pandas as pd
 
+def test_get_current_season(browser):
+	current_season = kpmisc.get_current_season(browser)
+
+	# Test that the current season is indeed published on the homepage
+	assert browser.request(url='https://kenpom.com/?y=' + str(current_season), method='GET', allow_redirects=False).status_code == 200
+
+	# Test that there are isn't a season beyond the "current" one (that it is indeed the latest)
+	assert browser.request(url='https://kenpom.com/?y=' + str(current_season + 1), method='GET', allow_redirects=False).status_code == 302
+
 def test_get_pomeroy_ratings(browser):
     expected = ['1', 'Virginia', 'ACC', '35-3', '+34.22', '123.4', '2', '89.2', '5', '59.4', '353', '+.050', '62', '+11.18', '22', '109.2', '34', '98.1', '14', '-3.24', '255', '1']
     df = kpmisc.get_pomeroy_ratings(browser, season=2019)

tests/test_team.py

@@ -1,6 +1,7 @@
 import pytest
 import datetime
 import kenpompy.team as kpteam
+import kenpompy.misc as kpmisc
 import pandas as pd
 
 def test_get_valid_teams(browser):
@@ -37,7 +38,7 @@ def test_get_schedule(browser):
 	assert df.shape == (34, 9)
 
 	date = datetime.date.today()
-	currentYear = date.strftime("%Y")
+	currentYear = kpmisc.get_current_season(browser)
 	nextYear = str(int(currentYear)+1)
 
 	with pytest.raises(ValueError):