Last active
September 28, 2019 16:38
-
-
Save hamletbatista/5f3b61008fa978f66ff75ecf21af85f8 to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
script_sel="#new_gchart_slideshow > div.forecast-viewport > ul > li > script" | |
all_rides = r.html.find(script_sel) | |
def extract_dates(data): | |
start_row = "dateData.addRows\(" | |
end_row = "\);" | |
columns = [] | |
title = None | |
results = re.search(start_row+"([^;]+)"+end_row, data) | |
if results != None: | |
columns = results.group(1) | |
start_title = "title: \"" | |
end_title = "\"," | |
results = re.search(start_title+'([^"]+)'+end_title, data) | |
if results != None: | |
title = results.group(1) | |
return (title, columns) | |
#all_rides raw extracted text looks like this | |
print(all_rides[0].text) | |
#'$(function() { google.load("visualization", "1", {packages:["corechart"], callback: drawChart137}); }); function drawChart137() { // Create and populate the data table. var dateData = new google.visualization.DataTable(); dateData.addColumn(\'datetime\', \'time\'); //dateData.addColumn(\'number\', \'null\'); dateData.addColumn(\'number\', \'Wait Times the Crowd Calendar Predicted\'); dateData.addRows([[new Date( 2019,6,15,08,00,00 ), 4],[new Date( 2019,6,15,08,15,00 ), 4],[new Date( 2019,6,15,08,30,00 ), 4],[new Date( 2019,6,15,08,45,00 ), 4],[new Date( 2019,6,15,09,00,00 ), 12],[new Date( 2019,6,15,09,15,00 ), 6],[new Date( 2019,6,15,09,30,00 ), 7],[new Date( 2019,6,15,09,45,00 ), 14],[new Date( 2019,6,15,10,00,00 ), 16],[new Date( 2019,6,15,10,15,00 ), 16],[new Date( 2019,6,15,10,30,00 ), 16],[new Date( 2019,6,15,10,45,00 ), 16],[new Date( 2019,6,15,11,00,00 ), 16],[new Date( 2019,6,15,11,15,00 ), 23],[new Date( 2019,6,15,11,30,00 ), 21],[new Date( 2019,6,15,11,45,00 ), 16],[new Date( 2019,6,15,12,00,00 ), 26],[new Date( 2019,6,15,12,15,00 ), 26],[new Date( 2019,6,15,12,30,00 ), 26],[new Date( 2019,6,15,12,45,00 ), 25],[new Date( 2019,6,15,13,00,00 ), 25],[new Date( 2019,6,15,14,30,00 ), 18],[new Date( 2019,6,15,14,45,00 ), 23],[new Date( 2019,6,15,15,00,00 ), 25],[new Date( 2019,6,15,15,15,00 ), 20],[new Date( 2019,6,15,15,30,00 ), 18],[new Date( 2019,6,15,15,45,00 ), 16],[new Date( 2019,6,15,16,00,00 ), 16],[new Date( 2019,6,15,16,15,00 ), 18],[new Date( 2019,6,15,16,30,00 ), 18],[new Date( 2019,6,15,16,45,00 ), 18],[new Date( 2019,6,15,17,00,00 ), 18],[new Date( 2019,6,15,17,15,00 ), 19],[new Date( 2019,6,15,17,30,00 ), 19],[new Date( 2019,6,15,17,45,00 ), 19],[new Date( 2019,6,15,18,00,00 ), 19],[new Date( 2019,6,15,18,15,00 ), 20],[new Date( 2019,6,15,18,30,00 ), 20],[new Date( 2019,6,15,18,45,00 ), 24],[new Date( 2019,6,15,19,00,00 ), 24],[new Date( 2019,6,15,19,15,00 ), 25],[new Date( 2019,6,15,19,30,00 ), 21],[new Date( 2019,6,15,19,45,00 ), 21],[new Date( 2019,6,15,20,00,00 ), 21],[new Date( 2019,6,15,20,15,00 ), 21],[new Date( 2019,6,15,20,30,00 ), 21],[new Date( 2019,6,15,20,45,00 ), 21],[new Date( 2019,6,15,21,00,00 ), 21],[new Date( 2019,6,15,21,15,00 ), 21],[new Date( 2019,6,15,21,30,00 ), 21],[new Date( 2019,6,15,21,45,00 ), 21],[new Date( 2019,6,15,22,00,00 ), 21],[new Date( 2019,6,15,22,15,00 ), 25],[new Date( 2019,6,15,22,30,00 ), 12],[new Date( 2019,6,15,22,45,00 ), 12],[new Date( 2019,6,15,23,00,00 ), 12],[new Date( 2019,6,15,23,15,00 ), 12],[new Date( 2019,6,15,23,30,00 ), 12],[new Date( 2019,6,15,23,45,00 ), 12]]); var userOptions = { title: "Alice in Wonderland - 7/15/19", series: [ // CC Predicted forecasts {color: \'blue\', pointSize: 0, visibleInLegend: true, lineWidth:2 }, ], width: $(\'div#center.column\').width(), height: 320, chartArea:{left:30,top:50,width:"80%",height:"70%"}, legend: {position: \'top\', textStyle: {color: \'black\', fontSize: 11}}, fontSize:10, hAxis: { slantedText: true, slantedTextAngle: 45, viewWindowMode:\'pretty\', format: \'h aa\', maxValue: new Date(2019,6,16,00,00,00) }, vAxis: {format: \'0\', title: \'minutes\', maxValue: 31} // Nothing specified for axis 0 }; var userChart = new google.visualization.AreaChart(document.getElementById(\'google_chart_137\')); userChart.draw(dateData, userOptions); } $("#new_gchart_slideshow").width( $("div#center.column").width() );' | |
#After we process it, we get a cleaner dataset | |
print(extract_dates(all_rides[0].text)) | |
#('Alice in Wonderland - 7/15/19', | |
#'[[new Date( 2019,6,15,08,00,00 ), 4],[new Date( 2019,6,15,08,15,00 ), 4],[new Date( 2019,6,15,08,30,00 ), 4],[new Date( 2019,6,15,08,45,00 ), 4],[new Date( 2019,6,15,09,00,00 ), 12],[new Date( 2019,6,15,09,15,00 ), 6],[new Date( 2019,6,15,09,30,00 ), 7],[new Date( 2019,6,15,09,45,00 ), 14],[new Date( 2019,6,15,10,00,00 ), 16],[new Date( 2019,6,15,10,15,00 ), 16],[new Date( 2019,6,15,10,30,00 ), 16],[new Date( 2019,6,15,10,45,00 ), 16],[new Date( 2019,6,15,11,00,00 ), 16],[new Date( 2019,6,15,11,15,00 ), 23],[new Date( 2019,6,15,11,30,00 ), 21],[new Date( 2019,6,15,11,45,00 ), 16],[new Date( 2019,6,15,12,00,00 ), 26],[new Date( 2019,6,15,12,15,00 ), 26],[new Date( 2019,6,15,12,30,00 ), 26],[new Date( 2019,6,15,12,45,00 ), 25],[new Date( 2019,6,15,13,00,00 ), 25],[new Date( 2019,6,15,14,30,00 ), 18],[new Date( 2019,6,15,14,45,00 ), 23],[new Date( 2019,6,15,15,00,00 ), 25],[new Date( 2019,6,15,15,15,00 ), 20],[new Date( 2019,6,15,15,30,00 ), 18],[new Date( 2019,6,15,15,45,00 ), 16],[new Date( 2019,6,15,16,00,00 ), 16],[new Date( 2019,6,15,16,15,00 ), 18],[new Date( 2019,6,15,16,30,00 ), 18],[new Date( 2019,6,15,16,45,00 ), 18],[new Date( 2019,6,15,17,00,00 ), 18],[new Date( 2019,6,15,17,15,00 ), 19],[new Date( 2019,6,15,17,30,00 ), 19],[new Date( 2019,6,15,17,45,00 ), 19],[new Date( 2019,6,15,18,00,00 ), 19],[new Date( 2019,6,15,18,15,00 ), 20],[new Date( 2019,6,15,18,30,00 ), 20],[new Date( 2019,6,15,18,45,00 ), 24],[new Date( 2019,6,15,19,00,00 ), 24],[new Date( 2019,6,15,19,15,00 ), 25],[new Date( 2019,6,15,19,30,00 ), 21],[new Date( 2019,6,15,19,45,00 ), 21],[new Date( 2019,6,15,20,00,00 ), 21],[new Date( 2019,6,15,20,15,00 ), 21],[new Date( 2019,6,15,20,30,00 ), 21],[new Date( 2019,6,15,20,45,00 ), 21],[new Date( 2019,6,15,21,00,00 ), 21],[new Date( 2019,6,15,21,15,00 ), 21],[new Date( 2019,6,15,21,30,00 ), 21],[new Date( 2019,6,15,21,45,00 ), 21],[new Date( 2019,6,15,22,00,00 ), 21],[new Date( 2019,6,15,22,15,00 ), 25],[new Date( 2019,6,15,22,30,00 ), 12],[new Date( 2019,6,15,22,45,00 ), 12],[new Date( 2019,6,15,23,00,00 ), 12],[new Date( 2019,6,15,23,15,00 ), 12],[new Date( 2019,6,15,23,30,00 ), 12],[new Date( 2019,6,15,23,45,00 ), 12]]') | |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment