Skip to content

Instantly share code, notes, and snippets.

@psychemedia
Created April 10, 2011 21:46
  • Star 6 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
Star You must be signed in to star a gist
Save psychemedia/912758 to your computer and use it in GitHub Desktop.
Scraperwiki script for scraping data from FIA/F1 timing sheets/PDFs
import scraperwiki
import urllib2
import lxml.etree
'''
Code to pull data out of the timing related press releases issued by FIA for Formula One races.
This code is provided solely for your own use, without guarantee, so you can publish F1 timing data,
according to license conditions specified by FIA,
without having to rekey the timing data as published on the PDF press releases yourself.
If you want to run the code in your own Python environment, you can what the pdftoxml function calls here:
https://bitbucket.org/ScraperWiki/scraperwiki/src/7d6c7a5393ed/scraperlibs/scraperwiki/utils.py
Essentially, it seems to be a call to the binary /usr/bin/pdftohtml ? [h/t @frabcus]
??pdf2html - this one? http://sourceforge.net/projects/pdftohtml/
'''
'''
To run the script, you need to provide a couple of bits of info...
Check out the PDF URLs on the F1 Media Centre timings page:
http://www.fia.com/en-GB/mediacentre/f1_media/Pages/timing.aspx
You should see a common slug identifying the race:
'''
#Enter slug for race here
race='chn'
#chn, mal, aus
'''
...and then something relevant for the rest of the filename
'''
#enter slug for timing sheet here
typ='race-analysis'
#typ can be any of the following (if they use the same convention each race...)
'''
session1-classification.+
session1-times.x
session2-classification.+
session2-times.x
session3-classification.+
session3-times.x
x qualifying-classification
qualifying-trap.+
qualifying-speeds.+
qualifying-sectors.+
qualifying-times.x
race-laps.+
race-speeds.+
race-sectors.+
race-trap.+
race-analysis.x
race-summary.+
race-history.+
race-chart.+
**Note that race-analysis and *-times may have minor glitches**
The report list is a bit casual and occasionally a lap mumber is omitted and appears at the end of the list
A tidying pass on the data that I'm reporting is probably required...
BROKEN (IMPOSSIBLE AFTER SIGNING?)
race-classification <- under development; getting null response? Hmmm -seems to have turned to an photocopied image?
qualifying-classification <- seems like this gets signed and photocopied too :-(
IMPOSSIBLE?
race-grid
'''
#only go below here if you need to do maintenance on the script...
#...which you will have to do in part to get the data out in a usable form
#...at the moment, I only go so far as to preview what's there
#Here's where we construct the URL for the timing sheet.
#I assume a similar naming convention is used for each race?
url = "http://www.fia.com/en-GB/mediacentre/f1_media/Documents/"+race+"-"+typ+".pdf"
pdfdata = urllib2.urlopen(url).read()
print "The pdf file has %d bytes" % len(pdfdata)
xmldata = scraperwiki.pdftoxml(pdfdata)
'''
print "After converting to xml it has %d bytes" % len(xmldata)
print "The first 2000 characters are: ", xmldata[:2000]
'''
root = lxml.etree.fromstring(xmldata)
pages = list(root)
print "The pages are numbered:", [ page.attrib.get("number") for page in pages ]
def tidyup(txt):
txt=txt.strip()
txt=txt.strip('\n')
txt=txt.strip('<b>')
txt=txt.strip('</b>')
return txt
def gettext_with_bi_tags(el):
res = [ ]
if el.text:
res.append(el.text)
for lel in el:
res.append("<%s>" % lel.tag)
res.append(gettext_with_bi_tags(lel))
res.append("</%s>" % lel.tag)
if el.tail:
res.append(el.tail)
return "".join(res)
#I use the stub() routine to preview the raw scrape for new documents...
def stub():
page = pages[0]
scraping=1
for el in list(page)[:200]:
if el.tag == "text":
if scraping:
print el.attrib,gettext_with_bi_tags(el)
#The scraper functions themselves
#I just hope the layout of the PDFs, and the foibles, are the same for all races!
def race_history():
lapdata=[]
txt=''
for page in pages:
lapdata=race_history_page(page,lapdata)
txt=txt+'new page'+str(len(lapdata))+'\n'
#Here's the data
for lap in lapdata:
print lap
print lapdata
print txt
print 'nlaps timing',str(len(lapdata))
def race_history_page(page,lapdata=[]):
scraping=0
cnt=0
cntz=[2,2]
laps={}
lap=''
results=[]
microresults=[]
headphase=0
phase=0
pos=1
for el in list(page):
if el.tag == "text":
if scraping:
#print el.attrib,gettext_with_bi_tags(el)
txt=tidyup(gettext_with_bi_tags(el))
if txt.startswith("LAP") or txt.startswith("Page"):
if lap!='' and microresults!=[]:
results.append(microresults)
laps[lap]=results
lapdata.append(results)
pos=2
else:
print ';;;;'
pos=1
lap=txt
headphase=1
results=[]
results.append(txt.split(' ')[1])
microresults=[]
cnt=0
if headphase==1 and txt.startswith("TIME"):
headphase=0
elif headphase==0:
if cnt<cntz[phase] or (pos==1 and txt=='PIT'):
microresults.append(txt)
cnt=cnt+1
else:
cnt=0
results.append(microresults)
#print microresults,phase,cnt,headphase,pos,'....'
microresults=[txt]
if phase==0:
phase=1
else:
txt=gettext_with_bi_tags(el)
txt=txt.strip()
if txt.startswith("<b>2011 FORMULA 1"):
scraping=1
#print laps
return lapdata
def race_chart():
laps=[]
for page in pages:
laps=race_chart_page(page,laps)
#Here's the data
for lap in laps:
print lap
print laps
def race_chart_page(page,laps):
cnt=0
cntz=[2,2]
scraping=0
lap=''
results=[]
headphase=0
phase=0
pos=1
for el in list(page):
if el.tag == "text":
if scraping:
#print el.attrib,gettext_with_bi_tags(el)
txt=tidyup(gettext_with_bi_tags(el))
if txt.startswith("GRID"):
lap=txt
results=[txt]
elif txt.startswith("LAP"):
if lap !='':
laps.append(results)
lap=txt
results=[txt]
elif txt.startswith("Page"):
laps.append(results)
else:
results.append(txt)
else:
txt=gettext_with_bi_tags(el)
txt=txt.strip()
if txt.startswith("<b>2011 FORMULA 1"):
scraping=1
#print laps
return laps
def race_summary():
stops=[]
for page in pages:
stops=race_summary_page(page,stops)
#Here's the data
for stop in stops:
print stop
print stops
def race_summary_page(page,stops=[]):
scraping=0
cnt=0
cntz=6
results=[]
pos=1
for el in list(page):
if el.tag == "text":
if scraping:
#print el.attrib,gettext_with_bi_tags(el)
txt=gettext_with_bi_tags(el)
if cnt<cntz:
if cnt==0:
results.append([])
txt=txt.split("<b>")
for t in txt:
if t !='':
results[pos-1].append(tidyup(t))
cnt=cnt+1
else:
cnt=0
txt=txt.split("<b>")
for t in txt:
results[pos-1].append(tidyup(t))
#print pos,results[pos-1]
pos=pos+1
else:
txt=gettext_with_bi_tags(el)
txt=txt.strip()
if txt.startswith("<b>2011 FORMULA 1"):
scraping=1
for result in results:
if not result[0].startswith("Page"):
stops.append(result)
return stops
def qualifying_times():
pos=1
dpos=[]
#pos,dpos=qualifying_times_page(pages[0],pos,dpos)
for page in pages:
pos,dpos=qualifying_times_page(page,pos,dpos)
#Here's the data
for pos in dpos:
print pos
dposcorr=[]
for pos in dpos:
dupe=[]
print pos
prev=0
fixed=0
for p in pos:
if p.count(':')>0:
if prev==1:
print "oops - need to do a shuffle here and insert element at [-1] here"
dupe.append(pos[-1])
prev=1
else:
prev=0
if len(dupe)<len(pos):
dupe.append(p)
print 'corr?',dupe
print dposcorr.append(dupe)
print dpos
print 'hackfix',dposcorr
def linebuffershuffle(oldbuffer, newitem):
oldbuffer[2]=oldbuffer[1]
oldbuffer[1]=oldbuffer[0]
oldbuffer[0]=newitem
return oldbuffer
def qualifying_times_page(page,pos,dpos):
#There are still a few issues with this one:
#Some of the lap numbers appear in the wrong position in results list
scraping=0
cnt=0
cntz=5
drivers=[]
results=[]
phase=0
linebuffer=["","",""]
for el in list(page):
if el.tag == "text":
txt=gettext_with_bi_tags(el)
txt=tidyup(txt)
items=txt.split(" <b>")
for item in items:
linebuffer=linebuffershuffle(linebuffer, item)
if scraping:
#print el.attrib,gettext_with_bi_tags(el)
if phase==0 and txt.startswith("NO"):
phase=1
cnt=0
results=[]
print linebuffer
results.append(linebuffer[2])
results.append(linebuffer[1])
elif phase==1 and cnt<3:
cnt=cnt+1
elif phase==1:
phase=2
results.append(txt)
elif phase==2 and txt.startswith("NO"):
phase=1
#print results,linebuffer[2],linebuffer[1]
results.remove(linebuffer[2])
results.remove(linebuffer[1])
#print '>>>',pos,results
dpos.append(results)
pos=pos+1
drivers.append(results)
results=[]
cnt=0
results.append(linebuffer[2])
results.append(linebuffer[1])
elif phase==2 and txt.startswith("Page"):
#print '>>>',pos,results
dpos.append(results)
drivers.append(results)
pos=pos+1
elif phase==2:
items=txt.split(" <b>")
for item in items:
results.append(item)
else:
txt=gettext_with_bi_tags(el)
txt=txt.strip()
if txt.startswith("<b>2011 FORMULA 1"):
scraping=1
return pos,dpos
def race_analysis():
pos=1
dpos=[]
dposcorr=[]
for page in pages:
pos,dpos=race_analysis_page(page,pos,dpos)
#Here's the data
for pos in dpos:
print pos
dupe=[]
prev=0
fixed=0
for p in pos:
if p.count(':')>0:
if prev==1:
print "oops - need to do a shuffle here and insert element at [-1] here"
dupe.append(pos[-1])
prev=1
else:
prev=0
if len(dupe)<len(pos):
dupe.append(p)
print dupe
print dposcorr.append(dupe)
print dpos
print dposcorr
def race_analysis_page(page,pos,dpos):
#There are still a few issues with this one:
#Some of the lap numbers appear in the wrong position in results list
scraping=0
cnt=0
cntz=5
drivers=[]
results=[]
phase=0
linebuffer=["","",""]
for el in list(page):
if el.tag == "text":
txt=gettext_with_bi_tags(el)
txt=tidyup(txt)
items=txt.split(" <b>")
for item in items:
linebuffer=linebuffershuffle(linebuffer, item)
if scraping:
#print el.attrib,gettext_with_bi_tags(el)
if phase==0 and txt.startswith("LAP"):
phase=1
cnt=0
results=[]
results.append(linebuffer[2])
results.append(linebuffer[1])
elif phase==1 and cnt<3:
cnt=cnt+1
elif phase==1:
phase=2
results.append(txt)
elif phase==2 and txt.startswith("LAP"):
phase=1
#print results,linebuffer[2],linebuffer[1]
results.remove(linebuffer[2])
results.remove(linebuffer[1])
#print '>>>',pos,results
dpos.append(results)
pos=pos+1
drivers.append(results)
results=[]
cnt=0
results.append(linebuffer[2])
results.append(linebuffer[1])
elif phase==2 and txt.startswith("Page"):
#print '>>>',pos,results
dpos.append(results)
drivers.append(results)
pos=pos+1
elif phase==2:
items=txt.split(" <b>")
for item in items:
results.append(item)
else:
txt=gettext_with_bi_tags(el)
txt=txt.strip()
if txt.startswith("<b>2011 FORMULA 1"):
scraping=1
return pos,dpos
def session1_classification():
page = pages[0]
scraping=0
cnt=0
cntz=[7,8,9]
results=[]
pos=1
phase=0
for el in list(page):
if el.tag == "text":
txt=gettext_with_bi_tags(el)
if scraping:
#print el.attrib,gettext_with_bi_tags(el)
txt=tidyup(txt)
if cnt<cntz[phase]:
if cnt==0:
results.append([])
txt=txt.split("<b>")
for t in txt:
results[pos-1].append(t.strip())
cnt=cnt+1
else:
if phase<2:
phase=phase+1
cnt=0
results[pos-1].append(txt)
#print pos,results[pos-1]
pos=pos+1
else:
txt=gettext_with_bi_tags(el)
txt=txt.strip()
if txt.startswith("<b>TIME OF"):
scraping=1
#Here is the data
for pos in results:
print pos
def qualifying_sectors():
sectors=["<b>SECTOR 1</b>\n","<b>SECTOR 2</b>\n","<b>SECTOR 3</b>\n"]
sector=1
scraping=0
results=[]
sectorResults=[]
pos=1
cnt=0
cntz=2
for el in list(page):
if el.tag == "text":
if scraping:
#print el.attrib,gettext_with_bi_tags(el)
txt=gettext_with_bi_tags(el)
if txt in sectors:
sector=sector+1
sectorResults.append(results)
#print sectorResults
#print "Next sector"
scraping=0
continue
if cnt<cntz:
if cnt==0:
results.append([])
txt=txt.strip()
txt=txt.split("<b>")
for t in txt:
t=tidyup(t)
results[pos-1].append(t)
cnt=cnt+1
else:
cnt=0
txt=txt.strip()
txt=txt.split("<b>")
for t in txt:
t=tidyup(t)
results[pos-1].append(t)
#print pos,results[pos-1]
pos=pos+1
else:
txt=gettext_with_bi_tags(el)
txt=txt.strip()
if txt.startswith("<b>TIME"):
scraping=1
results=[]
pos=1
cnt=0
sectorResults.append(results)
#print sectorResults
#Here's the data
for result in sectorResults:
print result
def qualifying_speeds():
sessions=["<b>INTERMEDIATE 1</b>\n","<b>INTERMEDIATE 2</b>\n","<b>FINISH LINE</b>\n"]
session=1
scraping=0
results=[]
sessionResults=[]
pos=1
cnt=0
cntz=2
for el in list(page):
if el.tag == "text":
if scraping:
#print el.attrib,gettext_with_bi_tags(el)
txt=gettext_with_bi_tags(el)
if txt in sessions:
session=session+1
sessionResults.append(results)
#print sessionResults
#print "Next session"
scraping=0
continue
if cnt<cntz:
if cnt==0:
results.append([])
txt=txt.strip()
txt=txt.split("<b>")
for t in txt:
t=tidyup(t)
results[pos-1].append(t)
cnt=cnt+1
else:
cnt=0
txt=txt.strip()
txt=txt.split("<b>")
for t in txt:
txt=tidyup(t)
results[pos-1].append(t)
#print pos,results[pos-1]
pos=pos+1
else:
txt=gettext_with_bi_tags(el)
txt=txt.strip()
if txt.startswith("<b>KPH"):
scraping=1
results=[]
pos=1
cnt=0
sessionResults.append(results)
#Here's the data
for session in sessionResults:
for pos in session:
print pos
def qualifying_trap():
page = pages[0]
scraping=0
cnt=0
cntz=3
results=[]
pos=1
for el in list(page):
if el.tag == "text":
if scraping:
#print el.attrib,gettext_with_bi_tags(el)
txt=gettext_with_bi_tags(el)
if cnt<cntz:
if cnt==0:
results.append([])
txt=txt.split("<b>")
for t in txt:
results[pos-1].append(tidyup(t))
cnt=cnt+1
else:
cnt=0
txt=txt.split("<b>")
for t in txt:
results[pos-1].append(tidyup(t))
#print pos,results[pos-1]
pos=pos+1
else:
txt=gettext_with_bi_tags(el)
txt=txt.strip()
if txt.startswith("<b>TIME OF"):
scraping=1
#Here's the data
for pos in results:
print pos
def qualifying_classification():
# print the first hundred text elements from the first page
page0 = pages[0]
scraping=0
session=1
cnt=0
pos=1
results=[]
cntz=[12,9,6]
posz=[10,17,24]
for el in list(page):
if el.tag == "text":
if scraping:
#print el.attrib,gettext_with_bi_tags(el)
txt=gettext_with_bi_tags(el)
if session<4:
if cnt<cntz[session-1]:
if cnt==0:
results.append([])
txt=txt.strip()
txt=txt.split(" ")
for j in txt:
results[pos-1].append(j)
cnt=cnt+1
else:
results[pos-1].append(txt)
cnt=cnt+1
else:
if pos==posz[session-1]:
session=session+1
#print "session",session
cnt=0
results[pos-1].append(txt)
#print pos,results[pos-1]
pos=pos+1
else:
txt=gettext_with_bi_tags(el)
txt=txt.strip()
if txt.startswith("<b>2011"):
scraping=1
#Here's the data
for result in results:
print result
def race_classification():
#under development
page = pages[0]
scraping=0
cnt=0
cntz=[8,9,10,8]
results=[]
pos=1
phase=0
for el in list(page):
print "broken?",el
if el.tag == "text":
txt=gettext_with_bi_tags(el)
if scraping:
#print el.attrib,gettext_with_bi_tags(el)
txt=tidyup(txt)
if cnt<cntz[phase]:
if cnt==0:
results.append([])
txt=txt.split("<b>")
for t in txt:
results[pos-1].append(t.strip())
cnt=cnt+1
else:
if phase<2:
phase=phase+1
cnt=0
if txt.startswith("NOT CLASS"):
phase=3
else:
results[pos-1].append(txt)
print pos,results[pos-1]
pos=pos+1
else:
txt=gettext_with_bi_tags(el)
txt=txt.strip()
print "...",txt
if txt.startswith("<b>LAP<"):
scraping=1
if typ=="qualifying-classification":
qualifying_classification()
elif typ=="qualifying-trap" or typ=="race-trap":
qualifying_trap()
elif typ=="qualifying-speeds" or typ=="race-speeds":
qualifying_speeds()
elif typ=="qualifying-sectors" or typ=="race-sectors":
qualifying_sectors()
elif typ=="session1-classification" or typ=="session2-classification" or typ=="session3-classification" or typ=="race-laps":
session1_classification()
if typ=="race-classification":
race_classification()
elif typ=="qualifying-times" or typ=="session3-times" or typ=="session2-times" or typ=="session1-times":
print "Trying qualifying times"
qualifying_times()
if typ=="race-analysis":
race_analysis()
elif typ=="race-summary":
race_summary()
elif typ=="race-history":
race_history()
elif typ=="race-chart":
race_chart()
# If you have many PDF documents to extract data from, the trick is to find what's similar
# in the way that the information is presented in them in terms of the top left bottom right
# pixel locations. It's real work, but you can use the position visualizer here:
# http://scraperwikiviews.com/run/pdf-to-html-preview-1/
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment