converting scrapy to lxml

Question

I have scrapy code that looks like this

for row in response.css("div#flexBox_flex_calendar_mainCal table tr.calendar_row"):
    print "================" 
        print row.xpath(".//td[@class='time']/text()").extract()
        print row.xpath(".//td[@class='currency']/text()").extract()
        print row.xpath(".//td[@class='impact']/span/@title").extract()
        print row.xpath(".//td[@class='event']/span/text()").extract()
        print row.xpath(".//td[@class='actual']/text()").extract()
        print row.xpath(".//td[@class='forecast']/text()").extract()
        print row.xpath(".//td[@class='previous']/text()").extract()
    print "================" 

I am able to get the same stuff using pure python like this,

from lxml import html
import requests

page = requests.get('http://www.forexfactory.com/calendar.php?day=dec1.2011')

tree = html.fromstring(page.text)

print tree.xpath(".//td[@class='time']/text()")
print tree.xpath(".//td[@class='currency']/text()")
print tree.xpath(".//td[@class='impact']/span/@title")
print tree.xpath(".//td[@class='event']/span/text()")
print tree.xpath(".//td[@class='actual']/text()")
print tree.xpath(".//td[@class='forecast']/text()")
print tree.xpath(".//td[@class='previous']/text()")

However I need to do this row by row. My first attempt to port to lxml doesn't work:

from lxml import html
import requests

page = requests.get('http://www.forexfactory.com/calendar.php?day=dec1.2011')

tree = html.fromstring(page.text)

for row in tree.css("div#flexBox_flex_calendar_mainCal table tr.calendar_row"):
    print row.xpath(".//td[@class='time']/text()")
    print row.xpath(".//td[@class='currency']/text()")
    print row.xpath(".//td[@class='impact']/span/@title")
    print row.xpath(".//td[@class='event']/span/text()")
    print row.xpath(".//td[@class='actual']/text()")
    print row.xpath(".//td[@class='forecast']/text()")
    print row.xpath(".//td[@class='previous']/text()")

What is the correct way to port this scrapy code to pure lxml?

EDIT: I have gotten a little closer. I can see a table{} object, I just don't know how to walk it.

import urllib2
from lxml import etree


#import requests

def wgetUrl(target):
    try:
        req = urllib2.Request(target)
        req.add_header('User-Agent', 'Mozilla/5.0 (Windows; U; Windows NT 5.1; en-GB; rv:1.9.0.3 Gecko/2008092417 Firefox/3.0.3')
        response = urllib2.urlopen(req)
        outtxt = response.read()
        response.close()
    except:
        return ''

    return outtxt


url = 'http://www.forexfactory.com/calendar.php?day='
date = 'dec1.2011'

data = wgetUrl(url + date)
parser = etree.HTMLParser()

tree   = etree.fromstring(data, parser)

for elem in tree.xpath("//div[@id='flexBox_flex_calendar_mainCal']"):
    print elem[0].tag, elem[0].attrib, elem[0].text
    # elem[1] is where the table is
    print elem[1].tag, elem[1].attrib, elem[1].text
    print elem[1]

Answer 1

I like to use lxml for scraping. I usually do not use its xpath functionality though and opt for their ElementPath library instead. It is very similar in syntax. Below is how I would port your scrapy code.

Going line by line:

initialization:

from lxml import etree

# analogous function xpath(.../text()).extract() for lxml etree nodes
def extract_text(elem):        
    if elem is None:
        print None
    else
        return ''.join(i for i in elem.itertext())

data = wgetUrl(url+date)  # wgetUrl, url, date you defined in your question
tree = etree.HTML(content)

line 1

# original
for row in response.css("div#flexBox_flex_calendar_mainCal table tr.calendar_row"):

# ported
for row in tree.findall(r'.//div[@id="flexBox_flex_calendar_mainCal"]//table/tr[@class="calendar_row"]'):

line 2

print "================" 

line 3

# original
print row.xpath(".//td[@class='time']/text()").extract()
# ported
print extract_text(row.find(r'.//td[@class="time"]'))

line 4

# original
print row.xpath(".//td[@class='currency']/text()").extract()
# ported
print extract_text(row.find(r'.//td[@class="currency"]'))

line 5

# original
print row.xpath(".//td[@class='impact']/span/@title").extract()
# ported
td = row.find(r'.//td[@class="impact"]/span')
if td is not None and 'title' in td.attrib:
    print td.attrib['title']

line 6

# original
print row.xpath(".//td[@class='event']/span/text()").extract()
# ported
print extract_text(row.find(r'.//td[@class="event"]/span'))

line 7

# original
print row.xpath(".//td[@class='actual']/text()").extract()
# ported
print extract_text(row.find(r'.//td[@class="actual"]'))

line 8

# original
print row.xpath(".//td[@class='forecast']/text()").extract()
# ported
print extract_text(row.find(r'.//td[@class="forecast"]'))

line 9

# original
print row.xpath(".//td[@class='previous']/text()").extract()
# ported
print extract_text(row.find(r'.//td[@class="previous"]'))

line 10

print "================" 

And all together now:

from lxml import etree

def wgetUrl(target):
    # same as you defined it

# analogous function xpath(.../text()).extract() for lxml etree nodes
def extract_text(elem):        
    if elem is None:
        print None
    else
        return ''.join(i for i in elem.itertext())

content = wgetUrl(your_url)  # wgetUrl as the function you defined in your question
node = etree.HTML(content)


for row in node.findall(r'.//div[@id="flexBox_flex_calendar_mainCal"]//table/tr[@class="calendar_row"]'):
    print "================" 
    print extract_text(row.find(r'.//td[@class="time"]'))
    print extract_text(row.find(r'.//td[@class="currency"]'))
    td = row.find(r'.//td[@class="impact"]/span')
    if td is not None and 'title' in td.attrib:
        print td.attrib['title']
    print extract_text(row.find(r'.//td[@class="event"]/span'))
    print extract_text(row.find(r'.//td[@class="actual"]'))
    print extract_text(row.find(r'.//td[@class="forecast"]'))
    print extract_text(row.find(r'.//td[@class="previous"]'))
    print "================"