lxml/Python : get previous-sibling

I have the following html: ``` <div id = "big"> <span>header 1</span> <ul id = "outer"> <li id = "inner">aaa</li> <li id = "inner">bbb</li> </ul> <span>header 2</span> <ul id = "outer"> <li id = "inner">ccc</li> <li id = "inner">ddd</li> </ul> </div> ``` I want it to loop it in the order: ``` header 1 aaa bbb header 2 ccc ddd ``` I have tried looping through each ul and then printing the header and the li values. However, I don't know how to get the span header that is associated with a ul. ``` sets = tree.xpath("//div[@id='big']//ul[@id='outer']") for set in sets: # Print header. Not sure how to get it header = set.xpath(".//li/preceding-sibling::span") print header # Print texts. This works. values = set.xpath(".//li//text()") for v in values: print v ``` Just looping all text nodes won't work because I need to know if it is a header or li value.