Issues - fizx/parsley

fizx / parsley

Description:

Parsley is a simple language for extracting structured data from web pages. Parsley consists of an powerful selector language wrapped with a JSON structure that can represent page-wide formatting.

Sort by: Priority | Votes | Last Updated

Loading…

1.

0 votes

OR selectors together
1 comment Created about 1 month ago by amccloud

Title

Body
One of the features that I wanted to suggest is some sort of logical OR operator. For example, http://www.bestbuy.com/site/Swiss+Gear+-+SHERPA+Laptop+Sleeve+-+Black/Blue/8819264.p?skuId=8819264&id=1206750211247 I want the price. I'd like to have the sale price and if none is provided the original price. I've tried 3 diff ways to achieve this. The first, which I believe would probably be the best syntax for this to be implemented, comma delimited selectors. `{ "price": "#productpreview .salenum, #productpreview .pricenum" }` Then I tried `{ "price?": "#productpreview .salenum", "price": "#productpreview .pricenum" }` finally because my crawling software (Scrapy) takes the first object from the list `{ "price": ["#productpreview .salenum, #productpreview .pricenum"] }` ... but order is not preserved so that isn't reliable. If order was preserved this would be my second choice. Any thoughts? or has this already been implemented? Andrew
or cancel

One of the features that I wanted to suggest is some sort of logical OR operator.

For example,
http://www.bestbuy.com/site/Swiss+Gear+-+SHERPA+Laptop+Sleeve+-+Black/Blue/8819264.p?skuId=8819264&id=1206750211247

I want the price. I'd like to have the sale price and if none is provided the original price. I've tried 3 diff ways to achieve this. The first, which I believe would probably be the best syntax for this to be implemented, comma delimited selectors.

{ "price": "#productpreview .salenum, #productpreview .pricenum" }

Then I tried
{ "price?": "#productpreview .salenum", "price": "#productpreview .pricenum" }

finally because my crawling software (Scrapy) takes the first object from the list
{ "price": ["#productpreview .salenum, #productpreview .pricenum"] }

... but order is not preserved so that isn't reliable. If order was preserved this would be my second choice.

Any thoughts? or has this already been implemented?

Andrew

Comments

fizx Mon Nov 02 10:05:09 -0800 2009 | link

Right, so the logical OR operators (css comma or pipe in xpath), maintain order, but with respect to the page. So the original price will always come first, because it's higher on the page.

This is getting to be a little to much logic for Parsley, which was never designed to be Turing-complete. I can see implementing a reverse() function, which would give you a workaround here. My approach to this problem is reflected in the csvget project, where each row is optionally post-processed in Ruby.

Comment
Right, so the logical OR operators (css comma or pipe in xpath), maintain order, but with respect to the page. So the original price will always come first, because it's higher on the page. This is getting to be a little to much logic for Parsley, which was never designed to be Turing-complete. I can see implementing a reverse() function, which would give you a workaround here. My approach to this problem is reflected in the csvget project, where each row is optionally post-processed in Ruby.
or cancel

Parsed with GitHub Flavored Markdown

Labels

OR selectors together

Comments