Class | Scrubyt::BaseFilter |
In: |
lib/scrubyt/core/scraping/filters/base_filter.rb
|
Parent: | Object |
A Scrubyt extractor is almost like a waterfall: water is pouring from the top until it reaches the bottom. The biggest difference is that instead of water, a HTML document travels through the space.
Of course Scrubyt would not make much sense if the same document would arrive at the bottom that was poured in at the top - since in this case we might use an indentity transformation (i.e. do nothing with the input) as well.
This is where filters came in: as they name says, they filter the stuff that is pouring from above, to leave the interesting parts and discard the rest. The working of a filter will be explained most easily by the help of an example. Let‘s consider that we would like to extract information from a webshop; Concretely we are interested in the name of the items and the URL pointing to the image of the item.
To accomplish this, first we select the items with the pattern item (a pattern is a logical grouping of fillters; see Pattern documentation) Then our new context is the result extracted by the ‘item’ pattern; For every ‘item’ pattern, further extract the name and the image of the item; and finally, extract the href attribute of the image. Let‘s see an illustration:
root --> This pattern is called a 'root pattern', It is invisible to you | and basically it represents the document; it has no filters +-- item --> Filter what's coming from above (the whole document) to get | relevant pieces of data (in this case webshop items) +-- name --> Again, filter what's coming from above (a webshop item) and | leave only item names after this operation +-- image --> This time filter the image of the item | +-- href --> And finally, from the image elements, get the attribute 'href'
EXAMPLE_TYPE_XPATH | = | 0 | XPath example, like html/body/tr/td[1] etc. | |
EXAMPLE_TYPE_STRING | = | 1 | String from the document, for example ‘Canon EOS 300 D’. | |
EXAMPLE_TYPE_IMAGE | = | 2 |
Image example, like ‘![]() |
|
EXAMPLE_TYPE_CHILDREN | = | 3 | No example - the actual XPath is determined from the children XPaths (their LCA) | |
EXAMPLE_TYPE_REGEXP | = | 4 | Regexp example, like /\d+@*\d+[a-z]/ | |
EXAMPLE_TYPE_COMPOUND | = | 5 | Compound example, like :contains => ‘goodies‘ |
method_missing | -> | throw_method_missing |
TODO still used? |
constraints | [RW] | |
example | [RW] | |
example_type | [RW] | |
final_result | [RW] | |
parent_pattern | [RW] | |
regexp | [RW] | |
temp_sink | [RW] | |
xpath | [RW] |