Module definining types used by the library.
- type Tag = String
- class DataFormat a
- data XML = XML
- data JSON = JSON
- class Response r => CoerceResponse x r | x -> r, r -> x where
- class (Site s, DataFormat r) => PostParser s r where
- class (Site s, DataFormat r) => Counted s r where
- class (Site s, DataFormat r) => Postable s r where
- class Site s
- data Rating
- data Limit
- data XMLResponse = XMLResponse String
- data JSONResponse = JSONResponse String
- class Response r where
- class Post a where
- height :: a -> Integer
- score :: a -> Integer
- file_url :: a -> String
- parent_id :: a -> Maybe Integer
- sample_url :: a -> String
- sample_width :: a -> Integer
- sample_height :: a -> Integer
- preview_url :: a -> String
- rating :: a -> Rating
- tags :: a -> [String]
- id :: a -> Integer
- width :: a -> Integer
- change :: a -> String
- md5 :: a -> String
- creator_id :: a -> Integer
- has_children :: a -> Maybe Bool
- created_at :: a -> String
- status :: a -> String
- source :: a -> String
- has_notes :: a -> Maybe Bool
- has_comments :: a -> Maybe Bool
- preview_width :: a -> Integer
- preview_height :: a -> Integer
- betweenPosts :: Post b => a -> PostConstructor b -> b
- type PostConstructor b = Integer -> Integer -> String -> Maybe Integer -> String -> Integer -> Integer -> String -> Rating -> [String] -> Integer -> Integer -> String -> String -> Integer -> Maybe Bool -> String -> String -> String -> Maybe Bool -> Maybe Bool -> Integer -> Integer -> b
Tags used for searching in sites. No special escaping is done. Note that many sites would treat a tag like "striped panties" as two separate tags and you wouldn't get the results you were after.
Data format used by various
Sites. See instances for currently used
Used as one of the data formats.
Used as one of the data formats.
Thanks to this class, we're able to provide instances converting
Response. This is useful if we need a
while we only have a type that's an instance of
Response. Note that the
functional dependency currently requires that there is only one way to coerce
between two types.
toResponse XML "<SomeXML></SomeXML>"XMLReponse "<SomeXML></SomeXML>"
Class specifying a parser that can fetch posts. A post usually consists of
links to the image, samples, and some meta-data. See
GenericPost for the kind of thing we usually
get out. The reason for this class is that sometimes we might get different
information based on the
DataFormat we use so we use type families to
denote this rather than forcing the library user to make do with our best
guess on what goes into the post. It also allows us to use different post
types for sites that provide different information.
Given a parser working with
DataFormat specified by an instance of
this class, we require through
CoerceResponse that it is able to parse
responses in the format so what we actually pass into this function is
Site this parser works with (so that we can pick the appropriate data
type for the posts) and a
Response matching the
DataFormat (through a
class instance). For
PostParser instance, example use
might go like
do fc <-
readFile"gelbooruResponse.xml" -- the type of images is actually inferred for us let images ∷ [
GelbooruPost] images = parseResponse
Gelboorufc return images
The cool thing is that we can't feed anything but
XMLResponse to an
DataFormat and a list of
Tags, an instance of this
class should be able to return a
String at which we can find data in
DataFormat format that honours our tags. This is effectively a URL
builder for POST requests.
Provides information about whether there's a hard limit on the amount of
posts we can fetch from the site at once. The reason for this function here
rather than in
Site is that we might be parsing data without an API we
can post to at all and we're getting our data through other means.
Describes a site for a parser. The reason why this isn't a simple data type is to allow us to write additional parsers in the future without modifying this library if we wish to do so.
Denotes whethere there's a hard limit on the number of posts
we can fetch at a time from a site. NoLimit implies that we can fetch
everything at once and not that we don't know. See
Counted for a way to
potentially retrieve number of posts present on the site.
Specifies what is considered a response. You'll almost certainly also
CoerceResponse instances if you're adding some
here. This class assumes that all responses carry the response in a string we
can extract. Note that this is not for use as network response if you're
scraping, only for putting data into after you have done all the error
checking and whatnot.
Class representing a best-case scenario post. We use this to convert between different posts for each site while providing uniform access. The methods are just the attributes of posts seen on Gelbooru-like sites.
type PostConstructor b = Integer -> Integer -> String -> Maybe Integer -> String -> Integer -> Integer -> String -> Rating -> [String] -> Integer -> Integer -> String -> String -> Integer -> Maybe Bool -> String -> String -> String -> Maybe Bool -> Maybe Bool -> Integer -> Integer -> bSource
A cludge for use with