css-syntax alternatives and similar packages
Based on the "CSV" category.
Alternatively, view css-syntax alternatives based on common mentions on social networks and blogs.
cassava9.7 0.0 css-syntax VS cassavaA CSV parsing and encoding library optimized for ease of use and high performance
css-text8.2 0.0 css-syntax VS css-textCSS parser and renderer.
pipes-csv7.2 0.0 css-syntax VS pipes-csvStreaming csv parser using cassava and pipes
csv-enumerator5.6 0.0 css-syntax VS csv-enumeratorAn enumerator-based, flexible, constant-memory and fast CSV parser library for Haskell.
csv-table4.4 0.0 css-syntax VS csv-tableScripts for manipulating tables stored as CSV files
hxt-css3.9 0.0 css-syntax VS hxt-cssA CSS selector engine for the Haskell XML Toolbox (XHT)
cassava-streams2.8 0.0 css-syntax VS cassava-streamsio-streams interface for the cassava CSV library.
Access the most powerful time series database as a service
Do you think we are missing an alternative of css-syntax or a related project?
This package provides functions to parse a CSS file into a stream of tokens and from that into rules or declarations. The exact algorithm is defined in the css-syntax module (it is the same algorithm that Blink uses since mid 2015 when it replaced their old Bison based parser).
Note: Only the tokenizer is currently implemented. Parsing the token stream into rules or declarations isn't available as of yet.
I needed a library which would allow me to process a CSS file, modify all image URLs and write the result to disk. Existing CSS libraries for Haskell either focus only on generating CSS (through a DLS) or parse the source into a format that is too high-level and makes this kind of processing difficult.
A token stream is just the right level of abstraction for that task. The spec
<function-token>, which is what I needed to
extract the image URLs from the source.
More advanced processing likely requires a higher level abstraction than the token stream provides. E.g. to expand vendor prefixes you'll need to parse the token stream into a list of rules and declarations, so you can pick the declarations you want to process.
I (the second author) needed to preprocess HTML in realtime to make it responsive. Besides other things it requires parsing
style=... attribute that can have any amount of junk so I optimized a parser/serializer a lot while still passing all the tests.
The tokenizer uses fast low-level parser (20-50MB/s on average CSS files) to convert the input to a list of tokens. This process removes all comments and collapses consecutive whitespace into a single space character (U+0020). There may be other occasions where the tokenizer looses information from the input stream.
Serializer converts list of tokens back to string. Serialization round-trips: tokenizing produces same tokens list as tokenizing, serializing and tokenizing again. Tokenize-serialize pair works at about 15MB/s or more.
In the following example I replace all URLs in the source CSS file with links to predefined image.
import qualified Data.Text.IO as T import Data.CSS.Syntax.Tokens (tokenize, serialize) main :: IO () main = do source <- T.readFile "path-to-your.css" let tokens = tokenize source putStrLn $ "The CSS file has " ++ show (length tokens) ++ " tokens" let newTokens = map f tokens T.writeFile "path-to-lolcatted.css" $ serialize newTokens f :: Token -> Token f (Url _) = Url "http://lol.cat/img.png" f x = x