globber.glob — spootnik/globber 0.4.2

Globbing searches in clojure.

all-indices clj

(all-indices s pat)

Find all indices of a substring in a given string

ast-type clj

(ast-type ast)

Determine if the ast can be unrolled to a string, can be exploded to a set of strings or needs to run through our algorithm.

ast-typenames clj

Yield a keyword name for the AST type

char->token clj

(char->token [prev c])

Our simple lexer, understands escaped characters.

close-branch clj

(close-branch branch)

We succesfuly parsed a branch, now yield the best possible AST for it.

close-charclass clj

(close-charclass cc nchar)

We're finished running through a class. Yield a set with some additional metadata to indicate whether the class is negated.

This also handles edge-cases (zero or single element classes).

explode-ast clj

(explode-ast ast)

Explode a non-compound AST to all its possible combinations.

explode-compound-ast clj

(explode-compound-ast partitions)

Transform a partitioned compound AST by exploding eligible sub-ASTs

expr->tokens clj

(expr->tokens expr)

Looks at characters two at a time to yield a list of tokens.

filter-compound-ast clj

(filter-compound-ast partitions candidates)

Given a compound AST, yield the set of matching candidates.

filter-compound-partition clj

(filter-compound-partition [eaters patterns] [pos candidate])

Match a list of candidates against a single partition of a compound AST. When a match occurs, yield a list of the new candidates this may have generated and if they are eligible as a terminal match.

find-pattern clj

(find-pattern pos candidate s eaters pattern)

Match a string against a list of eaters and a pattern

glob clj

(glob expression candidates)

Perform globbing, matching the supplied expression against the list of supplied candidates.

• candidates is a collection of strings and must be seqable.
• expression adheres mostly to the bash notion of globbing

Globbing syntax:

• Any stray character is matched exactly
• Wildcards ('*') mean any number (including zero) of arbitrary chars
• Anyones ('?') mean a single arbitrary char
• Character classes are enclosed in square brackets and may contain arbitrary list of characters to match. If a character class begins with a ('!') or ('^') character, the class will be negated, i.e: will only match characters absent from the class. Empty charclasses, i.e: ('[]'), ('[!]'), and ('[^]') match their representation, not their content.
• Trees match any of their branches. Trees are delimited by curly brackets and content are separated by commas. Empty trees, i.e: ('{}'), ('{,}'), ('{,,}') match their representation, not their content.

Examples:

(glob "foobar" ["foobar"])  ;; => ("foobar")
(glob "fo[a-z]" ["foobar"]) ;; => ("foobar")

make-class clj

(make-class beg end)

Build a list of all characters between beg and end

merge-ast clj

(merge-ast)

(merge-ast left right)

Merge two non-compound ASTs by yielding all possible combinations.

negated-cc? clj

(negated-cc? ast)

Is this representing a negated character class?

partition-compound-ast clj

(partition-compound-ast tokens)

Partition a compound AST into a list of (eaters, ast) tuples where eaters is a list of eager character eaters and a potentially explodable AST

scan-branch clj

(scan-branch tokens)

Eat tokens until the end of a branch.

scan-branches clj

(scan-branches tokens)

A tree has started, scan all its branches and yield the best possible AST for it.

scan-charclass clj

(scan-charclass tokens)

Eat tokens until the character class is closed.

stringify-ast clj

(stringify-ast ast)

Reduce a string-only AST to its equivalent string

tokens->ast clj

(tokens->ast tokens)

Transform a list of tokens to an AST, if possible.