DPath

DPath is a completed project for my EECS 433 Databases course. It is an implementation of an XQuery-like language for file system queries. For example, the DPath query below would match every PNG file in the current directory tree:

.//.[ends-with(name(), ".png")]

Usage

Setup

This will get you where you need to go, although there may be a "better way".

$ export GOPATH=~/go
$ export PATH=$GOPATH/bin:$PATH
$ go get -t github.com/blynn/nex github.com/brenns10/dpath
# this will give an error, ignore it
$ cd $GOPATH/src/github.com/brenns10/dpath
$ go generate
$ go build
$ go test

Usage

You can go install once you've done go generate, which will put the dpath command in your Go binary directory, which is hopefully in your $PATH. From there, try some queries:

$ dpath './/.'
# recursively lists this subdirectory and everything under it

$ dpath './/.[empty(ancestor-or-self::#".git")]'
# recursively lists this subdirectory, filtering out things within a .git
# directory.

$ dpath '/home/stephen//.[ends-with(name(), ".png")]'
# lists PNG files in home directory

$ dpath '/home/stephen//.[contains(name(), "smb196")]'
# finds files I named with my school ID, likely for homework submissions

$ dpath '/home/stephen//.[matches(name(), ".*\w{3}\d{1,3}.*")]'
# finds files that have any school ID in them
# this turns up lots of junk because it's a common pattern

$ dpath './/.[starts-with(name(), parent::*/name())]'
# or just:
$ dpath './/.[starts-with(name(), ../name())]'
# finds files that start with their containing directory's name
file:/home/stephen/go/src/github.com/brenns10/dpath/dpath.nex
file:/home/stephen/go/src/github.com/brenns10/dpath/dpath.y
file:/home/stephen/go/src/github.com/brenns10/dpath/dpath
file:/home/stephen/go/src/github.com/brenns10/dpath/dpath.nn.go
file:/home/stephen/go/src/github.com/brenns10/dpath/.git/objects/e9/e9f542b2423e029b7adc72f71265e2eabb63a6

$ dpath './/file()[@size > 4096]'
# find files greater than 4K bytes

A full description of the usage is described in SYNTAX.md, although that document assumes a basic knowledge of XPath.

If you're interested in how this implementation works, I maintain the file GUIDE.md, which should give some high-level explanation of how the language comes together. The low-level details can be read about in comments and code.

Status

Completed

• Lexer
• Parser
• Arithmetic expressions on int, double involving operators + - * div idiv mod
• Range expressions for numeric types, e.g. (1 to 5) which evaluates to (1, 2, 3, 4, 5).
• Value (= != <= < >= >) and General (eq ne le lt ge gt) comparisons. The difference being that Value requires singletons, whereas General will look for any pair of atomics in the input sequences that satisfy the comparisons.
• Boolean logic expressions (and or)
• Predicate syntax on sequences, e.g. (1 to 5)[. mod 2 eq 0] which evaluates to (2, 4).
• Path expressions on the following axes: child, parent, descendant, descendant-or-self, ancestor, ancestor-or-self, attribute.
• The shorthand notations *, .., //, #"spaces etc here"
• Functions: boolean(), concat(), round(), substring(), string(), string-length(), ends-with(), starts-with(), contains(), matches(), empty(), exists(), name(), path(), count().
• Selectors: file(), dir()