Skip to content

Latest commit

 

History

History
241 lines (110 loc) · 3.3 KB

GRAMMAR.md

File metadata and controls

241 lines (110 loc) · 3.3 KB
Raw

General Syntax

sq consists of two principle symbols Verbs and Nouns. Sets of Verbs and Nouns can be chained together to form Clauses. Data flows from left to right with the output of one Verb becoming the input for the next.

Character Set

abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789./=:|

regex: [a-zA-Z0-9./=:|]

The characters :=| are reserved for language delimiters

Verbs

Verbs are analogous to functions with Nouns being their arguments.


	head			# An example Verb

	ls:mybucket 		# Verb with Noun

	get:mybucket:myobject	# Verb with two Nouns

More than one Verb can be chained together using pipe | into a Clause. The output of each Verb becomes the Nouns for the next Verb in the chain, allowing complex behavior to be built.


	# List a bucket and HEAD each object
	ls:mybucket | head

	# Copy only mp4 files from one bucket to another
	ls:mymedia | gr:.*\.mp4$ | cp:mymp4s

	# Count the number of object successfully replicated
	lr:mycrrbucket | rstat:SUCCESS | count

Nouns

Nouns come in two flavors, positional and keyword, however all Nouns have an associated keyword. Internally all positional Nouns are mapped to keywords for identification, as such positional Nouns can be thought of as a less verbose form of keyword Nouns that are provided to the user for convenience.


	# this
	ls:mybucket:myobject

	# is the same as this
	ls:bucket=mybucket:key=myobject

Not all Nouns can be specified as positional for every Verb. Consult the Language Reference for more details

Language Reference

Available Verbs

Verbs are shown in the following format: verb:<required argument>:[optional arguement]

Listing

List objects from storage


ls:[bucket]

List buckets or objects if bucket provided


lv:<bucket>:[key]

List object versions


lr:[bucket]

List all CRRs or CRRs scoped by bucket

Operations

Perform operations on objects


head:<bucket>:<key>:[version-id]

HEAD an object


get:<bucket>:<key>:[versionid]

GET an object


put:<bucket>:[key]:[file=/path/to/file]:[content=[zero,rand]]:[size=100B]

PUT object in bucket at key


mpu:<bucket>:[key]:<parts=10>:[content=[zero,rand]]

Similar to put this uploads an object as an mpu using parts object parts

Filter

Filter lists of objects or buckets

Filters can be negated by adding a ! as its first argument

eg gr:!:rejecthispattern


gr:<regex>

gr:!:<regex>

Match names based on a regex


rstat:<status>

Filter based upon replication status


all

Match everything

Stats

Generate counts based upon filters, these will be printed at the end of the run

Optionally these can take a label which will be printed along with the count


stat:<filter>:[label]

Counts emitted items based upon filter

stat:gr:^matchthisline$


count:[label]

Special filter that behaves like stat:all. Simply count all items

Generators

Create resources

All generators use uuid4 to create unique names if no prefix is given


bkt:<count>:[prefix]

Generate count buckets beginning with prefix if given


obj:<count>:[prefix]

Generate count objects beginning with prefix if given