containerd/filters/parser.go
Stephen J Day d69ef98bfd
filters: add package for filter syntax
With this PR, we add the syntax to use for filtration of items over the
containerd API. This package defines a syntax and parser that can be
used across types and use cases in a uniform manner.

The syntax is fairly familiar, if you've used container ecosystem
projects.  At the core, we base it on the concept of protobuf field
paths, augmenting with the ability to quote portions of the field path
to match arbitrary labels. These "selectors" come in the following
syntax:

```
<fieldpath>[<operator><value>]
```

A basic example is as follows:

```
name=foo
```

This would match all objects that have a field `name` with the value
`foo`. If we only want to test if the field is present, we can omit the
operator. This is most useful for matching labels in containerd. The
following will match objects that has the field labels and have the
label "foo" defined:

```
labels.foo
```

We also allow for quoting of parts of the field path to allow matching
of arbitrary items:

```
labels."very complex label"==something
```

We also define `!=` and `~=` as operators. The `!=` operator will match
all objects that don't match the value for a field and `~=` will compile
the target value as a regular expression and match the field value
against that.

Selectors can be combined using a comma, such that the resulting
selector will require all selectors are matched for the object to match.
The following example will match objects that are named `foo` and have
the label `bar`:

```
name==foo,labels.bar
```

This filter syntax will be used across all APIs that allow listing of
objects and for filtering which event a cleint see. By using a common
syntax, we hope to keep API access uniform.

For the most part, this takes inspiration from docker, swarm and k8s,
but has the limitation that it only allows selection of an inner
product. We may expand to operators that implement `or`, `in` or
`notin`, but it is not clear that this is useful at this level of the
stack.

Signed-off-by: Stephen J Day <stephen.day@docker.com>
2017-06-20 13:03:35 -07:00

240 lines
4.5 KiB
Go

package filters
import (
"fmt"
"io"
"strconv"
"github.com/pkg/errors"
)
/*
Parse the strings into a filter that may be used with an adaptor.
The filter is made up of zero or more selectors.
The format is a comma separated list of expressions, in the form of
`<fieldpath><op><value>`, known as selectors. All selectors must match the
target object for the filter to be true.
We define the operators "==" for equality, "!=" for not equal and "~=" for a
regular expression. If the operator and value are not present, the matcher will
test for the presence of a value, as defined by the target object.
The formal grammar is as follows:
selectors := selector ("," selector)*
selector := fieldpath (operator value)
fieldpath := field ('.' field)*
field := quoted | [A-Za-z] [A-Za-z0-9_]+
operator := "==" | "!=" | "~="
value := quoted | [^\s,]+
quoted := <go string syntax>
*/
func Parse(s string) (Filter, error) {
// special case empty to match all
if s == "" {
return Always, nil
}
p := parser{input: s}
return p.parse()
}
type parser struct {
input string
scanner scanner
}
func (p *parser) parse() (Filter, error) {
p.scanner.init(p.input)
ss, err := p.selectors()
if err != nil {
return nil, errors.Wrap(err, "filters")
}
return ss, nil
}
func (p *parser) selectors() (Filter, error) {
s, err := p.selector()
if err != nil {
return nil, err
}
ss := All{s}
loop:
for {
tok := p.scanner.peek()
switch tok {
case ',':
pos, tok, _ := p.scanner.scan()
if tok != tokenSelectorSeparator {
return nil, p.mkerr(pos, "expected a separator")
}
s, err := p.selector()
if err != nil {
return nil, err
}
ss = append(ss, s)
case tokenEOF:
break loop
default:
panic("unconsumed input")
}
}
return ss, nil
}
func (p *parser) selector() (selector, error) {
fieldpath, err := p.fieldpath()
if err != nil {
return selector{}, err
}
switch p.scanner.peek() {
case tokenSelectorSeparator, tokenEOF:
return selector{
fieldpath: fieldpath,
operator: operatorPresent,
}, nil
}
op, err := p.operator()
if err != nil {
return selector{}, err
}
value, err := p.value()
if err != nil {
if err == io.EOF {
return selector{}, io.ErrUnexpectedEOF
}
return selector{}, err
}
return selector{
fieldpath: fieldpath,
value: value,
operator: op,
}, nil
}
func (p *parser) fieldpath() ([]string, error) {
f, err := p.field()
if err != nil {
return nil, err
}
fs := []string{f}
loop:
for {
tok := p.scanner.peek() // lookahead to consume field separtor
switch tok {
case '.':
pos, tok, _ := p.scanner.scan() // consume separator
if tok != tokenFieldSeparator {
return nil, p.mkerr(pos, "expected a field separator (`.`)")
}
f, err := p.field()
if err != nil {
return nil, err
}
fs = append(fs, f)
default:
// let the layer above handle the other bad cases.
break loop
}
}
return fs, nil
}
func (p *parser) field() (string, error) {
pos, tok, s := p.scanner.scan()
switch tok {
case tokenField:
return s, nil
case tokenQuoted:
return p.unquote(pos, s)
}
return "", p.mkerr(pos, "expected field or quoted")
}
func (p *parser) operator() (operator, error) {
pos, tok, s := p.scanner.scan()
switch tok {
case tokenOperator:
switch s {
case "==":
return operatorEqual, nil
case "!=":
return operatorNotEqual, nil
case "~=":
return operatorMatches, nil
default:
return 0, p.mkerr(pos, "unsupported operator %q", s)
}
}
return 0, p.mkerr(pos, `expected an operator ("=="|"!="|"~=")`)
}
func (p *parser) value() (string, error) {
pos, tok, s := p.scanner.scan()
switch tok {
case tokenValue, tokenField:
return s, nil
case tokenQuoted:
return p.unquote(pos, s)
}
return "", p.mkerr(pos, "expected value or quoted")
}
func (p *parser) unquote(pos int, s string) (string, error) {
uq, err := strconv.Unquote(s)
if err != nil {
return "", p.mkerr(pos, "unquoting failed: %v", err)
}
return uq, nil
}
type parseError struct {
input string
pos int
msg string
}
func (pe parseError) Error() string {
if pe.pos < len(pe.input) {
before := pe.input[:pe.pos]
location := pe.input[pe.pos : pe.pos+1] // need to handle end
after := pe.input[pe.pos+1:]
return fmt.Sprintf("[%s >|%s|< %s]: %v", before, location, after, pe.msg)
}
return fmt.Sprintf("[%s]: %v", pe.input, pe.msg)
}
func (p *parser) mkerr(pos int, format string, args ...interface{}) error {
return errors.Wrap(parseError{
input: p.input,
pos: pos,
msg: fmt.Sprintf(format, args...),
}, "parse error")
}