IMPROVED:
* Better documentation for remap
This commit is contained in:
brent saner 2025-08-12 00:06:51 -04:00
parent 368ae0cb8e
commit bae0abe960
Signed by: bts
GPG Key ID: 8C004C2F93481F6B
3 changed files with 166 additions and 15 deletions

View File

@ -1,4 +1,4 @@
/*
Package remap provides convenience functions around regular expressions.
Package remap provides convenience functions around regular expressions, primarily offering maps for named capture groups.
*/
package remap

View File

@ -1,11 +1,67 @@
package remap
/*
Map returns a map[string]<match bytes> for regexes with named capture groups matched in bytes b.
Note that this supports non-unique group names; regexp.Regexp allows for patterns with multiple groups
using the same group name. Each match for each group is in a slice keyed under that group name, with
that slice ordered by the indexing done by the regex match itself.
matches and/or its values may be nil or empty under the following condition tree:
Map returns a map[string][]<match bytes> for regexes with named capture groups matched in bytes b.
Note that this supports non-unique group names; [regexp.Regexp] allows for patterns with multiple groups
using the same group name (though your IDE might complain; I know GoLand does).
Each match for each group is in a slice keyed under that group name, with that slice
ordered by the indexing done by the regex match itself.
In summary, the parameters are as follows:
# inclNoMatch
If true, then attempt to return a non-nil matches (as long as b isn't nil).
Group keys will be populated and explicitly defined as nil.
For example, if a pattern
^(?P<g1>foo)(?P<g1>bar)(?P<g2>baz)$
is provided but b does not match then matches will be:
map[string][][]byte{
"g1": nil,
"g2": nil,
}
# inclNoMatchStrict
If true (and inclNoMatch is true), instead of a single nil the group's values will be
a slice of nil values explicitly matching the number of times the group name is specified
in the pattern.
For example, if a pattern:
^(?P<g1>foo)(?P<g1>bar)(?P<g2>baz)$
is provided but b does not match then matches will be:
map[string][][]byte{
"g1": [][]byte{
nil,
nil,
},
"g2": [][]byte{
nil,
},
}
# mustMatch
If true, matches will be nil if the entirety of b does not match the pattern (and thus
no capture groups matched) (overrides inclNoMatch) -- explicitly:
matches == nil
Otherwise if false (and assuming inclNoMatch is false), matches will be:
map[string][][]byte{}{}
# Condition Tree
In detail, matches and/or its values may be nil or empty under the following condition tree:
IF b is nil:
THEN matches will always be nil
@ -148,9 +204,100 @@ func (r *ReMap) Map(b []byte, inclNoMatch, inclNoMatchStrict, mustMatch bool) (m
}
/*
MapString is exactly like ReMap.Map(), but operates on (and returns) strings instead. (matches will always be nil if s == ``.)
MapString is exactly like ReMap.Map(), but operates on (and returns) strings instead.
(matches will always be nil if s == .)
A small deviation, though; empty strings instead of nils (because duh) will occupy placeholders (if inclNoMatchStrict is specified).
A small deviation, though; empty strings instead of nils (because duh) will occupy slice placeholders (if `inclNoMatchStrict` is specified).
This unfortunately *does not provide any indication* if an empty string positively matched the pattern (a "hit") or if it was simply
not matched at all (a "miss"). If you need definitive determination between the two conditions, it is instead recommended to either
*not* use inclNoMatchStrict or to use ReMap.Map() instead and convert any non-nil values to strings after.
Particularly:
# inclNoMatch
If true, then attempt to return a non-nil matches (as long as s isn't empty).
Group keys will be populated and explicitly defined as nil.
For example, if a pattern
^(?P<g1>foo)(?P<g1>bar)(?P<g2>baz)$
is provided but s does not match then matches will be:
map[string][]string{
"g1": nil,
"g2": nil,
}
# inclNoMatchStrict
If true (and inclNoMatch is true), instead of a single nil the group's values will be
a slice of eempty string values explicitly matching the number of times the group name is specified
in the pattern.
For example, if a pattern:
^(?P<g1>foo)(?P<g1>bar)(?P<g2>baz)$
is provided but s does not match then matches will be:
map[string][]string{
"g1": []string{
"",
"",
},
"g2": []string{
"",
},
}
# mustMatch
If true, matches will be nil if the entirety of s does not match the pattern (and thus
no capture groups matched) (overrides inclNoMatch) -- explicitly:
matches == nil
Otherwise if false (and assuming inclNoMatch is false), matches will be:
map[string][]string{}{}
# Condition Tree
In detail, matches and/or its values may be nil or empty under the following condition tree:
IF s is empty:
THEN matches will always be nil
ELSE:
IF all of s does not match pattern
IF mustMuch is true
THEN matches == nil
ELSE
THEN matches == map[string][]string{} (non-nil but empty)
ELSE IF pattern has no named capture groups
IF inclNoMatch is true
THEN matches == map[string][]string{} (non-nil but empty)
ELSE
THEN matches == nil
ELSE
IF there are no named group matches
IF inclNoMatch is true
THEN matches is non-nil; matches[<group name>, ...] is/are defined but nil (_, ok = matches[<group name>]; ok == true)
ELSE
THEN matches == nil
ELSE
IF <group name> does not have a match
IF inclNoMatch is true
IF inclNoMatchStrict is true
THEN matches[<group name>] is defined and non-nil, but populated with placeholder nils
(matches[<group name>] == []string{""[, ""...]})
ELSE
THEN matches[<group name>] is guaranteed defined but may be nil (_, ok = matches[<group name>]; ok == true)
ELSE
THEN matches[<group name>] is not defined (_, ok = matches[<group name>]; ok == false)
ELSE
matches[<group name>] == []{<match>[, <match>...]}
*/
func (r *ReMap) MapString(s string, inclNoMatch, inclNoMatchStrict, mustMatch bool) (matches map[string][]string) {

View File

@ -1,7 +1,7 @@
package remap
import (
`regexp`
"regexp"
)
type (
@ -10,14 +10,18 @@ type (
*regexp.Regexp
}
// TODO?
/*
ExplicitStringMatch is used with ReMap.MapStringExplicit to indicate if a
capture group result is a hit (a group matched, but e.g. the match value is empty string)
or not (a group did not match)
or not (a group did not match).
*/
/*
ExplicitStringMatch struct {
Group string
IsMatch bool
Value string
}
*/
ExplicitStringMatch struct {
Group string
IsMatch bool
Value string
}
)