mdiff
v1.4.2
Published
An LCS Implementaion for Arrays and Strings
Downloads
5,401
Maintainers
Readme
mdiff
A Minimalistic Diff Implementation
Based on the algorithm proposed in "An O(ND) Difference Algorithm and its Variations" (Myers, 1986). Works with Arrays and Strings.
Usage
$ npm install mdiff
import mdiff from 'mdiff';
let a = 'ABCABBA';
let b = 'CBABAC';
let diff = mdiff(a, b);
console.log("lcs='%s'", diff.getLcs());
console.log("Common:");
let d = diff.scanCommon((aS, aE, bS, bE) => {
console.log(" '%s' == '%s'", a.slice(aS, aE), b.slice(bS, bE));
});
console.log("edit-distance=%s", d);
console.log("Diff:");
d = diff.scanDiff((aS, aE, bS, bE) => {
console.log(" '%s' -> '%s'", a.slice(aS, aE), b.slice(bS, bE));
});
console.log("edit-distance=%s", d);
API
var diff = mdiff(a, b, options)
Creates a diff-object.
a
,b
(strings or arrays): the items to be compared.options
equal
(function(aValue, bValue)
): a comparator that determines if two entries are supposed to be equal (default is just===
).indexEqual
(function(aIdx, bIdx)
): a comparator that determines if entries for two indices are supposed to be equal (default is to comparea[aIdx]
withb[bIdx]
byequal
).
diff.scanCommon(cb, dMax)
Calls cb
for each common slice and returns the edit-distance d
.
cb
(function(aS, aE, bS, bE)
): reports the corresponding slices with these guarantees:- non-emptiness:
aE - aS == bE - bS > 0
- monotony:
aS
is not less thanaE
from the previous call.bS
is not less thanbE
from the previous call. - equality:
a.slice(aS, aE)
is equalb.slice(bS, bE)
with respect toequal
. - minimality: The sum of all
aS - aE
is equal to(a.length + b.length - d) / 2
and there is no lesserd
so that 'equality' holds.
- non-emptiness:
dMax
(optional): the maximum edit-distance to be handled. If the items' edit-distance happens to exceed this value, the function never callscb
and returnsnull
.
diff.scanDiff(cb, dMax)
Calls cb
for each difference and returns the edit-distance d
. This is just the complement of scanCommon
.
cb
(function(aS, aE, bS, bE)
): reports differences with these guarantees:- non-emptiness:
aE - aS > 0
orbE - bS > 0
- monotony:
aS
is not less thanaE
from the previous call.bS
is not less thanbE
from the previous call. - equality: If in
a
all slicesa.slice(aS, aE)
are replaced by their correspondingb.slice(bS, bE)
, the result is equal tob
with respect toequal
. - minimality: The sum of all
aS - aE
plus the sum of allbS - bE
is equal tod
and there is no lesserd
so that 'equality' holds.
- non-emptiness:
dMax
(optional): the maximum edit-distance to be handled. If the items' edit-distance happens to exceed this value, the function never callscb
and returnsnull
.
diff.getLcs(dMax)
Returns the longest common subsequence of a
and b
(as a
's type) or null
, if the items' edit-distance exceeds dMax
.