/// <summary>
/// Initializes a new instance of the <see cref="StringMap" /> class.
/// </summary>
/// <param name="input">The input.</param>
/// <param name="options">The options.</param>
/// <exception cref="ArgumentNullException">
/// <paramref name="input" /> is <see langword="null" />.</exception>
/// <exception cref="ArgumentOutOfRangeException">
/// <paramref name="options" /> are invalid.</exception>
internal StringMap(
[NotNull] string input,
TextOptions options = TextOptions.Default)
{
if (input == null)
{
throw new ArgumentNullException(nameof(input));
}
if (options == (TextOptions.Trim | TextOptions.CollapseWhiteSpace))
{
throw new ArgumentOutOfRangeException(nameof(options));
}
Original = input;
Options = options;
int length = input.Length;
if (length < 1)
{
Mapped = string.Empty;
Map = _emptyMap;
return;
}
if (options == TextOptions.None)
{
Mapped = input;
Map = new[] { 0, length, length };
return;
}
StringBuilder builder = new StringBuilder(Original.Length);
List <int> map = new List <int>(4);
int i = 0;
int offset = 0;
int count = 0;
char c;
bool skip;
// Logic for adding a map
Action <int, int> addMap = (s, e) =>
{
int len = e - s;
count += len;
map.Add(s);
map.Add(len);
map.Add(count);
};
// ReSharper disable EventExceptionNotDocumented
switch (options)
{
/*
* Simplest case is to skip all white space - this is less useful in that it effectively ignores word
* boundaries. Note this implicitly 'trims', 'normalizes line endings' and 'collapses white space' as
* ALL white space is ignored.
*/
case TextOptions.IgnoreWhiteSpace:
while (i < length)
{
c = Original[i++];
if (char.IsWhiteSpace(c))
{
if (i - offset > 1)
{
addMap(offset, i - 1);
}
offset = i;
continue;
}
builder.Append(c);
}
if (i - offset > 0)
{
addMap(offset, i);
}
break;
/*
* Collapse white space ignores all but the first white space character, effectively preserving the
* word boundaries. Note this implicitly 'normalizes line endings' as only the first character of
* white space is preserved, so only on line ending character will be kept.
*
* It makes not sense to allow this in conjunction with trim as line breaks are effectively treated the
* same as any white space.
*/
case TextOptions.CollapseWhiteSpace:
skip = false;
while (i < length)
{
c = Original[i++];
if (char.IsWhiteSpace(c))
{
if (!skip)
{
skip = true;
builder.Append(c);
continue;
}
if (i - offset > 1)
{
addMap(offset, i - 1);
}
offset = i;
continue;
}
builder.Append(c);
skip = false;
}
if (i - offset > 0)
{
addMap(offset, i);
}
break;
/*
* Trim and NormalizeLineEndings both require line detection and so we treat together, they can also
* be used in combination.
*/
case TextOptions.Trim:
case TextOptions.NormalizeLineEndings:
case TextOptions.Trim | TextOptions.NormalizeLineEndings:
bool normalizeLineEndings = options.HasFlag(TextOptions.NormalizeLineEndings);
bool trim = options.HasFlag(TextOptions.Trim);
// Skip is used to indicate we've had the first character in a line.
skip = false;
// The index after the last non-white space character
StringBuilder trailingWhiteSpace = new StringBuilder(16);
while (i < length)
{
c = Original[i++];
if (char.IsWhiteSpace(c))
{
bool isNewLine = c == '\n';
// Detect line endings
if (isNewLine || c == '\r')
{
if (i < 2 || !skip || !normalizeLineEndings || !isNewLine)
{
if (!skip && (i - offset > 1))
{
int end = i;
// Trim trailing white space if any
if (trim && trailingWhiteSpace.Length > 0)
{
end -= trailingWhiteSpace.Length;
trailingWhiteSpace.Clear();
}
if (end - offset > 1)
{
addMap(offset, end - 1);
}
offset = i - 1;
}
// Preserve first line ending when normalizing or all line endings.
skip = true;
builder.Append(c);
continue;
}
// Skip second line ending character if == '\n'
if (i - offset > 1)
{
addMap(offset, i - 1);
}
offset = i;
continue;
}
if (trim)
{
if (i < 2 || skip)
{
skip = true;
// White space at start of line.
if (i - offset > 1)
{
addMap(offset, i - 1);
}
offset = i;
continue;
}
// Hold whitespace in temporary builder.
trailingWhiteSpace.Append(c);
continue;
}
}
if (trailingWhiteSpace.Length > 0)
{
builder.Append(trailingWhiteSpace);
trailingWhiteSpace.Clear();
}
builder.Append(c);
skip = false;
}
i -= trailingWhiteSpace.Length;
trailingWhiteSpace.Clear();
if (i - offset > 0)
{
addMap(offset, i);
}
break;
default:
throw new ArgumentOutOfRangeException(nameof(options), options, null);
}
// ReSharper restore EventExceptionNotDocumented
if (map.Count < 1)
{
Mapped = string.Empty;
Map = _emptyMap;
return;
}
// Store mapped string and the map.
Mapped = builder.ToString();
Map = map.ToArray();
#if false
// Create map
List <int> map = new List <int>();
bool ignoreWhiteSpace = options.HasFlag(TextOptions.IgnoreWhiteSpace);
bool collapseWhiteSpace = ignoreWhiteSpace || options.HasFlag(TextOptions.CollapseWhiteSpace);
bool normalizeLineEndings = ignoreWhiteSpace || options.HasFlag(TextOptions.NormalizeLineEndings);
bool trim = ignoreWhiteSpace || options.HasFlag(TextOptions.Trim);
int o = 0;
int l = 0;
int lnws = -1;
int i = 0;
int end;
// Logic for adding a map
Action <int, int> addMap = (int s, int e) =>
{
int mc = map.Count;
int len = e - s;
Count += len;
map.Add(s);
map.Add(len);
map.Add(Count);
};
// TODO Not working for most tests!
// ReSharper disable EventExceptionNotDocumented
while (i < length)
{
char c = Original[i++];
// Check for white space
if (char.IsWhiteSpace(c))
{
if (collapseWhiteSpace && (ignoreWhiteSpace || lnws < i - 2))
{
if (lnws - o >= 0 && lnws >= 0)
{
addMap(o, 1 + lnws);
}
o = ignoreWhiteSpace ? i : i - 1;
lnws = -1;
continue;
}
if ((c == '\r' || c == '\n') && (i < length))
{
c = Original[i];
if (c == '\r' || c == '\n')
{
end = trim ? lnws : i - 1;
if (end - o >= 0)
{
addMap(o, 1 + end);
}
i++;
o = i;
lnws = -1;
// TODO We have a line ending
continue;
}
}
if (trim && lnws < 0)
{
o = i;
}
}
else
{
lnws = i - 1;
}
}
end = ignoreWhiteSpace || trim ? lnws : i - 1;
if (end - o >= 0)
{
addMap(o, 1 + end);
}
// ReSharper restore EventExceptionNotDocumented
// Store map
Map = map.ToArray();
#endif
}
/// <summary>
/// Initializes a new instance of the <see cref="StringDifferences" /> class.
/// </summary>
/// <param name="a">The 'A' string.</param>
/// <param name="offsetA">The offset to the start of a window in the first string.</param>
/// <param name="lengthA">The length of the window in the first string.</param>
/// <param name="b">The 'B' string.</param>
/// <param name="offsetB">The offset to the start of a window in the second string.</param>
/// <param name="lengthB">The length of the window in the second string.</param>
/// <param name="textOptions">The text options.</param>
/// <param name="comparer">The character comparer.</param>
/// <exception cref="ArgumentNullException"><paramref name="a" /> is <see langword="null" />.</exception>
/// <exception cref="ArgumentNullException"><paramref name="b" /> is <see langword="null" />.</exception>
/// <exception cref="ArgumentNullException"><paramref name="comparer" /> is <see langword="null" />.</exception>
/// <exception cref="ArgumentOutOfRangeException">The <paramref name="offsetA" /> is out of range.</exception>
/// <exception cref="ArgumentOutOfRangeException">The <paramref name="lengthA" /> is out of range.</exception>
/// <exception cref="ArgumentOutOfRangeException">The <paramref name="offsetB" /> is out of range.</exception>
/// <exception cref="ArgumentOutOfRangeException">The <paramref name="lengthB" /> is out of range.</exception>
/// <exception cref="Exception">The <paramref name="comparer" /> throws an exception.</exception>
internal StringDifferences(
[NotNull] string a,
int offsetA,
int lengthA,
[NotNull] string b,
int offsetB,
int lengthB,
TextOptions textOptions,
[NotNull] Func <char, char, bool> comparer)
{
if (a == null)
{
throw new ArgumentNullException(nameof(a));
}
if (b == null)
{
throw new ArgumentNullException(nameof(b));
}
if (comparer == null)
{
throw new ArgumentNullException(nameof(comparer));
}
A = a;
B = b;
if (textOptions != TextOptions.None)
{
// Wrap the comparer with an additional check to handle special characters.
Func <char, char, bool> oc = comparer;
if (textOptions.HasFlag(TextOptions.IgnoreWhiteSpace))
{
// Ignore white space - treat all whitespace as the same (note this will handle line endings too).
comparer = (x, y) => char.IsWhiteSpace(x) ? char.IsWhiteSpace(y) : oc(x, y);
}
else if (textOptions.HasFlag(TextOptions.NormalizeLineEndings))
{
// Just normalize line endings - treat '\r' and '\n\ as the same
comparer = (x, y) => x == '\r' || x == '\n' ? y == '\r' || y == '\n' : oc(x, y);
}
}
// Map strings based on text options
StringMap aMap = a.ToMapped(textOptions);
StringMap bMap = b.ToMapped(textOptions);
// Perform diff on mapped string
Differences <char> chunks = aMap.Diff(bMap, comparer);
// Special case simple equality
if (chunks.Count < 2)
{
Chunk <char> chunk = chunks.Single();
// ReSharper disable once PossibleNullReferenceException
_chunks = new[] { new StringChunk(chunk.AreEqual, a, 0, b, 0) };
return;
}
// To reverse the mapping we first calculate the split points in the original strings, and find
// the last reference to the original strings in each chunk.
int[] aEnds = new int[chunks.Count];
int[] bEnds = new int[chunks.Count];
int lastA = 0;
int lastB = 0;
for (int i = 0; i < chunks.Count; i++)
{
Chunk <char> chunk = chunks[i];
ReadOnlyWindow <char> chunkA = chunk.A;
ReadOnlyWindow <char> chunkB = chunk.B;
if (chunk.A != null)
{
aEnds[i] = aMap.GetOriginalIndex(chunkA.Offset + chunkA.Count - 1) + 1;
lastA = i;
}
else
{
aEnds[i] = -1;
}
if (chunk.B != null)
{
bEnds[i] = bMap.GetOriginalIndex(chunkB.Offset + chunkB.Count - 1) + 1;
lastB = i;
}
else
{
bEnds[i] = -1;
}
}
// Now we're ready to build up a new chunk array based on the original strings
StringChunk[] stringChunks = new StringChunk[chunks.Count];
int aStart = 0;
int bStart = 0;
for (int i = 0; i < chunks.Count; i++)
{
int aEnd = i == lastA ? aMap.OriginalCount : aEnds[i];
int bEnd = i == lastB ? bMap.OriginalCount : bEnds[i];
string ac = aEnd > -1 ? a.Substring(aStart, aEnd - aStart) : null;
string bc = bEnd > -1 ? b.Substring(bStart, bEnd - bStart) : null;
stringChunks[i] = new StringChunk(chunks[i].AreEqual, ac, aEnd > -1 ? aStart : -1, bc, bEnd > -1 ? bStart : -1);
if (aEnd > -1)
{
aStart = aEnd;
}
if (bEnd > -1)
{
bStart = bEnd;
}
}
_chunks = stringChunks;
}
