github_mirrors/vectorscan
3
0
Fork 1
mirror of https://github.com/VectorCamp/vectorscan.git synced 2025-10-01 12:07:42 +03:00
Code Issues Packages Projects Releases Wiki Activity

Add support for Hamming distance approx matching

Browse Source
This commit is contained in:
Matthew Barr
2017-10-10 15:26:35 +11:00
committed by Xiang Wang
parent 5827bd1c2b
commit 1891f14755
13 changed files with 169 additions and 56 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
src/compiler/compiler.cpp
View File

@@ -78,7 +78,8 @@ void validateExt(const hs_expr_ext &ext) {
static const unsigned long long ALL_EXT_FLAGS = HS_EXT_FLAG_MIN_OFFSET |
HS_EXT_FLAG_MAX_OFFSET |
HS_EXT_FLAG_MIN_LENGTH |
HS_EXT_FLAG_EDIT_DISTANCE;
HS_EXT_FLAG_EDIT_DISTANCE |
HS_EXT_FLAG_HAMMING_DISTANCE;
if (ext.flags & ~ALL_EXT_FLAGS) {
throw CompileError("Invalid hs_expr_ext flag set.");
}
@@ -96,6 +97,13 @@ void validateExt(const hs_expr_ext &ext) {
throw CompileError("In hs_expr_ext, min_length must be less than or "
"equal to max_offset.");
}
if ((ext.flags & HS_EXT_FLAG_EDIT_DISTANCE) &&
(ext.flags & HS_EXT_FLAG_HAMMING_DISTANCE)) {
throw CompileError("In hs_expr_ext, cannot have both edit distance and "
"Hamming distance.");
}
}
ParsedExpression::ParsedExpression(unsigned index_in, const char *expression,
@@ -103,7 +111,7 @@ ParsedExpression::ParsedExpression(unsigned index_in, const char *expression,
const hs_expr_ext *ext)
: expr(index_in, flags & HS_FLAG_ALLOWEMPTY, flags & HS_FLAG_SINGLEMATCH,
false, flags & HS_FLAG_PREFILTER, SOM_NONE, report, 0, MAX_OFFSET,
0, 0) {
0, 0, 0) {
ParseMode mode(flags);
component = parse(expression, mode);
@@ -158,6 +166,9 @@ ParsedExpression::ParsedExpression(unsigned index_in, const char *expression,
if (ext->flags & HS_EXT_FLAG_EDIT_DISTANCE) {
expr.edit_distance = ext->edit_distance;
}
if (ext->flags & HS_EXT_FLAG_HAMMING_DISTANCE) {
expr.hamm_distance = ext->hamming_distance;
}
}
// These are validated in validateExt, so an error will already have been

7
src/compiler/expression_info.h
View File

@@ -45,11 +45,13 @@ public:
ExpressionInfo(unsigned int index_in, bool allow_vacuous_in,
bool highlander_in, bool utf8_in, bool prefilter_in,
som_type som_in, ReportID report_in, u64a min_offset_in,
u64a max_offset_in, u64a min_length_in, u32 edit_distance_in)
u64a max_offset_in, u64a min_length_in, u32 edit_distance_in,
u32 hamm_distance_in)
: index(index_in), report(report_in), allow_vacuous(allow_vacuous_in),
highlander(highlander_in), utf8(utf8_in), prefilter(prefilter_in),
som(som_in), min_offset(min_offset_in), max_offset(max_offset_in),
min_length(min_length_in), edit_distance(edit_distance_in) {}
min_length(min_length_in), edit_distance(edit_distance_in),
hamm_distance(hamm_distance_in) {}
/**
* \brief Index of the expression represented by this graph.
@@ -95,6 +97,7 @@ public:
* 0 if not used.
*/
u32 edit_distance;
u32 hamm_distance;
};
}