vlib.regex module in pure V

2023-08-10 21:13:21 +03:00 · 2020-01-13 13:30:41 +01:00 · 2020-01-13 13:30:41 +01:00 · 8ea0c08a38
commit 8ea0c08a38
parent 6733b12ec3
2 changed files with 2158 additions and 0 deletions
--- a/vlib/regex/regex.v
+++ b/vlib/regex/regex.v
--- a/vlib/regex/regex_test.v
+++ b/vlib/regex/regex_test.v
@ -0,0 +1,157 @@
+import regex
+
+struct TestItem {
+	src string
+	q string
+	s int = 0
+	e int = 0
+}
+
+const(
+match_test_suite = [
+
+	// positive
+	TestItem{"this is a good.",r"this",0,4},
+	TestItem{"this is a good.",r"good",10,14},
+	TestItem{"this is a good.",r"go+d",10,14},
+	TestItem{"this is a good.",r"g[oae]+d",10,14},
+	TestItem{"this is a goed.",r"g[oae]+d",10,14},
+	TestItem{"this is a good.",r"g[oae]*d",10,14},
+	TestItem{"this is a goaezd.",r"g[ea-cm-z]*d",10,16},
+	TestItem{"this is a good.",r"this (\w+) a",0,9},
+	TestItem{"this is a good.",r"this( \w+){2} g",0,11},
+	TestItem{"this is a good.",r"( ?\w+){,1}",0,4},
+	TestItem{"this is a good.",r"( ?\w+)+",0,14},
+	TestItem{"this is a good.",r"this( \w+)+",0,14},
+	TestItem{"this is a good sample.",r"( ?\w+){,2}",0,7},
+	TestItem{"this is a good sample.",r"( ?\w+){,3}",0,9},
+	TestItem{"this is a good sample.",r"( ?\w+){,4}",0,14},
+	TestItem{"this is a good sample.",r"( ?\w+){,5}",0,21},
+	TestItem{"this is a good sample.",r"( ?\w+){2,3}",0,9},
+	TestItem{"this is a good sample.",r"(\s?\w+){2,3}",0,9},
+	TestItem{"this is a good sample.",r".*i(\w)+",0,4},
+	TestItem{"this these those.",r"(th[ei]se?\s|\.)+",0,11},
+	TestItem{"this these those ",r"(th[eio]se? ?)+",0,17},
+	TestItem{"this these those ",r"(th[eio]se? )+",0,17},
+	TestItem{"this,these,those. over",r"(th[eio]se?[,. ])+",0,17},
+	TestItem{"soday,this,these,those. over",r"(th[eio]se?[,. ])+",6,23},
+	TestItem{"soday,this,these,those. over",r".*,(th[eio]se?[,. ])+",0,23},
+	TestItem{"soday,this,these,thesa.thesi over",r".*,(th[ei]se?[,. ])+(thes[ai][,. ])+",0,29},
+	TestItem{"cpapaz",r"(c(pa)+z)",0,6},
+	TestItem{"this is a cpapaz over",r"(c(pa)+z)",10,16},
+	TestItem{"this is a cpapapez over",r"(c(p[ae])+z)",10,18},
+	TestItem{"test@post.pip.com",r"[a-z0-9_]+@([a-z0-9_]+\.?)+",0,17},
+	TestItem{"test1@post.pip.com, pera",r"[\w]+@([\w]+\.)+\w+",0,18},
+	TestItem{"pippo@pera.com ",r"[a-z0-9_]+@([a-z0-9_]+\.?)+",0,14},
+	TestItem{"adce aabe",r"(a(ab)+)|(a(dc)+)e",0,4},
+	TestItem{"zadce aabe",r"(a(ab)+)|(a(dc)+)e",1,5},
+	TestItem{"abbz accz addz.",r"c|(d)|e|(ab+)",0,3},
+	TestItem{"this those these ciao",r"((t[hieo]+se?)\s*)+",0,17},
+	TestItem{"this ciao",r"((t[hieo]+se?)\s*)+",0,5},
+	TestItem{"this cpapaz adce aabe",r"(c(pa)+z)(\s[\a]+){2}",5,21},
+	TestItem{"1234this cpapaz adce aabe",r"(c(pa)+z)(\s[\a]+){2}$",9,25},
+	TestItem{"this cpapaz adce aabe third",r"(c(pa)+z)(\s[\a]+){2}",5,21},
+	TestItem{"cpapaz ole. pippo,",r".*(c(pa)+z)(\s+\a+[\.,]?)+",0,18},
+	TestItem{"cpapaz ole. pippo",r".*(c(pa)+z)(\s+\a+[\.,]?)+",0,17},
+	TestItem{"cpapaz ole. pippo, 852",r".*(c(pa)+z)(\s+\a+[\.,]?)+",0,18},
+	TestItem{"123cpapaz ole. pippo",r".*(c(pa)+z)(\s+\a+[\.,]?)+",0,20},
+	TestItem{"...cpapaz ole. pippo",r".*(c(pa)+z)(\s+\a+[\.,]?)+",0,20},
+	TestItem{"123cpapaz ole. pippo",r"(c(pa)+z)(\s+\a+[\.,]?)+",3,20},
+	TestItem{"cpapaz ole. pippo,",r".*c.+ole.*pi",0,14},
+	TestItem{"cpapaz ole. pipipo,",r".*c.+ole.*p([ip])+o",0,18},
+	TestItem{"cpapaz ole. pipipo",r"^.*c.+ol?e.*p([ip])+o$",0,18},
+
+	// negative
+	TestItem{"zthis ciao",r"((t[hieo]+se?)\s*)+",-1,0},
+	TestItem{"this is a good.",r"thes",-1,0},
+	TestItem{"test1post.pip.com, pera",r"[\w]+@([\w]+\.)+\w+",-1,0},
+	TestItem{"this cpapaz adce",r"(c(pa)+z)(\s[\a]+){2}",-1,0},
+	TestItem{"this cpapaz adce aabe third",r"(c(pa)+z)(\s[\a]+){2}$",-1,0},
+	TestItem{"1234this cpapaz adce aabe ter",r"(c(pa)+z)(\s[\a]+){2}$",-1,0},
+	TestItem{"cpapaz ole. pipipo,",r"^.*c.+ol?e.*p([ip])+o$",-1,0},
+	
+
+	// check unicode
+	TestItem{"this is a Ⅰ Ⅱ Ⅲ Ⅳ Ⅴ Ⅵ test",r".*a [Ⅰ-Ⅵ ]+",0,34},
+	TestItem{"123Ⅰ Ⅱ Ⅲ Ⅳ Ⅴ Ⅵ test",r"[Ⅰ-Ⅴ\s]+",3,23},
+]
+)
+
+fn test_regex(){
+	for c,to in match_test_suite {
+		// debug print
+		//println("#$c [$to.src] q[$to.q] $to.s")
+
+		// test the find
+		if to.s > 0 {
+			mut re, re_err, err_pos := regex.regex(to.q)
+			if re_err == regex.COMPILE_OK {
+				//q_str := re.get_query()
+				//println("Query: $q_str")
+				start,end := re.find(to.src)
+
+				if start != to.s || end != to.e {
+					err_str := re.get_parse_error_string(start)
+					println("ERROR : $err_str")
+					assert false
+				} else {
+					//tmp_str := text[start..end]
+					//println("found in [$start, $end] => [$tmp_str]")
+					assert true
+				}
+
+			} else {
+				println("query: $to.q")
+				lc := "-".repeat(err_pos-1)
+				println("err  : $lc^")
+				err_str := re.get_parse_error_string(re_err)
+				println("ERROR: $err_str")
+				assert false
+			}
+			continue
+		}
+		
+		// test the match
+		mut re := regex.new_regex()
+		//re.debug = true
+		
+		re_err,err_pos := re.compile(to.q)
+		if re_err == regex.COMPILE_OK {
+			//println("#$c [$to.src] q[$to.q]")
+			start, end := re.match_string(to.src)
+
+			mut tmp_str := ""
+			if start >= 0 && end  > start{
+				tmp_str = to.src[start..end]
+			}
+			
+			if start != to.s || end != to.e {
+				println("#$c [$to.src] q[$to.q] res[$tmp_str] $start, $end")	
+				println("ERROR!")
+				C.printf("ERROR!! res:(%d, %d) refh:(%d, %d)\n",start, end, to.s, to.e)
+				assert false
+				break
+			} else {
+				assert true
+			}
+
+			// rerun to test consistency
+			tmp_str1 := to.src.clone()
+			start1, end1 := re.match_string(tmp_str1)
+			if start1 != start || end1 != end {
+				println("two run ERROR!!")
+				assert false
+				break
+			}
+
+		} else {
+			println("query: $to.q")
+			lc := "-".repeat(err_pos-1)
+			println("err  : $lc")
+			err_str := re.get_parse_error_string(re_err)
+			println("ERROR: $err_str")
+			assert false
+			break
+		}
+	}
+}