replacing submodule

jamesturk · Sep 7, 2024 · 1dc5bd5 · 1dc5bd5
1 parent 3370375
commit 1dc5bd5
Show file tree

Hide file tree

Showing 16 changed files with 23,708 additions and 1 deletion.
diff --git a/csvs/README.md b/csvs/README.md
@@ -0,0 +1 @@
+Test data for jellyfish string comparison and phonetic encoding algorithms.
diff --git a/csvs/damerau_levenshtein.csv b/csvs/damerau_levenshtein.csv
@@ -0,0 +1,9 @@
+,,0
+abc,,3
+bc,abc,1
+fuor,four,1
+abcd,acb,2
+cape sand recycling ,edith ann graham,17
+jellyifhs,jellyfish,2
+ifhs,fish,2
+"Hello, world!","Hello,Â world!",2
diff --git a/csvs/hamming.csv b/csvs/hamming.csv
@@ -0,0 +1,8 @@
+,,0
+,abc,3
+abc,abc,0
+acc,abc,1
+abcd,abc,1
+abc,abcd,1
+testing,this is a test,13
+Saturday,Sunday,7
diff --git a/csvs/jaccard.csv b/csvs/jaccard.csv
@@ -0,0 +1,16 @@
+abc,xyz,0.0,
+abc,abc,1.0,
+abc,abcd,0.0,
+abcd,abce,0.0,
+abcd,abcde,0.0,
+french,quebec,0.0,
+france,quebec,0.0,
+france,france,1.0,
+The quick brown fox jumps over the lazy dog,The quick brown fox jumps over the lazy cat,0.8,
+The quick brown fox jumps over the lazy dog,The slow green turtle crawls under the lazy cat,0.2,
+John Smith,Smith; John,0.33333,
+John Smith,Smith John,1.0,
+John Smith,John Jacob Smith,0.666667,
+night,nacht,0.0,
+night,nacht,0.2,2
+night,nacht,0.33333,3
diff --git a/csvs/jaro_distance.csv b/csvs/jaro_distance.csv
@@ -0,0 +1,5 @@
+dixon,dicksonx,0.767
+martha,marhta,0.944
+dwayne,duane,0.822
+0ð00,0ð00,1
+"Sint-Pietersplein 6, 9000 Gent","Test 10, 1010 Brussel",0.518
diff --git a/csvs/jaro_winkler.csv b/csvs/jaro_winkler.csv
@@ -0,0 +1,17 @@
+dixon,dicksonx,0.813
+martha,marhta,0.961
+dwayne,duane,0.84
+William,Williams,0.975
+,foo,0
+a,a,1
+abc,xyz,0
+aaaa,aaaaa,0.96
+orangutan-kumquat,orangutan kumquat,0.976
+jaz,jal,0.822
+@,@@,0.85
+0,0@,0.85
+a,ab,0.85
+012345,0123456,0.971
+012abc,012abcd,0.971
+012abc,013abcd,0.879
+a1bc,a1be,0.883
diff --git a/csvs/jaro_winkler_longtol.csv b/csvs/jaro_winkler_longtol.csv
@@ -0,0 +1,10 @@
+dixon,dicksonx,0.830
+martha,marhta,0.971
+dwayne,duane,0.869
+William,Williams,0.980
+,foo,0
+a,a,1
+abc,xyz,0
+aaaa,aaaaa,0.96
+orangutan-kumquat,orangutan kumquat,0.986
+1abcdefg,1abcdefh,0.96
diff --git a/csvs/levenshtein.csv b/csvs/levenshtein.csv
@@ -0,0 +1,6 @@
+,,0
+abc,,3
+,abc,3
+bc,abc,1
+kitten,sitting,3
+Saturday,Sunday,3
diff --git a/csvs/match_rating_codex.csv b/csvs/match_rating_codex.csv
@@ -0,0 +1,12 @@
+Byrne,BYRN
+Boern,BRN
+Smith,SMTH
+Smyth,SMYTH
+Catherine,CTHRN
+Kathryn,KTHRYN
+Kathrynoglin,KTHGLN
+Ad,AD
+Ed,ED
+William,WLM
+ä,Ä
+Frédéric,FRÉÉRC
diff --git a/csvs/match_rating_comparison.csv b/csvs/match_rating_comparison.csv
@@ -0,0 +1,7 @@
+Bryne,Boern,True
+Smith,Smyth,True
+Catherine,Kathryn,True
+Michael,Mike,False
+Tim,Timothy,None
+Ed,Ad,True
+Marie Helene,Maria Rio,True
diff --git a/csvs/metaphone.csv b/csvs/metaphone.csv
@@ -0,0 +1,33 @@
+DGIB,JB
+metaphone,MTFN
+wHErE,WR
+shell,XL
+this is a difficult string,0S IS A TFKLT STRNK
+aeromancy,ERMNS
+Antidisestablishmentarianism,ANTTSSTBLXMNTRNSM
+sunlight labs,SNLT LBS
+sonlite laabz,SNLT LBS
+Çáŕẗéř,KRTR
+kentucky,KNTK
+KENTUCKY,KNTK
+NXNXNX,NKSNKSNKS
+Aapti,PT
+Aarti,RT
+CIAB,XB
+NQ,NK
+sian,XN
+gek,JK
+Hb,HB
+Bho,BH
+Tiavyi,XFY
+Xhot,XHT
+Xnot,SNT
+g,K
+8 queens,KNS
+Utah,UT
+WH,W
+walt,WLT
+ANDREW,ANTR
+why,W
+whynot,WNT
+acceptingness,AKSPTNKNS
diff --git a/csvs/nysiis.csv b/csvs/nysiis.csv
@@ -0,0 +1,36 @@
+Worthy,WARTY
+Ogata,OGAT
+montgomery,MANTGANARY
+Costales,CASTAL
+Tu,T
+martincevic,MARTANCAFAC
+Catherine,CATARAN
+Katherine,CATARAN
+Katerina,CATARAN
+Johnathan,JANATAN
+Jonathan,JANATAN
+John,JAN
+Teresa,TARAS
+Theresa,TARAS
+Jessica,JASAC
+Joshua,JAS
+Bosch,BAS
+Lapher,LAFAR
+wiyh,WY
+MacArthur,MCARTAR
+Pheenard,FANAD
+Schmittie,SNATY
+Knaqze,NAGS
+Knokno,NAN
+Knoko,NAC
+Macaw,MC
+,
+T,T
+S,S
+P,P
+K,C
+M,M
+E,E
+PFEISTER,FASTAR
+SARAH,SAR
+ç,Ç
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1 @@
		Test data for jellyfish string comparison and phonetic encoding algorithms.