fix wrong result when allowing only deletions

Also fix several buggy tests which this fix surfaced.
taleinat · Apr 6, 2020 · 8d13f0a · 8d13f0a
1 parent 9dab06d
commit 8d13f0a
Show file tree

Hide file tree

Showing 5 changed files with 29 additions and 19 deletions.
diff --git a/src/fuzzysearch/_generic_search.c b/src/fuzzysearch/_generic_search.c
diff --git a/src/fuzzysearch/_generic_search.pyx b/src/fuzzysearch/_generic_search.pyx
@@ -178,7 +178,7 @@ cdef _c_find_near_matches_generic_linear_programming(
                         # if skipping n_dels sub-sequence chars reaches the end
                         # of the sub-sequence, yield a match
                         if cand.subseq_index + n_skipped == subseq_len:
-                            add_match(cand.start, index + 1, cand.l_dist + n_skipped)
+                            add_match(cand.start, index, cand.l_dist + n_skipped)
                             break
                         # otherwise, if skipping n_skipped sub-sequence chars
                         # reaches a sub-sequence char identical to this sequence
@@ -187,7 +187,7 @@ cdef _c_find_near_matches_generic_linear_programming(
                             # if this is the last char of the sub-sequence, yield
                             # a match
                             if cand.subseq_index + n_skipped + 1 == subseq_len:
-                                add_match(cand.start, index + 1, cand.l_dist + n_skipped)
+                                add_match(cand.start, index, cand.l_dist + n_skipped)
                             # otherwise add a candidate skipping n_skipped
                             # subsequence chars
                             else:

diff --git a/src/fuzzysearch/generic_search.py b/src/fuzzysearch/generic_search.py
@@ -142,7 +142,7 @@ def make_match(start, end, dist):
                     # if skipping n_dels sub-sequence chars reaches the end
                     # of the sub-sequence, yield a match
                     if cand.subseq_index + n_skipped == subseq_len:
-                        yield make_match(cand.start, index + 1,
+                        yield make_match(cand.start, index,
                                          cand.l_dist + n_skipped)
                         break
                     # otherwise, if skipping n_skipped sub-sequence chars
@@ -152,7 +152,7 @@ def make_match(start, end, dist):
                         # if this is the last char of the sub-sequence, yield
                         # a match
                         if cand.subseq_index + n_skipped + 1 == subseq_len:
-                            yield make_match(cand.start, index + 1,
+                            yield make_match(cand.start, index,
                                              cand.l_dist + n_skipped)
                         # otherwise add a candidate skipping n_skipped
                         # subsequence chars

diff --git a/tests/test_generic_search.py b/tests/test_generic_search.py
@@ -257,6 +257,16 @@ def test_no_deletion(self):
             []
         )
 
+    def test_only_deletions(self):
+        self.expectedOutcomes(
+            self.search(b('TESTabc'), b('TEST123'), 0, 0, 5, None),
+            [
+                Match(start=0, end=4, dist=3, matched=b('TEST')),
+                Match(start=1, end=4, dist=4, matched=b('EST')),
+                Match(start=2, end=4, dist=5, matched=b('ST')),
+            ],
+        )
+
     def test_invalid_none_arguments(self):
         # check that an exception is raised when max_l_dist is None as well as
         # at least one other limitation
@@ -306,8 +316,8 @@ def test_non_string_sequences(self):
                                       [Match(start=0, end=3, dist=1,
                                              matched=klass([1, 2, 4]))])
                 self.expectedOutcomes(self.search(klass([1, 2, 3]), klass([1, 2, 4]), 0, 0, 1, 1),
-                                      [Match(start=0, end=3, dist=1,
-                                             matched=klass([1, 2, 4]))])
+                                      [Match(start=0, end=2, dist=1,
+                                             matched=klass([1, 2]))])
 
     def test_list_of_words_one_missing(self):
         subsequence = "jumped over the a lazy dog".split()
@@ -427,7 +437,7 @@ def test_missing_second_item_complex(self):
                 Match(start=1, end=5, dist=1, matched=b('bcde')),
                 Match(start=2, end=5, dist=1, matched=b('cde')),
                 Match(start=3, end=5, dist=1, matched=b('de')),
-                Match(start=2, end=5, dist=3, matched=b('bcd')),
+                Match(start=2, end=5, dist=2, matched=b('bcd')),
             }.issubset(set(
                 self.search(b('bde'), b('abcdefg'), 1, 1, 1, 3),
             ))

diff --git a/tests/test_generic_search_cython.py b/tests/test_generic_search_cython.py
@@ -125,7 +125,7 @@ def test_missing_second_item_complex(self):
                     Match(start=1, end=5, dist=1, matched=b('bcde')),
                     Match(start=2, end=5, dist=1, matched=b('cde')),
                     Match(start=3, end=5, dist=1, matched=b('de')),
-                    Match(start=2, end=5, dist=3, matched=b('cde')),
+                    Match(start=2, end=5, dist=2, matched=b('bcd')),
                 }.issubset(set(
                     self.search(b('bde'), b('abcdefg'), 1, 1, 1, 3),
                 ))