Merge branch 'main' into fix_statistics

data61 · Jul 7, 2023 · 79ea87a · 79ea87a
2 parents 61537f3 + 083ec4c
commit 79ea87a
Show file tree

Hide file tree

Showing 4 changed files with 7 additions and 7 deletions.
diff --git a/.github/workflows/typechecking.yml b/.github/workflows/typechecking.yml
@@ -18,7 +18,7 @@ jobs:
     - name: mypy (pinned)
       run: |
         pip install --upgrade pip
-        pip install -U mypy==0.931
+        pip install -U mypy==1.3.0
         pip install -U pydantic
         pip install -U types-setuptools
         mypy blocklib --ignore-missing-imports --no-implicit-optional --disallow-untyped-calls

diff --git a/blocklib/pprlpsig.py b/blocklib/pprlpsig.py
@@ -116,11 +116,11 @@ def filter_reversed_index(self, data: Sequence[Sequence], reversed_index: Dict):
         if filter_type == "ratio":
             min_occur_ratio = self.filter_config.min
             max_occur_ratio = self.filter_config.max
-            reversed_index = {k: v for k, v in reversed_index.items() if n * max_occur_ratio > len(v) > n * min_occur_ratio}
+            reversed_index = {k: v for k, v in reversed_index.items() if n * max_occur_ratio >= len(v) >= n * min_occur_ratio}
         elif filter_type == "count":
             min_occur_count = self.filter_config.min
             max_occur_count = self.filter_config.max
-            reversed_index = {k: v for k, v in reversed_index.items() if max_occur_count > len(v) > min_occur_count}
+            reversed_index = {k: v for k, v in reversed_index.items() if max_occur_count >= len(v) >= min_occur_count}
         else:
             raise NotImplementedError("Don't support {} filter yet.".format(filter_type))
 

diff --git a/tests/test_candidate_block_generator.py b/tests/test_candidate_block_generator.py
@@ -52,7 +52,7 @@ def test_generate_candidate_blocks_psig(self):
             "record-id-col": 0,
             "filter": {
                 "type": "ratio",
-                "max": 0.5,
+                "max": 0.49,
                 "min": 0.0,
             },
             "blocking-filter": {

diff --git a/tests/test_pprlpsig.py b/tests/test_pprlpsig.py
@@ -67,7 +67,7 @@ def test_build_reversed_index(self):
             "record-id-col": 0,
             "filter": {
                 "type": "ratio",
-                "max": 0.5,
+                "max": 0.49,
                 "min": 0.2,
             },
             "blocking-filter": {
@@ -102,7 +102,7 @@ def test_build_reversed_index_feature_name(self):
             "record-id-col": 0,
             "filter": {
                 "type": "ratio",
-                "max": 0.5,
+                "max": 0.49,
                 "min": 0.2,
             },
             "blocking-filter": {
@@ -133,7 +133,7 @@ def test_build_reversed_index_feature_name(self):
             "record-id-col": 0,
             "filter": {
                 "type": "ratio",
-                "max": 0.5,
+                "max": 0.49,
                 "min": 0.2,
             },
             "blocking-filter": {