2026.06.24.1 changes

PAMI/fuzzyFrequentPattern/basic/FFIMiner.py

@@ -164,7 +164,7 @@ class FFIMiner(_ab._fuzzyFrequentPatterns):
     _memoryRSS = float()
     _sep = "\t"
 
-    def __init__(self, iFile: str, minSup: float, sep: str="\t") -> None:
+    def __init__(self, iFile: str, minSup: float, sep: str = "\t", k: int = 1) -> None:
         super().__init__(iFile, minSup, sep)
         self._Database = None
         self._startTime = 0
@@ -173,6 +173,7 @@ def __init__(self, iFile: str, minSup: float, sep: str="\t") -> None:
         self._minSup = minSup
         self._iFile = iFile
         self._sep = sep
+        self._k = k
         self._finalPatterns = {}
         self._memoryUSS = 0
         self._memoryRSS = 0
@@ -219,7 +220,7 @@ def _creatingItemsets(self) -> None:
                 except IOError:
                     print("File Not Found")
                     quit()
-    
+
     def startMine(self):
         self.mine()
 
@@ -243,7 +244,7 @@ def _convert(self, value) -> Union[int, float]:
             else:
                 value = int(value)
         return value
-    
+
     def dfs(self, cands):
         """
         Perform depth-first search (DFS) to find frequent patterns in a database.
@@ -296,6 +297,20 @@ def mine(self):
                     items[item] = {}
                 items[item][lineNo] = fuzzyValue
 
+        if self._k >= 1:
+            #default k = 1 (max cardinality)
+            labelsByBaseItem = {}
+            for label in items:
+                token = label[0]
+                baseItem = token.rsplit('.', 1)[0] if '.' in token else token
+                labelsByBaseItem.setdefault(baseItem,[]).append(label)
+            for baseItem,labels in labelsByBaseItem.items():
+                if len(labels) <= self._k:
+                    continue
+                ranked = sorted(labels, key=lambda lbl: sum(items[lbl].values()), reverse=True)
+                for lbl in ranked[self._k:]:
+                    del items[lbl]
+
         self._minSup = self._convert(self._minSup)
         self._Database = items.copy()
 

PAMI/fuzzyGeoreferencedFrequentPattern/basic/FFSPMiner.py

@@ -278,13 +278,14 @@ class FFSPMiner(_ab._fuzzySpatialFrequentPatterns):
     _nFile = " "
     _sep = "\t"
 
-    def __init__(self, iFile: str, nFile: str, minSup: float, sep: str="\t") -> None:
+    def __init__(self, iFile: str, nFile: str, minSup: float, sep: str = "\t", k: int = 1) -> None:
         super().__init__(iFile, nFile, minSup, sep)
         self.oFile = None
         self._mapItemNeighbours = {}
         self._startTime = 0
         self._endTime = 0
         self._mapItemSum = {}
+        self._k = k
         self._joinsCnt = 0
         self._BufferSize = 200
         self._itemSetBuffer = []
@@ -446,6 +447,20 @@ def mine(self) -> None:
                     self._mapItemSum[item] += quantities[i]
                 else:
                     self._mapItemSum[item] = quantities[i]
+
+        if self._k >= 1:
+            #default k = 1 (max cardinality)
+            labelsByBaseItem = {}
+            for label in self._mapItemSum:
+                baseItem = label.rsplit('.', 1)[0] if '.' in label else label
+                labelsByBaseItem.setdefault(baseItem, []).append(label)
+            for baseItem, labels in labelsByBaseItem.items():
+                if len(labels) <= self._k:
+                    continue
+                ranked = sorted(labels, key=lambda lbl: self._mapItemSum[lbl], reverse=True)
+                for lbl in ranked[self._k:]:
+                    del self._mapItemSum[lbl]
+
         listOfFFList = []
         mapItemsToFFLIST = {}
         #self._minSup = self._convert(self._minSup)
@@ -466,7 +481,7 @@ def mine(self) -> None:
                 pair.item = items[i]
                 pair.quantity = quantities[i]
                 item = pair.item
-                if self._mapItemSum[item] >= self._minSup:
+                if self._mapItemSum.get(item, 0) >= self._minSup:
                     if pair.quantity > 0:
                         revisedTransaction.append(pair)
             revisedTransaction.sort(key=_ab._functools.cmp_to_key(self._compareItems))

PAMI/fuzzyGeoreferencedPeriodicFrequentPattern/basic/FGPFPMiner.py

@@ -268,13 +268,14 @@ class FGPFPMiner(_ab._fuzzySpatialFrequentPatterns):
     _sep = "\t"
 
 
-    def __init__(self, iFile, nFile, minSup, maxPer, sep):
+    def __init__(self, iFile, nFile, minSup, maxPer, sep, k=1):
         super().__init__(iFile, nFile, minSup, maxPer, sep)
         self.oFile = None
         self._mapItemNeighbours = {}
         self._startTime = 0
         self._endTime = 0
         self._itemsCnt = 0
+        self._k = k
         self._itemSupData = {}
         self._mapItemSum = {}
         self._joinsCnt = 0
@@ -463,6 +464,19 @@ def mine(self):
         Step3. At the end, sort the list of stored Candidate Frequent-Periodic Patterns in ascending order
         """
 
+        if self._k >= 1:
+            #default k = 1 (max cardinality)
+            labelsByBaseItem = {}
+            for label in self._itemSupData:
+                baseItem = label.rsplit('.', 1)[0] if '.' in label else label
+                labelsByBaseItem.setdefault(baseItem, []).append(label)
+            for baseItem, labels in labelsByBaseItem.items():
+                if len(labels) <= self._k:
+                    continue
+                ranked = sorted(labels, key=lambda lbl: self._itemSupData[lbl], reverse=True)
+                for lbl in ranked[self._k:]:
+                    del self._itemSupData[lbl]
+
         listOfFFList = []
         mapItemsToFFLIST = {}
         #region_label = []

PAMI/fuzzyPeriodicFrequentPattern/basic/FPFPMiner.py

@@ -309,12 +309,13 @@ class FPFPMiner(_ab._fuzzyPeriodicFrequentPatterns):
     _fuzzyValues = []
     _ts = []
 
-    def __init__(self, iFile: Union[str, _ab._pd.DataFrame], minSup: Union[int, float], period: Union[int, float], sep: str="\t") -> None:
+    def __init__(self, iFile: Union[str, _ab._pd.DataFrame], minSup: Union[int, float], period: Union[int, float], sep: str = "\t", k: int = 1) -> None:
         super().__init__(iFile, minSup, period, sep)
         self._oFile = ""
         self._BufferSize = 200
         self._itemSetBuffer = []
         self._mapItemSum = {}
+        self._k = k
         self._finalPatterns = {}
         self._joinsCnt = 0
         self._itemsCnt = 0
@@ -451,6 +452,20 @@ def mine(self) -> None:
                     self._mapItemSum[item] += quantities[i]
                 else:
                     self._mapItemSum[item] = quantities[i]
+
+        if self._k >= 1:
+            #default k = 1 (max cardinality)
+            labelsByBaseItem = {}
+            for label in self._mapItemSum:
+                baseItem = label.rsplit('.', 1)[0] if '.' in label else label
+                labelsByBaseItem.setdefault(baseItem, []).append(label)
+            for baseItem, labels in labelsByBaseItem.items():
+                if len(labels) <= self._k:
+                    continue
+                ranked = sorted(labels, key=lambda lbl: self._mapItemSum[lbl], reverse=True)
+                for lbl in ranked[self._k:]:
+                    del self._mapItemSum[lbl]
+
         listOfFFIList = []
         mapItemsToFFLIST = {}
         # self._minSup = self._convert(self._minSup)
@@ -475,7 +490,7 @@ def mine(self) -> None:
                 pair.item = items[i]
                 item = pair.item
                 pair.quantity = quantities[i]
-                if self._mapItemSum[item] >= self._minSup:
+                if self._mapItemSum.get(item, 0) >= self._minSup:
                     if pair.quantity > 0:
                         revisedTransaction.append(pair)
             revisedTransaction.sort(key=_ab._functools.cmp_to_key(self._compareItems))

setup.py

@@ -5,7 +5,7 @@
 
 setuptools.setup(
     name='pami',
-    version='2026.06.14.1',
+    version='2026.06.24.1',
     author='Rage Uday Kiran',
     author_email='uday.rage@gmail.com',
     description='This software is being developed at the University of Aizu, Aizu-Wakamatsu, Fukushima, Japan',