Source code for PAMI.faultTolerantFrequentPattern.basic.FTFPGrowth

# FTFPGrowth algorithm aims to discover all fault-tolerant frequent patterns that may exist in a transactional database.
#
# **Importing this algorithm into a python program**
# --------------------------------------------------
#
#
#             from PAMI.faultTolerantFrequentPattern.basic import FTFPGrowth as alg
#
#             obj = alg.FTFPGrowth(inputFile,minSup,itemSup,minLength,faultTolerance)
#
#             obj.mine()
#
#             faultTolerantFrequentPatterns = obj.getPatterns()
#
#             print("Total number of fault-tolerant frequent patterns:", len(faultTolerantFrequentPatterns))
#
#             obj.save(oFile)
#
#             Df = obj.getPatternInDataFrame()
#
#             memUSS = obj.getMemoryUSS()
#
#             print("Total Memory in USS:", memUSS)
#
#             memRSS = obj.getMemoryRSS()
#
#             print("Total Memory in RSS", memRSS)
#
#             run = obj.getRuntime()
#
#             print("Total ExecutionTime in seconds:", run)
#




__copyright__ = """
Copyright (C)  2021 Rage Uday Kiran

     This program is free software: you can redistribute it and/or modify
     it under the terms of the GNU General Public License as published by
     the Free Software Foundation, either version 3 of the License, or
     (at your option) any later version.

     This program is distributed in the hope that it will be useful,
     but WITHOUT ANY WARRANTY; without even the implied warranty of
     MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
     GNU General Public License for more details.

     You should have received a copy of the GNU General Public License
     along with this program.  If not, see <https://www.gnu.org/licenses/>.
     Copyright (C)  2021 Rage Uday Kiran
     
"""



from PAMI.faultTolerantFrequentPattern.basic import abstract as _fp
from typing import List, Dict, Tuple, Set, Union, Any, Generator
import pandas as pd
from deprecated import deprecated

_minSup = str()
_fp._sys.setrecursionlimit(20000)


class _Node:
    """
    A class used to represent the node of frequentPatternTree

    :Attributes:

        itemId: int
            storing item of a node
        counter: int
            To maintain the support of node
        parent: node
            To maintain the parent of node
        children: list
            To maintain the children of node

    :Methods:

        addChild(node)
            Updates the nodes children list and parent for the given node
    """

    def __init__(self, item: int, children: Dict[int, '_Node']) -> None:
        self.itemId = item
        self.counter = 1
        self.parent = None
        self.children = children

    def addChild(self, node: '_Node') -> None:
        """
        Retrieving the child from the tree

        :param node: Child Node

        :type node: Nodes

        :return: Updates the children nodes and parent nodes

        """
        self.children[node.itemId] = node
        node.parent = self


class _Tree:
    """
    A class used to represent the frequentPatternGrowth tree structure

    :Attributes:

        root : Node
            The first node of the tree set to Null.
        summaries : dictionary
            Stores the nodes itemId which shares same itemId
        info : dictionary
            frequency of items in the transactions

    :Methods:

        addTransaction(transaction, freq)
            adding items of  transactions into the tree as nodes and freq is the count of nodes
        getFinalConditionalPatterns(node)
            getting the conditional patterns from fp-tree for a node
        getConditionalPatterns(patterns, frequencies)
            sort the patterns by removing the items with lower minSup
        generatePatterns(prefix)
            generating the patterns from fp-tree
    """

    def __init__(self) -> None:
        self.headerList = []
        self.root = _Node(None, {})
        self.summaries = {}
        self.info = {}

    def addTransaction(self, transaction: List[int], count: int) -> None:
        """
        Adding transaction into tree

        :param transaction: it represents the one transaction in database

        :type transaction: list

        :param count: frequency of item

        :type count: int

        :return: None

        """

        # This method takes transaction as input and returns the tree
        currentNode = self.root
        for i in range(len(transaction)):
            if transaction[i] not in currentNode.children:
                newNode = _Node(transaction[i], {})
                newNode.freq = count
                currentNode.addChild(newNode)
                if transaction[i] in self.summaries:
                    self.summaries[transaction[i]].append(newNode)
                else:
                    self.summaries[transaction[i]] = [newNode]
                currentNode = newNode
            else:
                currentNode = currentNode.children[transaction[i]]
                currentNode.freq += count

    def getFinalConditionalPatterns(self, alpha: int) -> Tuple[List[List[int]], List[int], Dict[int, int]]:
        """
        Generates the conditional patterns for a node

        :param alpha: node to generate conditional patterns

        :type alpha: int

        :return:returns conditional patterns, frequency of each item in conditional patterns

        :rtype: tuple, List, Dict

        """
        finalPatterns = []
        finalFreq = []
        for i in self.summaries[alpha]:
            set1 = i.freq
            set2 = []
            while i.parent.itemId is not None:
                set2.append(i.parent.itemId)
                i = i.parent
            if len(set2) > 0:
                set2.reverse()
                finalPatterns.append(set2)
                finalFreq.append(set1)
        finalPatterns, finalFreq, info = self.getConditionalTransactions(finalPatterns, finalFreq)
        return finalPatterns, finalFreq, info

    @staticmethod
    def getConditionalTransactions(ConditionalPatterns: List[List[int]], conditionalFreq: List[int]) -> Tuple[List[List[int]], List[int], Dict[int, int]]:
        """
        To calculate the frequency of items in conditional patterns and sorting the patterns

        :Parameters:
            ConditionalPatterns: paths of a node
            conditionalFreq: frequency of each item in the path

        :type ConditionalPatterns: List

        :type conditionalFreq: List

        :return: conditional patterns and frequency of each item in transactions

        :rtype: Tuple, List, Dict

        """
        global _minSup
        pat = []
        freq = []
        data1 = {}
        for i in range(len(ConditionalPatterns)):
            for j in ConditionalPatterns[i]:
                if j in data1:
                    data1[j] += conditionalFreq[i]
                else:
                    data1[j] = conditionalFreq[i]
        up_dict = {k: v for k, v in data1.items() if v >= _minSup}
        count = 0
        for p in ConditionalPatterns:
            p1 = [v for v in p if v in up_dict]
            trans = sorted(p1, key=lambda x: (up_dict.get(x), -x), reverse=True)
            if len(trans) > 0:
                pat.append(trans)
                freq.append(conditionalFreq[count])
            count += 1
        return pat, freq, up_dict

    def generatePatterns(self, prefix: List[int]) -> Generator[Tuple[List[int], int], None, None]:
        """
        To generate the frequent patterns

        :parameters:
            prefix: an empty list

        :type prefix: List

        :return:Frequent patterns that are extracted from fp-tree

        :rtype: Typle, None, None
        """
        for i in sorted(self.summaries, key=lambda x: (self.info.get(x), -x)):
            pattern = prefix[:]
            pattern.append(i)
            yield pattern, self.info[i]
            patterns, freq, info = self.getFinalConditionalPatterns(i)
            conditionalTree = _Tree()
            conditionalTree.info = info.copy()
            for pat in range(len(patterns)):
                conditionalTree.addTransaction(patterns[pat], freq[pat])
            if len(patterns) > 0:
                for q in conditionalTree.generatePatterns(pattern):
                    yield q


[docs] class FTFPGrowth(_fp._faultTolerantFrequentPatterns): """ :Description: FPGrowth is one of the fundamental algorithm to discover frequent patterns in a transactional database. It stores the database in compressed fp-tree decreasing the memory usage and extracts the patterns from tree.It employs downward closure property to reduce the search space effectively. :Reference: Han, J., Pei, J., Yin, Y. et al. Mining Frequent Patterns without Candidate Generation: A Frequent-Pattern Tree Approach. Data Mining and Knowledge Discovery 8, 53–87 (2004). https://doi.org/10.1023 :param iFile: file : Name of the Input file to mine complete set of fault Tolerant frequent patterns :param oFile: str : Name of the output file to store complete set of falut Tolerant frequent patterns :param minSup: float or int or str : The user can specify minSup either in count or proportion of database size. If the program detects the data type of minSup is integer, then it treats minSup is expressed in count. Otherwise, it will be treated as float. Example: minSup=10 will be treated as integer, while minSup=10.0 will be treated as float :param sep : str : This variable is used to distinguish items from one another in a transaction. The default separator is tab space or \t. However, the users can override their default separator. :Attributes: startTime: float : To record the start time of the mining process endTime: float : To record the completion time of the mining process memoryUSS : float To store the total amount of USS memory consumed by the program memoryRSS : float To store the total amount of RSS memory consumed by the program Database : list To store the transactions of a database in list mapSupport : Dictionary To maintain the information of item and their frequency lno : int it represents the total no of transactions tree : class it represents the Tree class finalPatterns : dict it represents to store the patterns :Methods: mine() Mining process will start from here getPatterns() Complete set of patterns will be retrieved with this function save(oFile) Complete set of frequent patterns will be loaded in to an output file getPatternsAsDataFrame() Complete set of frequent patterns will be loaded in to a dataframe getMemoryUSS() Total amount of USS memory consumed by the mining process will be retrieved from this function getMemoryRSS() Total amount of RSS memory consumed by the mining process will be retrieved from this function getRuntime() Total amount of runtime taken by the mining process will be retrieved from this function creatingItemSets() Scans the dataset or dataframes and stores in list format frequentOneItem() Extracts the one-frequent patterns from transactions **Executing the code on terminal:** ---------------------------------------- .. code-block:: console Format: (.venv) $ python3 FPGrowth.py <inputFile> <outputFile> <minSup> Example Usage: (.venv) $ python3 FPGrowth.py sampleDB.txt patterns.txt 10.0 .. note:: minSup will be considered in times of minSup and count of database transactions **Sample run of the importing code:** ------------------------------------------- .. code-block:: python from PAMI.faultTolerantFrequentPattern.basic import FTFPGrowth as alg obj = alg.FTFPGrowth(inputFile,minSup,itemSup,minLength,faultTolerance) obj.mine() patterns = obj.getPatterns() print("Total number of Frequent Patterns:", len(patterns)) obj.save(oFile) Df = obj.getPatternInDataFrame() memUSS = obj.getMemoryUSS() print("Total Memory in USS:", memUSS) memRSS = obj.getMemoryRSS() print("Total Memory in RSS", memRSS) run = obj.getRuntime() print("Total ExecutionTime in seconds:", run) **Credits:** --------------- The complete program was written by P.Likhitha under the supervision of Professor Rage Uday Kiran.\n """ __startTime = float() __endTime = float() _minSup = str() __finalPatterns = {} _iFile = " " _oFile = " " _sep = " " __memoryUSS = float() __memoryRSS = float() __Database = [] __mapSupport = {} __lno = 0 __tree = _Tree() __rank = {} __rankDup = {} def __init__(self, iFile: Union[str, pd.DataFrame], minSup: Union[int, float, str], itemSup: float, minLength: int, faultTolerance: int, sep: str='\t') -> None: super().__init__(iFile, minSup, itemSup, minLength, faultTolerance, sep) def __creatingItemSets(self) -> None: """ Storing the complete transactions of the database/input file in a database variable """ self.__Database = [] if isinstance(self._iFile, _fp._pd.DataFrame): if self._iFile.empty: print("its empty..") i = self._iFile.columns.values.tolist() if 'Transactions' in i: self.__Database = self._iFile['Transactions'].tolist() # print(self.Database) if isinstance(self._iFile, str): if _fp._validators.url(self._iFile): data = _fp._urlopen(self._iFile) for line in data: line.strip() line = line.decode("utf-8") temp = [i.rstrip() for i in line.split(self._sep)] temp = [x for x in temp if x] self.__Database.append(temp) else: try: with open(self._iFile, 'r', encoding='utf-8') as f: for line in f: line.strip() temp = [i.rstrip() for i in line.split(self._sep)] temp = [x for x in temp if x] self.__Database.append(temp) except IOError: print("File Not Found") quit() def __convert(self, value: Union[int, float, str]) -> Union[int, float]: """ To convert the type of user specified minSup value :param value: user specified minSup value :type value: Union[int, float, str] :return: converted type :rtype: int, float """ if type(value) is int: value = int(value) if type(value) is float: value = (len(self.__Database) * value) if type(value) is str: if '.' in value: value = float(value) value = (len(self.__Database) * value) else: value = int(value) return value def __frequentOneItem(self) -> List[str]: """ Generating One frequent items sets :return: one frequency items set :rtype: List """ self.__mapSupport = {} for tr in self.__Database: for i in range(0, len(tr)): if tr[i] not in self.__mapSupport: self.__mapSupport[tr[i]] = 1 else: self.__mapSupport[tr[i]] += 1 self.__mapSupport = {k: v for k, v in self.__mapSupport.items() if v >= self._minSup} genList = [k for k, v in sorted(self.__mapSupport.items(), key=lambda x: x[1], reverse=True)] self.__rank = dict([(index, item) for (item, index) in enumerate(genList)]) return genList def __updateTransactions(self, itemSet: List[str]) -> List[List[int]]: """ Updates the items in transactions with rank of items according to their support :Example: oneLength = {'a':7, 'b': 5, 'c':'4', 'd':3} rank = {'a':0, 'b':1, 'c':2, 'd':3} :param itemSet: list of one-frequent items :type itemSet: List :return: list of updated items in transactions with rank of items according to their support :rtype: List """ list1 = [] for tr in self.__Database: list2 = [] for i in range(len(tr)): if tr[i] in itemSet: list2.append(self.__rank[tr[i]]) if len(list2) >= 1: list2.sort() list1.append(list2) return list1 @staticmethod def __buildTree(transactions: List[List[int]], info: Dict[int, int]) -> _Tree: """ Builds the tree with updated transaction :param transactions: updated transactions :type transactions: List :param info: support details of each item in transactions :type info: Dict :returns: transactions compressed in fp-tree :rtype: Tree """ rootNode = _Tree() rootNode.info = info.copy() for i in range(len(transactions)): rootNode.addTransaction(transactions[i], 1) return rootNode def __savePeriodic(self, itemSet: List[int]) -> str: """ The duplication items and their ranks :param itemSet: frequent itemSet that generated :type itemSet: List :returns: patterns with original item names :rtype: String """ temp = str() for i in itemSet: temp = temp + self.__rankDup[i] + "\t" return temp
[docs] @deprecated("It is recommended to use 'mine()' instead of 'mine()' for mining process. Starting from January 2025, 'mine()' will be completely terminated.") def startMine(self) -> None: """ Main program to start the operation """ self.mine()
[docs] def mine(self) -> None: """ Main program to start the operation """ global _minSup self.__startTime = _fp._time.time() if self._iFile is None: raise Exception("Please enter the file path or file name:") if self._minSup is None: raise Exception("Please enter the Minimum Support") self.__creatingItemSets() self._minSup = self.__convert(self._minSup) _minSup = self._minSup itemSet = self.__frequentOneItem() updatedTransactions = self.__updateTransactions(itemSet) for x, y in self.__rank.items(): self.__rankDup[y] = x info = {self.__rank[k]: v for k, v in self.__mapSupport.items()} __Tree = self.__buildTree(updatedTransactions, info) patterns = __Tree.generatePatterns([]) self.__finalPatterns = {} for k in patterns: s = self.__savePeriodic(k[0]) self.__finalPatterns[str(s)] = k[1] print("Frequent patterns were generated successfully using frequentPatternGrowth algorithm") self.__endTime = _fp._time.time() self.__memoryUSS = float() self.__memoryRSS = float() process = _fp._psutil.Process(_fp._os.getpid()) self.__memoryUSS = process.memory_full_info().uss self.__memoryRSS = process.memory_info().rss
[docs] def getMemoryUSS(self) -> float: """ Total amount of USS memory consumed by the mining process will be retrieved from this function :return: returning USS memory consumed by the mining process :rtype: float """ return self.__memoryUSS
[docs] def getMemoryRSS(self) -> float: """ Total amount of RSS memory consumed by the mining process will be retrieved from this function :return: returning RSS memory consumed by the mining process :rtype: float """ return self.__memoryRSS
[docs] def getRuntime(self) -> float: """ Calculating the total amount of runtime taken by the mining process :return: returning total amount of runtime taken by the mining process :rtype: float """ return self.__endTime - self.__startTime
[docs] def getPatternsAsDataFrame(self) -> pd.DataFrame: """ Storing final frequent patterns in a dataframe :return: returning frequent patterns in a dataframe :rtype: pd.DataFrame """ dataframe = {} data = [] for a, b in self.__finalPatterns.items(): data.append([a.replace('\t', ' '), b]) dataframe = _fp._pd.DataFrame(data, columns=['Patterns', 'Support']) return dataframe
[docs] def save(self, outFile: str) -> None: """ Complete set of frequent patterns will be loaded in to an output file :param outFile: name of the output file :type outFile: csv file :return: None """ self._oFile = outFile writer = open(self._oFile, 'w+') for x, y in self.__finalPatterns.items(): s1 = x.strip() + ":" + str(y) writer.write("%s \n" % s1)
[docs] def getPatterns(self) -> Dict[str, int]: """ Function to send the set of frequent patterns after completion of the mining process :return: returning frequent patterns :rtype: dict """ return self.__finalPatterns
[docs] def printResults(self) -> None: """ This function is used to print the results """ print("Total number of Frequent Patterns:", len(self.getPatterns())) print("Total Memory in USS:", self.getMemoryUSS()) print("Total Memory in RSS", self.getMemoryRSS()) print("Total ExecutionTime in ms:", self.getRuntime())
if __name__ == "__main__": _ap = str() if len(_fp._sys.argv) == 7 or len(_fp._sys.argv) == 8: if len(_fp._sys.argv) == 8: _ap = FTFPGrowth(_fp._sys.argv[1], _fp._sys.argv[3], _fp._sys.argv[4], _fp._sys.argv[5], _fp._sys.argv[6], _fp._sys.argv[7]) if len(_fp._sys.argv) == 7: _ap = FTFPGrowth(_fp._sys.argv[1], _fp._sys.argv[3], _fp._sys.argv[4], _fp._sys.argv[5], _fp._sys.argv[6]) _ap.mine() _ap.mine() print("Total number of Frequent Patterns:", len(_ap.getPatterns())) _ap.save(_fp._sys.argv[2]) print("Total Memory in USS:", _ap.getMemoryUSS()) print("Total Memory in RSS", _ap.getMemoryRSS()) print("Total ExecutionTime in ms:", _ap.getRuntime()) else: print("Error! The number of input parameters do not match the total number of parameters provided")