text-summarizer-aweebtaku 1.0.2__tar.gz → 1.2.0__tar.gz

{text_summarizer_aweebtaku-1.0.2/text_summarizer_aweebtaku.egg-info → text_summarizer_aweebtaku-1.2.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: text-summarizer-aweebtaku
-Version: 1.0.2
+Version: 1.2.0
 Summary: A text summarization tool using GloVe embeddings and PageRank algorithm
 Home-page: https://github.com/AWeebTaku/Summarizer
 Author: Your Name
@@ -22,6 +22,7 @@ Requires-Dist: numpy
 Requires-Dist: nltk
 Requires-Dist: scikit-learn
 Requires-Dist: networkx
+Requires-Dist: requests
 Dynamic: author
 Dynamic: author-email
 Dynamic: classifier
@@ -60,8 +61,6 @@ A Python-based text summarization tool that uses GloVe word embeddings and PageR
 pip install text-summarizer-aweebtaku
 ```
 
-**Note:** This package includes the GloVe word embeddings file (~400MB), so the installation may take some time.
-
 ### Install from Source
 
 1. Clone the repository:
@@ -75,7 +74,14 @@ cd Summarizer
 pip install -e .
 ```
 
-**Note:** The GloVe word embeddings are included in the package, so no additional download is required.
+### Download GloVe Embeddings
+
+**No manual download required!** The package will automatically download GloVe embeddings (100d, ~400MB) on first use and cache them in your home directory (`~/.text_summarizer/`).
+
+If you prefer to use your own GloVe file, you can specify the path:
+```python
+summarizer = TextSummarizer(glove_path='path/to/your/glove.6B.100d.txt')
+```
 
 ## Usage
 
@@ -103,20 +109,20 @@ text-summarizer-gui
 
 ```python
 from text_summarizer import TextSummarizer
-import pandas as pd
 
-# Initialize summarizer
-summarizer = TextSummarizer(glove_path='glove.6B.100d.txt')
+# Initialize summarizer (automatic GloVe download)
+summarizer = TextSummarizer(num_sentences=3)
 
-# Load data
-df = pd.DataFrame([{'article_id': 1, 'article_text': 'Your text here...'}])
+# Simple text summarization
+text = "Your long text here..."
+summary = summarizer.summarize_text(text)
+print(summary)
 
-# Run summarization
+# Advanced usage with DataFrame
+import pandas as pd
+df = pd.DataFrame([{'article_id': 1, 'article_text': text}])
 scored_sentences = summarizer.run_summarization(df)
-
-# Get summary for article ID 1
 article_text, summary = summarizer.summarize_article(scored_sentences, 1, df)
-print(summary)
 ```
 
 ## Data Format

{text_summarizer_aweebtaku-1.0.2 → text_summarizer_aweebtaku-1.2.0}/README.md

@@ -24,8 +24,6 @@ A Python-based text summarization tool that uses GloVe word embeddings and PageR
 pip install text-summarizer-aweebtaku
 ```
 
-**Note:** This package includes the GloVe word embeddings file (~400MB), so the installation may take some time.
-
 ### Install from Source
 
 1. Clone the repository:
@@ -39,7 +37,14 @@ cd Summarizer
 pip install -e .
 ```
 
-**Note:** The GloVe word embeddings are included in the package, so no additional download is required.
+### Download GloVe Embeddings
+
+**No manual download required!** The package will automatically download GloVe embeddings (100d, ~400MB) on first use and cache them in your home directory (`~/.text_summarizer/`).
+
+If you prefer to use your own GloVe file, you can specify the path:
+```python
+summarizer = TextSummarizer(glove_path='path/to/your/glove.6B.100d.txt')
+```
 
 ## Usage
 
@@ -67,20 +72,20 @@ text-summarizer-gui
 
 ```python
 from text_summarizer import TextSummarizer
-import pandas as pd
 
-# Initialize summarizer
-summarizer = TextSummarizer(glove_path='glove.6B.100d.txt')
+# Initialize summarizer (automatic GloVe download)
+summarizer = TextSummarizer(num_sentences=3)
 
-# Load data
-df = pd.DataFrame([{'article_id': 1, 'article_text': 'Your text here...'}])
+# Simple text summarization
+text = "Your long text here..."
+summary = summarizer.summarize_text(text)
+print(summary)
 
-# Run summarization
+# Advanced usage with DataFrame
+import pandas as pd
+df = pd.DataFrame([{'article_id': 1, 'article_text': text}])
 scored_sentences = summarizer.run_summarization(df)
-
-# Get summary for article ID 1
 article_text, summary = summarizer.summarize_article(scored_sentences, 1, df)
-print(summary)
 ```
 
 ## Data Format

{text_summarizer_aweebtaku-1.0.2 → text_summarizer_aweebtaku-1.2.0}/requirements.txt

@@ -2,4 +2,5 @@ pandas
 numpy
 nltk
 scikit-learn
-networkx
+networkx
+requests

{text_summarizer_aweebtaku-1.0.2 → text_summarizer_aweebtaku-1.2.0}/setup.py

@@ -8,7 +8,7 @@ with open("requirements.txt", "r", encoding="utf-8") as fh:
 
 setup(
     name="text-summarizer-aweebtaku",
-    version="1.0.2",
+    version="1.2.0",
     author="Your Name",
     author_email="your.email@example.com",
     description="A text summarization tool using GloVe embeddings and PageRank algorithm",
@@ -37,6 +37,6 @@ setup(
     },
     include_package_data=True,
     package_data={
-        "text_summarizer": ["data/*.csv"],
+        "textsummarizer": ["data/*.csv"],
     },
 )

text_summarizer_aweebtaku-1.2.0/text_summarizer/summarizer.py

@@ -0,0 +1,323 @@
+import pandas as pd
+import numpy as np
+import nltk
+import os
+import zipfile
+from pathlib import Path
+from typing import Dict, List, Tuple, Optional, Union
+from nltk.tokenize import sent_tokenize
+from nltk.corpus import stopwords
+from sklearn.metrics.pairwise import cosine_similarity
+import networkx as nx
+
+# Download necessary NLTK data
+# nltk.download('punkt_tab')
+# nltk.download('stopwords')
+
+class TextSummarizer:
+    """A class for summarizing text documents using GloVe embeddings and PageRank."""
+
+    def __init__(self, glove_path: Optional[str] = None, num_sentences: int = 5):
+        self.num_sentences = num_sentences
+        self.word_embeddings: Dict[str, np.ndarray] = {}
+        self.stop_words: set = set(stopwords.words('english'))
+
+        # Set default GloVe path
+        if glove_path is None:
+            glove_path = self._get_default_glove_path()
+
+        self.glove_path = glove_path
+        self._load_embeddings()
+
+    def _get_default_glove_path(self):
+        """Get the default path for GloVe embeddings."""
+        # Use user's home directory for data
+        home_dir = Path.home()
+        glove_dir = home_dir / '.text_summarizer'
+        glove_dir.mkdir(exist_ok=True)
+        return glove_dir / 'glove.6B.100d.txt'
+
+    def _download_glove_embeddings(self):
+        """Download GloVe embeddings if not present with improved error handling."""
+        import requests
+
+        print("GloVe embeddings not found. Downloading from Stanford NLP...")
+
+        # Create directory if it doesn't exist
+        glove_file = Path(self.glove_path)
+        glove_file.parent.mkdir(exist_ok=True)
+
+        # Download the zip file
+        url = "https://nlp.stanford.edu/data/glove.6B.zip"
+        zip_path = glove_file.parent / "glove.6B.zip"
+
+        headers = {
+            'User-Agent': 'TextSummarizer/1.1.0 (https://github.com/AWeebTaku/Summarizer)',
+        }
+
+        try:
+            print("Downloading GloVe embeddings (862 MB)...")
+            with requests.get(url, headers=headers, stream=True, timeout=30) as response:
+                response.raise_for_status()
+
+                total_size = int(response.headers.get('content-length', 0))
+                downloaded_size = 0
+
+                with open(zip_path, 'wb') as f:
+                    for chunk in response.iter_content(chunk_size=8192):
+                        if chunk:
+                            f.write(chunk)
+                            downloaded_size += len(chunk)
+                            if total_size > 0:
+                                progress = (downloaded_size / total_size) * 100
+                                print(".1f", end='', flush=True)
+
+            print("\nDownload complete. Extracting...")
+
+            # Extract the specific file we need
+            with zipfile.ZipFile(zip_path, 'r') as zip_ref:
+                zip_ref.extract('glove.6B.100d.txt', glove_file.parent)
+
+            # Verify extraction
+            if not glove_file.exists():
+                raise FileNotFoundError("Failed to extract GloVe file from zip")
+
+            # Clean up zip file
+            zip_path.unlink()
+
+            print(f"GloVe embeddings extracted to {self.glove_path}")
+
+        except requests.exceptions.RequestException as e:
+            print(f"Network error during download: {e}")
+            raise Exception(f"Failed to download GloVe embeddings: {e}")
+        except zipfile.BadZipFile as e:
+            print(f"Invalid zip file downloaded: {e}")
+            if zip_path.exists():
+                zip_path.unlink()
+            raise Exception("Downloaded file is not a valid zip archive")
+        except Exception as e:
+            print(f"Unexpected error during download: {e}")
+            if zip_path.exists():
+                zip_path.unlink()
+            raise
+
+    def _load_embeddings(self):
+        """Load GloVe word embeddings from file with optimized memory usage."""
+        if not os.path.exists(self.glove_path):
+            self._download_glove_embeddings()
+
+        try:
+            print(f"Loading GloVe embeddings from {self.glove_path}...")
+            word_count = 0
+
+            with open(self.glove_path, 'r', encoding='utf-8', errors='ignore') as f:
+                for line_num, line in enumerate(f, 1):
+                    line = line.strip()
+                    if not line:
+                        continue
+
+                    try:
+                        values = line.split()
+                        if len(values) < 101:  # word + 100 dimensions
+                            continue
+
+                        word = values[0]
+                        coefs = np.asarray(values[1:101], dtype='float32')  # Only take first 100 dims
+                        self.word_embeddings[word] = coefs
+                        word_count += 1
+
+                        # Progress update every 50k words
+                        if word_count % 50000 == 0:
+                            print(f"Loaded {word_count} words...")
+
+                    except (ValueError, IndexError) as e:
+                        # Skip malformed lines
+                        continue
+
+            print(f"Successfully loaded {len(self.word_embeddings)} word embeddings.")
+
+            if len(self.word_embeddings) == 0:
+                raise ValueError("No valid embeddings found in GloVe file")
+
+        except FileNotFoundError:
+            raise FileNotFoundError(f"GloVe file not found at {self.glove_path}")
+        except Exception as e:
+            raise Exception(f"Error loading GloVe embeddings: {e}")
+
+    def preprocess_sentences(self, df: pd.DataFrame) -> List[Dict]:
+        """Tokenize articles into sentences and store metadata."""
+        all_sentences_data = []
+        sentence_counter_global = 0
+        for _, article_row in df.iterrows():
+            article_id = article_row['article_id']
+            article_text = article_row['article_text']
+            article_sentences = sent_tokenize(article_text)
+            for sent_idx, sentence_text in enumerate(article_sentences):
+                all_sentences_data.append({
+                    'global_sentence_idx': sentence_counter_global,
+                    'article_id': article_id,
+                    'sentence_text': sentence_text,
+                    'original_article_sentence_idx': sent_idx
+                })
+                sentence_counter_global += 1
+        return all_sentences_data
+
+    def clean_sentences(self, sentences):
+        """Clean sentences: remove non-alphabetic, lowercase, remove stopwords."""
+        if not sentences:
+            return []
+
+        # Use pandas for efficient string operations
+        clean_sentences = pd.Series(sentences).str.replace(r"[^a-zA-Z\s]", " ", regex=True)
+        clean_sentences = clean_sentences.str.lower()
+        clean_sentences = clean_sentences.apply(self._remove_stopwords)
+        return clean_sentences.tolist()
+
+    def _remove_stopwords(self, sentence):
+        """Remove stopwords from a sentence string."""
+        if not isinstance(sentence, str):
+            return ""
+        words = sentence.split()
+        filtered_words = [word for word in words if word not in self.stop_words]
+        return " ".join(filtered_words)
+
+    def compute_sentence_vectors(self, clean_sentences):
+        """Compute sentence vectors using GloVe embeddings with vectorized operations."""
+        if not clean_sentences:
+            return []
+
+        sentence_vectors = []
+        for sentence in clean_sentences:
+            words = sentence.split()
+            if words:
+                # Get embeddings for all words in sentence
+                vectors = []
+                for word in words:
+                    embedding = self.word_embeddings.get(word, np.zeros(100, dtype=np.float32))
+                    vectors.append(embedding)
+
+                if vectors:
+                    # Use mean of word vectors
+                    v = np.mean(vectors, axis=0)
+                else:
+                    v = np.zeros(100, dtype=np.float32)
+            else:
+                v = np.zeros(100, dtype=np.float32)
+            sentence_vectors.append(v)
+
+        return sentence_vectors
+
+    def compute_similarity_matrix(self, sentence_vectors):
+        """Compute cosine similarity matrix using vectorized operations."""
+        if not sentence_vectors:
+            return np.array([])
+
+        # Convert to numpy array for vectorized operations
+        vectors = np.array(sentence_vectors)
+        n = len(vectors)
+
+        # Normalize vectors for faster cosine similarity
+        norms = np.linalg.norm(vectors, axis=1, keepdims=True)
+        norms[norms == 0] = 1  # Avoid division by zero
+        normalized_vectors = vectors / norms
+
+        # Compute cosine similarity matrix using matrix multiplication
+        sim_mat = np.dot(normalized_vectors, normalized_vectors.T)
+
+        # Ensure diagonal is zero (no self-similarity)
+        np.fill_diagonal(sim_mat, 0)
+
+        return sim_mat
+
+    def rank_sentences(self, sim_mat):
+        """Rank sentences using PageRank with optimized parameters."""
+        if sim_mat.size == 0:
+            return {}
+
+        try:
+            # Create graph from similarity matrix
+            nx_graph = nx.from_numpy_array(sim_mat)
+
+            # Use optimized PageRank parameters
+            scores = nx.pagerank(
+                nx_graph,
+                alpha=0.85,  # Damping factor
+                max_iter=100,
+                tol=1e-6
+            )
+
+            return scores
+        except Exception as e:
+            print(f"Warning: PageRank failed, using uniform scores: {e}")
+            # Fallback: return uniform scores
+            n = sim_mat.shape[0]
+            return {i: 1.0/n for i in range(n)}
+
+    def summarize_article(self, scored_sentences, article_id, df):
+        """Generate summary for a specific article."""
+        article_sentences = [s for s in scored_sentences if s['article_id'] == article_id]
+        if not article_sentences:
+            return None, None
+
+        article_sentences.sort(key=lambda x: x['score'], reverse=True)
+        top_sentences = article_sentences[:self.num_sentences]
+        top_sentences.sort(key=lambda x: x['original_article_sentence_idx'])
+        summary = " ".join([s['sentence_text'] for s in top_sentences])
+
+        article_row = df[df['article_id'] == article_id]
+        if not article_row.empty:
+            article_text = article_row['article_text'].iloc[0]
+            return article_text, summary
+        return None, None
+
+    def summarize_all_articles(self, scored_sentences, df):
+        """Generate summaries for all articles."""
+        summaries = {}
+        for _, article_row in df.iterrows():
+            article_id = article_row['article_id']
+            article_text, summary = self.summarize_article(scored_sentences, article_id, df)
+            if article_text and summary:
+                summaries[article_id] = {'article': article_text, 'summary': summary}
+        return summaries
+
+    def run_summarization(self, df):
+        """Run the full summarization pipeline."""
+        sentences_data = self.preprocess_sentences(df)
+        sentences = [s['sentence_text'] for s in sentences_data]
+        clean_sentences = self.clean_sentences(sentences)
+        sentence_vectors = self.compute_sentence_vectors(clean_sentences)
+        sim_mat = self.compute_similarity_matrix(sentence_vectors)
+        scores = self.rank_sentences(sim_mat)
+
+        for i, sentence_data in enumerate(sentences_data):
+            sentence_data['score'] = scores[i]
+
+        return sentences_data
+
+    def summarize_text(self, text: str, num_sentences: Optional[int] = None) -> str:
+        """
+        Summarize a single text document.
+
+        Args:
+            text (str): The text to summarize
+            num_sentences (int, optional): Number of sentences in summary. Defaults to self.num_sentences.
+
+        Returns:
+            str: The summarized text
+        """
+        if not text or not text.strip():
+            return ""
+
+        if num_sentences is None:
+            num_sentences = self.num_sentences
+
+        # Create a temporary DataFrame
+        df = pd.DataFrame([{'article_id': 1, 'article_text': text}])
+
+        # Run summarization pipeline
+        scored_sentences = self.run_summarization(df)
+
+        # Get summary
+        _, summary = self.summarize_article(scored_sentences, 1, df)
+
+        return summary if summary else text

{text_summarizer_aweebtaku-1.0.2 → text_summarizer_aweebtaku-1.2.0/text_summarizer_aweebtaku.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: text-summarizer-aweebtaku
-Version: 1.0.2
+Version: 1.2.0
 Summary: A text summarization tool using GloVe embeddings and PageRank algorithm
 Home-page: https://github.com/AWeebTaku/Summarizer
 Author: Your Name
@@ -22,6 +22,7 @@ Requires-Dist: numpy
 Requires-Dist: nltk
 Requires-Dist: scikit-learn
 Requires-Dist: networkx
+Requires-Dist: requests
 Dynamic: author
 Dynamic: author-email
 Dynamic: classifier
@@ -60,8 +61,6 @@ A Python-based text summarization tool that uses GloVe word embeddings and PageR
 pip install text-summarizer-aweebtaku
 ```
 
-**Note:** This package includes the GloVe word embeddings file (~400MB), so the installation may take some time.
-
 ### Install from Source
 
 1. Clone the repository:
@@ -75,7 +74,14 @@ cd Summarizer
 pip install -e .
 ```
 
-**Note:** The GloVe word embeddings are included in the package, so no additional download is required.
+### Download GloVe Embeddings
+
+**No manual download required!** The package will automatically download GloVe embeddings (100d, ~400MB) on first use and cache them in your home directory (`~/.text_summarizer/`).
+
+If you prefer to use your own GloVe file, you can specify the path:
+```python
+summarizer = TextSummarizer(glove_path='path/to/your/glove.6B.100d.txt')
+```
 
 ## Usage
 
@@ -103,20 +109,20 @@ text-summarizer-gui
 
 ```python
 from text_summarizer import TextSummarizer
-import pandas as pd
 
-# Initialize summarizer
-summarizer = TextSummarizer(glove_path='glove.6B.100d.txt')
+# Initialize summarizer (automatic GloVe download)
+summarizer = TextSummarizer(num_sentences=3)
 
-# Load data
-df = pd.DataFrame([{'article_id': 1, 'article_text': 'Your text here...'}])
+# Simple text summarization
+text = "Your long text here..."
+summary = summarizer.summarize_text(text)
+print(summary)
 
-# Run summarization
+# Advanced usage with DataFrame
+import pandas as pd
+df = pd.DataFrame([{'article_id': 1, 'article_text': text}])
 scored_sentences = summarizer.run_summarization(df)
-
-# Get summary for article ID 1
 article_text, summary = summarizer.summarize_article(scored_sentences, 1, df)
-print(summary)
 ```
 
 ## Data Format

{text_summarizer_aweebtaku-1.0.2 → text_summarizer_aweebtaku-1.2.0}/text_summarizer_aweebtaku.egg-info/requires.txt

@@ -3,3 +3,4 @@ numpy
 nltk
 scikit-learn
 networkx
+requests

text_summarizer_aweebtaku-1.0.2/text_summarizer/summarizer.py

@@ -1,251 +0,0 @@
-import pandas as pd
-import numpy as np
-import nltk
-import os
-from nltk.tokenize import sent_tokenize
-from nltk.corpus import stopwords
-from sklearn.metrics.pairwise import cosine_similarity
-import networkx as nx
-import pkg_resources
-import urllib.request
-import zipfile
-import shutil
-
-# Download necessary NLTK data
-# nltk.download('punkt_tab')
-# nltk.download('stopwords')
-
-class TextSummarizer:
-    """A class for summarizing text documents using GloVe embeddings and PageRank."""
-
-    def __init__(self, glove_path=None, num_sentences=5):
-        if glove_path is None:
-            # Try to find GloVe file in package data first
-            try:
-                glove_path = pkg_resources.resource_filename('text_summarizer', 'glove.6B.100d.txt/glove.6B.100d.txt')
-            except (FileNotFoundError, ModuleNotFoundError):
-                # Fallback to default path
-                glove_path = 'glove.6B.100d.txt/glove.6B.100d.txt'
-        
-        # Download GloVe if it doesn't exist
-        if not os.path.exists(glove_path):
-            print("GloVe embeddings not found. Downloading...")
-            self._download_glove()
-        
-        self.glove_path = glove_path
-        self.num_sentences = num_sentences
-        self.word_embeddings = {}
-        self.stop_words = set(stopwords.words('english'))
-        self._load_embeddings()
-
-    def _download_glove(self):
-        """Download and extract GloVe embeddings."""
-        glove_dir = 'glove.6B.100d.txt'
-        glove_file = os.path.join(glove_dir, 'glove.6B.100d.txt')
-        zip_url = 'http://nlp.stanford.edu/data/glove.6B.zip'
-        zip_path = 'glove.6B.zip'
-        
-        try:
-            # Download the zip file
-            print(f"Downloading GloVe from {zip_url}...")
-            urllib.request.urlretrieve(zip_url, zip_path)
-            
-            # Extract the specific file we need
-            print("Extracting GloVe embeddings...")
-            with zipfile.ZipFile(zip_path, 'r') as zip_ref:
-                # Create directory if it doesn't exist
-                os.makedirs(glove_dir, exist_ok=True)
-                # Extract only the 100d file
-                zip_ref.extract('glove.6B.100d.txt', '.')
-                # Move to the expected directory
-                if os.path.exists('glove.6B.100d.txt'):
-                    shutil.move('glove.6B.100d.txt', glove_file)
-            
-            # Clean up
-            os.remove(zip_path)
-            print("GloVe embeddings downloaded successfully!")
-            
-        except Exception as e:
-            print(f"Failed to download GloVe embeddings: {e}")
-            print("Please download manually from: http://nlp.stanford.edu/data/glove.6B.zip")
-            raise
-
-    def _load_embeddings(self):
-        """Load GloVe word embeddings from file."""
-        try:
-            with open(self.glove_path, 'r', encoding='utf-8') as f:
-                for line in f:
-                    values = line.split()
-                    word = values[0]
-                    coefs = np.asarray(values[1:], dtype='float32')
-                    self.word_embeddings[word] = coefs
-        except FileNotFoundError:
-            raise FileNotFoundError(f"GloVe file not found at {self.glove_path}")
-
-    def load_data(self):
-        """Load data interactively."""
-        while True:
-            choice = input("Enter 'P' to paste a single article,\n'U' to upload a CSV with multiple articles,\n'C' to create a new CSV with multiple articles: ").upper()
-            df = pd.DataFrame()
-            save_csv = True
-
-            if choice == 'P':
-                article_text = input("Paste your article text here:\n")
-                df = pd.DataFrame([{'article_id': 1, 'article_text': article_text}])
-                print('DataFrame created from single article.')
-                save_csv = False
-                break
-            elif choice == 'U':
-                print("You chose to load an existing CSV file. It should contain 'article_id' and 'article_text' columns.")
-                save_csv = False
-                while True:
-                    file_name = input("Enter the name of the CSV file (e.g., 'tennis.csv') or type 'cancel' to go back: ").strip()
-                    if file_name.lower() == 'cancel':
-                        break
-                    if os.path.exists(file_name) and file_name.lower().endswith('.csv'):
-                        try:
-                            df = pd.read_csv(file_name)
-                            print(f'CSV file "{file_name}" loaded successfully.')
-                            break
-                        except Exception as e:
-                            print(f"Error reading file '{file_name}': {e}")
-                    else:
-                        print(f"File '{file_name}' not found or is not a CSV. Please try again.")
-                if not df.empty:
-                    break
-            elif choice == 'C':
-                print("You've chosen to create a CSV with multiple articles. Enter 'done' for article ID when finished.")
-                articles_data = []
-                article_counter = 1
-                while True:
-                    article_id_input = input(f"Enter article ID for article {article_counter} (or 'done' to finish): ").strip()
-                    if article_id_input.lower() == 'done':
-                        break
-                    try:
-                        article_id = int(article_id_input)
-                    except ValueError:
-                        print("Invalid Article ID. Please enter a number or 'done'.")
-                        continue
-                    article_text = input("Enter article text:\n").strip()
-                    if not article_text:
-                        print("Article text cannot be empty. Please try again.")
-                        continue
-                    articles_data.append({'article_id': article_id, 'article_text': article_text})
-                    article_counter += 1
-                if articles_data:
-                    df = pd.DataFrame(articles_data)
-                    print('DataFrame created from multiple articles.')
-                    break
-                else:
-                    print("No articles were entered. Please try again or choose another option.")
-            else:
-                print("Invalid choice. Please enter 'P', 'U', or 'C'.")
-
-        if not df.empty and save_csv:
-            df.to_csv('article.csv', index=False)
-            print('CSV file "article.csv" created/updated successfully.')
-        elif df.empty:
-            print("No DataFrame was created.")
-        return df
-
-    def preprocess_sentences(self, df):
-        """Tokenize articles into sentences and store metadata."""
-        all_sentences_data = []
-        sentence_counter_global = 0
-        for _, article_row in df.iterrows():
-            article_id = article_row['article_id']
-            article_text = article_row['article_text']
-            article_sentences = sent_tokenize(article_text)
-            for sent_idx, sentence_text in enumerate(article_sentences):
-                all_sentences_data.append({
-                    'global_sentence_idx': sentence_counter_global,
-                    'article_id': article_id,
-                    'sentence_text': sentence_text,
-                    'original_article_sentence_idx': sent_idx
-                })
-                sentence_counter_global += 1
-        return all_sentences_data
-
-    def clean_sentences(self, sentences):
-        """Clean sentences: remove non-alphabetic, lowercase, remove stopwords."""
-        clean_sentences = pd.Series(sentences).str.replace(r"[^a-zA-Z\s]", " ", regex=True)
-        clean_sentences = clean_sentences.str.lower()
-        clean_sentences = clean_sentences.apply(lambda s: self._remove_stopwords(s.split()))
-        return clean_sentences.tolist()
-
-    def _remove_stopwords(self, sen):
-        """Remove stopwords from a list of words."""
-        return " ".join([word for word in sen if word not in self.stop_words])
-
-    def compute_sentence_vectors(self, clean_sentences):
-        """Compute sentence vectors using GloVe embeddings."""
-        sentence_vectors = []
-        for sentence in clean_sentences:
-            words = sentence.split()
-            if words:
-                vectors = [self.word_embeddings.get(w, np.zeros(100)) for w in words]
-                v = np.mean(vectors, axis=0)
-            else:
-                v = np.zeros(100)
-            sentence_vectors.append(v)
-        return sentence_vectors
-
-    def compute_similarity_matrix(self, sentence_vectors):
-        """Compute cosine similarity matrix."""
-        n = len(sentence_vectors)
-        sim_mat = np.zeros((n, n))
-        for i in range(n):
-            for j in range(n):
-                if i != j:
-                    sim_mat[i][j] = cosine_similarity(
-                        sentence_vectors[i].reshape(1, -1),
-                        sentence_vectors[j].reshape(1, -1)
-                    )[0, 0]
-        return sim_mat
-
-    def rank_sentences(self, sim_mat):
-        """Rank sentences using PageRank."""
-        nx_graph = nx.from_numpy_array(sim_mat)
-        scores = nx.pagerank(nx_graph)
-        return scores
-
-    def summarize_article(self, scored_sentences, article_id, df):
-        """Generate summary for a specific article."""
-        article_sentences = [s for s in scored_sentences if s['article_id'] == article_id]
-        if not article_sentences:
-            return None, None
-
-        article_sentences.sort(key=lambda x: x['score'], reverse=True)
-        top_sentences = article_sentences[:self.num_sentences]
-        top_sentences.sort(key=lambda x: x['original_article_sentence_idx'])
-        summary = " ".join([s['sentence_text'] for s in top_sentences])
-
-        article_row = df[df['article_id'] == article_id]
-        if not article_row.empty:
-            article_text = article_row['article_text'].iloc[0]
-            return article_text, summary
-        return None, None
-
-    def summarize_all_articles(self, scored_sentences, df):
-        """Generate summaries for all articles."""
-        summaries = {}
-        for _, article_row in df.iterrows():
-            article_id = article_row['article_id']
-            article_text, summary = self.summarize_article(scored_sentences, article_id, df)
-            if article_text and summary:
-                summaries[article_id] = {'article': article_text, 'summary': summary}
-        return summaries
-
-    def run_summarization(self, df):
-        """Run the full summarization pipeline."""
-        sentences_data = self.preprocess_sentences(df)
-        sentences = [s['sentence_text'] for s in sentences_data]
-        clean_sentences = self.clean_sentences(sentences)
-        sentence_vectors = self.compute_sentence_vectors(clean_sentences)
-        sim_mat = self.compute_similarity_matrix(sentence_vectors)
-        scores = self.rank_sentences(sim_mat)
-
-        for i, sentence_data in enumerate(sentences_data):
-            sentence_data['score'] = scores[i]
-
-        return sentences_data