Create app.py

app.py

@@ -0,0 +1,92 @@
+import streamlit as st
+import networkx as nx
+import matplotlib.pyplot as plt
+from nltk import sent_tokenize
+from sklearn.feature_extraction.text import TfidfVectorizer
+from sklearn.cluster import KMeans
+import numpy as np
+
+# Helper function to split text into topics using KMeans clustering
+def split_text_into_topics(text, n_topics):
+    sentences = sent_tokenize(text)
+    vectorizer = TfidfVectorizer(stop_words='english')
+    X = vectorizer.fit_transform(sentences)
+    
+    kmeans = KMeans(n_clusters=n_topics, random_state=42)
+    kmeans.fit(X)
+    
+    clusters = kmeans.labels_.tolist()
+    topic_sentences = {i: [] for i in range(n_topics)}
+    
+    for i, sentence in enumerate(sentences):
+        topic_sentences[clusters[i]].append(sentence)
+        
+    return topic_sentences
+
+# Recursive function to split subtopics
+def recursive_split(topic_dict, depth, max_depth, subtopics):
+    if depth >= max_depth:
+        return
+    
+    new_topic_dict = {}
+    for topic, sentences in topic_dict.items():
+        if len(sentences) <= 1:
+            new_topic_dict[topic] = sentences
+        else:
+            sub_topics = split_text_into_topics(' '.join(sentences), subtopics)
+            new_topic_dict[topic] = sub_topics
+            
+    return new_topic_dict
+
+# Plotting function to visualize the tree structure
+def plot_tree(tree, parent=None, graph=None, level=0):
+    if graph is None:
+        graph = nx.Graph()
+    
+    for key, value in tree.items():
+        node_label = f'Topic {key}' if parent is None else f'Subtopic {key}'
+        graph.add_node(node_label, level=level)
+        if parent:
+            graph.add_edge(parent, node_label)
+        
+        if isinstance(value, dict):
+            plot_tree(value, parent=node_label, graph=graph, level=level+1)
+        else:
+            for i, sentence in enumerate(value):
+                sentence_label = f"{node_label} - Sentence {i+1}"
+                graph.add_node(sentence_label, level=level+1)
+                graph.add_edge(node_label, sentence_label)
+    
+    return graph
+
+# Streamlit App layout
+st.title('Text Topic Tree Generator')
+
+# Upload file
+uploaded_file = st.file_uploader("Upload a text file", type="txt")
+
+if uploaded_file is not None:
+    text = uploaded_file.read().decode('utf-8')
+    
+    # Select number of main topics and depth of subtopics
+    n_topics = st.slider('Select number of main topics', 2, 10, 5)
+    max_depth = st.slider('Select maximum depth of subtopics', 1, 5, 2)
+    subtopics_per_topic = st.slider('Select number of subtopics per topic', 2, 5, 3)
+    
+    # Split text into main topics
+    topic_dict = split_text_into_topics(text, n_topics)
+    
+    # Recursively split the topics into subtopics
+    full_tree = recursive_split(topic_dict, 0, max_depth, subtopics_per_topic)
+    
+    # Create and display the tree graph
+    graph = plot_tree(full_tree)
+    
+    # Plot the tree graph
+    pos = nx.spring_layout(graph)
+    levels = nx.get_node_attributes(graph, 'level')
+    plt.figure(figsize=(12, 8))
+    nx.draw(graph, pos, with_labels=True, node_size=3000, node_color="lightblue", font_size=10, font_weight="bold", labels={node: node for node in graph.nodes()})
+    plt.title("Tree Structure of Text Topics")
+    st.pyplot(plt)
+