Create app.py

app.py

@@ -0,0 +1,130 @@
+import matplotlib.pyplot as plt
+from wordcloud import WordCloud
+import nltk
+from collections import Counter
+from googletrans import Translator
+from nltk.corpus import stopwords
+import gradio as gr
+
+# Download necessary NLTK data
+nltk.download('punkt')
+nltk.download('averaged_perceptron_tagger')
+nltk.download('stopwords')
+
+translator = Translator()
+stop_words = set(stopwords.words('english'))
+
+# Define example sentences and synonyms for the word list
+word_data_examples = {
+    "village": ("The village was quiet at night.", "hamlet, community"),
+    "adventure": ("They went on an exciting adventure in the forest.", "expedition, quest"),
+    "map": ("We used a map to find the hidden treasure.", "chart, atlas"),
+    "cave": ("They explored a dark cave in the mountains.", "cavern, grotto"),
+    "among": ("She found her book among the pile of papers.", "amidst, between"),
+    "mountains": ("The mountains were covered with snow in winter.", "peaks, ranges"),
+    "children": ("The children played games in the park.", "kids, youngsters"),
+    "known": ("He was known for his kindness and bravery.", "recognized, famous"),
+    "hidden": ("They found a hidden door behind the bookshelf.", "concealed, secret"),
+    "local": ("The local market was full of fresh produce.", "regional, native"),
+    "discovery": ("The discovery of the old map excited everyone.", "finding, revelation"),
+    "eagle": ("An eagle soared high above the valley.", "raptor, bird of prey"),
+    "villagers": ("The villagers gathered in the square for the festival.", "residents, townsfolk"),
+    "legend": ("The legend of the lost city intrigued the adventurers.", "myth, lore"),
+    "tales": ("Grandma told us tales of her childhood.", "stories, narratives"),
+    "daring": ("His daring escape from the cave was legendary.", "bold, audacious"),
+    "spirit": ("The spirit of adventure was alive in their hearts.", "soul, essence"),
+    "exploring": ("They spent the summer exploring the forest.", "investigating, discovering"),
+    "old": ("The old castle was full of secrets.", "ancient, aged"),
+    "lost": ("He felt lost without his best friend.", "missing, misplaced"),
+    "ancient": ("They discovered ancient artifacts in the desert.", "archaic, antique"),
+    "inside": ("Inside the box was a beautiful necklace.", "within, interior"),
+    "treasure": ("They dreamed of finding hidden treasure.", "riches, valuables"),
+    "whispering": ("The trees were whispering secrets in the wind.", "murmuring, softly speaking"),
+    "hollow": ("They found a hollow tree to hide in during the storm.", "cavity, void"),
+    "decided": ("She decided to take the long way home.", "determined, resolved"),
+    "journey": ("Their journey took them across the country.", "trip, voyage"),
+    "together": ("They worked together to solve the mystery.", "jointly, collectively"),
+    "way": ("She found a new way to solve the puzzle.", "method, manner"),
+    "reached": ("They finally reached the top of the hill.", "arrived, attained"),
+    "chest": ("The chest was filled with gold coins.", "trunk, box"),
+    "boulder": ("A large boulder blocked the path.", "rock, stone"),
+    "artifacts": ("The museum displayed artifacts from ancient Egypt.", "relics, antiquities"),
+    "legends": ("The legends spoke of a hidden kingdom.", "myths, sagas"),
+    "explore": ("They wanted to explore the old mansion.", "investigate, examine"),
+    "secret": ("She kept the secret hidden from everyone.", "confidential, hidden"),
+    "small": ("The small kitten was very playful.", "tiny, little"),
+    "mountain": ("The mountain was covered in thick forests.", "peak, hill"),
+    "part": ("Each part of the puzzle was important.", "piece, segment"),
+    "everyday": ("He wore his everyday clothes to the party.", "daily, routine"),
+    "life": ("Life in the village was peaceful.", "existence, being"),
+    "nestled": ("The cabin was nestled in the woods.", "tucked, situated"),
+    "towering": ("The towering trees made the forest dark and cool.", "lofty, soaring"),
+    "peaks": ("The mountain peaks were covered in snow.", "summits, crests"),
+    "said": ("He said he would be back soon.", "stated, remarked"),
+    "protected": ("The ancient ruins were protected by law.", "guarded, sheltered"),
+    "massive": ("The massive ship docked at the port.", "enormous, huge"),
+    "supposedly": ("The treasure was supposedly buried under the tree.", "allegedly, reportedly"),
+    "watched": ("They watched the movie together.", "observed, viewed"),
+    "perch": ("The bird found a perch on the windowsill.", "roost, rest")
+}
+
+# Words to be excluded from both the word cloud and the word list
+exclude_words = set([
+    'i', 'you', 'he', 'she', 'it', 'we', 'they', 'me', 'him', 'her', 'us', 'them', 
+    'my', 'your', 'his', 'its', 'our', 'their', 'mine', 'yours', 'hers', 'ours', 'theirs',
+    'alex', 'mia', 'sam', 'echo', 'ridge', 'guardian', 'of', 'the', 'glen'
+])
+
+def process_text(text):
+    words = nltk.word_tokenize(text)
+    words = [word.lower() for word in words if word.isalnum() and word.lower() not in stop_words and word.lower() not in exclude_words]
+    word_freq = Counter(words)
+    pos_tags = nltk.pos_tag(words)
+    return word_freq, pos_tags
+
+def generate_wordcloud(word_freq):
+    wordcloud = WordCloud(width=800, height=400, background_color='white').generate_from_frequencies(word_freq)
+    plt.figure(figsize=(10, 5))
+    plt.imshow(wordcloud, interpolation='bilinear')
+    plt.axis('off')
+    plt.savefig('wordcloud.png')
+    return 'wordcloud.png'
+
+def translate_and_get_pos(word_freq, pos_tags):
+    pos_map = {
+        'NN': 'n.', 'NNS': 'n.', 'VB': 'v.', 'VBD': 'v.', 'VBG': 'v.', 'VBN': 'v.', 
+        'VBP': 'v.', 'VBZ': 'v.', 'JJ': 'adj.', 'JJR': 'adj.', 'JJS': 'adj.', 'RB': 'adv.', 
+        'RBR': 'adv.', 'RBS': 'adv.'
+    }
+
+    word_data = []
+    for word, freq in word_freq.items():
+        pos = [pos_tag[1] for pos_tag in pos_tags if pos_tag[0] == word]
+        if pos and (pos[0] in ['NNP', 'NNPS'] or word in exclude_words):
+            continue  # Skip proper nouns, pronouns, and specific excluded words
+        translation = translator.translate(word, dest='ko').text
+        pos = pos_map.get(pos[0], 'N/A') if pos else 'N/A'
+        example_sentence, synonyms = word_data_examples.get(word, (f"ex) The word '{word}' in a sentence.", ""))
+        word_data.append((word, freq, translation, pos, example_sentence, synonyms))
+    word_data.sort(key=lambda x: x[1], reverse=True)
+    return word_data[:50]
+
+def main(text):
+    word_freq, pos_tags = process_text(text)
+    wordcloud_image = generate_wordcloud(word_freq)
+    word_data = translate_and_get_pos(word_freq, pos_tags)
+    
+    word_data_str = "\n".join([f"{i+1}. {word}: {pos} {translation}, ex) {example_sentence} 동의어: {synonyms}." for i, (word, freq, translation, pos, example_sentence, synonyms) in enumerate(word_data)])
+    return wordcloud_image, word_data_str
+
+# Gradio interface
+interface = gr.Interface(
+    fn=main,
+    inputs="text",
+    outputs=["image", "text"],
+    title="Language Learning App",
+    description="Input text to generate a word cloud and a frequency list with Korean meanings, parts of speech, and example sentences."
+)
+
+# Launch the interface
+interface.launch()