app.py

# app.py

import random
import csv
import gradio as gr
import pandas as pd
import tempfile

# Word lists
nouns = [
    "dog", "cat", "child", "teacher", "artist", "bird", "river", "mountain",
    "book", "city", "car", "tree", "flower", "student", "computer", "phone",
    "house", "garden", "song", "idea", "scientist", "engineer", "doctor",
    "chef", "musician", "athlete", "writer", "poet", "farmer", "pilot"
]

adjectives = [
    "quick", "lazy", "beautiful", "tall", "short", "happy", "sad", "bright",
    "dark", "colorful", "quiet", "loud", "new", "old", "young", "ancient",
    "modern", "cold", "warm", "soft", "hard", "heavy", "light", "calm",
    "stormy", "fresh", "strong", "weak", "brave"
]

verbs = [
    "run", "jump", "paint", "read", "fly", "write", "sing", "build",
    "create", "discover", "learn", "teach", "drive", "grow", "think",
    "talk", "listen", "play", "see", "walk", "swim", "code", "design",
    "cook", "dance", "draw", "study", "explore", "invent", "research"
]

adverbs = [
    "quickly", "slowly", "gracefully", "happily", "sadly", "quietly", "loudly",
    "brightly", "softly", "carefully", "eagerly", "angrily", "easily", "hardly",
    "rarely", "often", "never", "always", "sometimes", "soon", "daily", "patiently",
    "politely", "proudly", "silently", "warmly", "well", "badly", "closely", "deeply"
]

prepositions = [
    "in", "on", "over", "under", "beside", "with", "without", "near",
    "between", "through", "against", "among", "around", "before", "after",
    "inside", "outside", "above", "below", "across", "behind", "beyond",
    "during", "for", "from", "into", "like", "off", "toward"
]

articles = ["the", "a", "an"]

conjunctions = ["and", "but", "so", "because", "when", "while", "although", "if", "unless", "since"]

# Semantic rules: mapping nouns to appropriate verbs
noun_verb_map = {
    "dog": ["run", "jump", "bark", "play", "walk"],
    "cat": ["meow", "sleep", "jump", "play", "purr"],
    "child": ["play", "learn", "read", "laugh", "grow"],
    "teacher": ["teach", "explain", "guide", "help", "learn"],
    "artist": ["paint", "draw", "create", "design", "imagine"],
    "bird": ["fly", "sing", "chirp", "nest", "soar"],
    "river": ["flow", "run", "wind", "bend", "swell"],
    "mountain": ["stand", "tower", "rise", "loom", "shadow"],
    "book": ["tell", "describe", "illustrate", "explain", "reveal"],
    "city": ["grow", "expand", "develop", "bustle", "sleep"],
    "car": ["drive", "speed", "stop", "park", "honk"],
    "tree": ["grow", "sway", "stand", "shed", "bloom"],
    "flower": ["bloom", "grow", "wilt", "open", "close"],
    "student": ["study", "learn", "read", "write", "graduate"],
    "computer": ["compute", "process", "run", "crash", "boot"],
    "phone": ["ring", "vibrate", "charge", "die", "connect"],
    "house": ["stand", "shelter", "protect", "age", "burn"],
    "garden": ["grow", "bloom", "flourish", "wilt", "produce"],
    "song": ["play", "sound", "echo", "resonate", "end"],
    "idea": ["form", "grow", "develop", "emerge", "inspire"],
    "scientist": ["research", "discover", "experiment", "study", "invent"],
    "engineer": ["design", "build", "develop", "test", "solve"],
    "doctor": ["heal", "diagnose", "treat", "operate", "care"],
    "chef": ["cook", "prepare", "taste", "create", "serve"],
    "musician": ["play", "compose", "perform", "sing", "record"],
    "athlete": ["run", "train", "compete", "win", "lose"],
    "writer": ["write", "create", "imagine", "edit", "publish"],
    "poet": ["write", "compose", "imagine", "express", "rhyme"],
    "farmer": ["grow", "plant", "harvest", "plow", "raise"],
    "pilot": ["fly", "navigate", "land", "take off", "command"]
}

# Sentence templates
templates = [
    "{Article} {adjective} {noun} {adverb} {verb}s {preposition} {article} {adjective} {noun2}.",
    "{Article} {noun} {verb}s {preposition} {article} {noun2} {conjunction} {verb2}s {adverb}.",
    "{Noun_plural} {adverb} {verb} {preposition} {noun2}.",
    "{Noun} {verb}s {preposition} {article} {noun2} {conjunction} {article} {noun} {verb2}s.",
    "{Article} {adjective} {noun} {verb}s {preposition} {noun2} {conjunction} {adverb} {verb2}s.",
    "{Noun} {verb}s {article} {noun2} {preposition} {noun}."
]

def generate_sentence():
    template = random.choice(templates)
    noun = random.choice(nouns)
    # Get appropriate verbs for noun
    verbs_for_noun = noun_verb_map.get(noun, verbs)
    verb = random.choice(verbs_for_noun)

    noun2 = random.choice(nouns)
    # Ensure noun2 is different from noun
    while noun2 == noun:
        noun2 = random.choice(nouns)
    # Get appropriate verbs for noun2
    verbs_for_noun2 = noun_verb_map.get(noun2, verbs)
    verb2 = random.choice(verbs_for_noun2)

    sentence = template.format(
        Article=random.choice(articles).capitalize(),
        article=random.choice(articles),
        adjective=random.choice(adjectives),
        noun=noun,
        noun2=noun2,
        Noun=noun.capitalize(),
        Noun_plural=noun.capitalize() + "s",
        verb=verb,
        verb2=verb2,
        adverb=random.choice(adverbs),
        preposition=random.choice(prepositions),
        conjunction=random.choice(conjunctions)
    )
    # Capitalize the first letter and ensure proper punctuation
    sentence = sentence.capitalize()
    if not sentence.endswith('.'):
        sentence += '.'
    return sentence

def generate_sentences(num_sentences):
    sentences = [generate_sentence() for _ in range(int(num_sentences))]
    df = pd.DataFrame(sentences, columns=["sentence"])
    # Save to a temporary CSV file
    temp_file = tempfile.NamedTemporaryFile(delete=False, suffix=".csv")
    df.to_csv(temp_file.name, index=False)
    return temp_file.name

def generate_and_download(num_sentences):
    csv_file = generate_sentences(num_sentences)
    return csv_file

# Gradio Interface
with gr.Blocks() as demo:
    gr.Markdown(
        """
        # Sentence Dataset Generator with Semantic Rules

        Enter the number of sentences you want to generate, and download a CSV file containing the sentences.

        This generator uses semantic rules to create more coherent and meaningful sentences.
        """
    )
    num_sentences = gr.Number(label="Number of Sentences", value=1000, precision=0)
    output = gr.File(label="Download CSV")
    generate_button = gr.Button("Generate Sentences")
    generate_button.click(
        fn=generate_and_download,
        inputs=num_sentences,
        outputs=output
    )

if __name__ == "__main__":
    demo.launch()