File size: 14,992 Bytes
85c7ef6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
import os
import random
import pickle
import re
import tempfile
import numpy as np
import cv2
import gradio as gr
from PIL import Image
from sentence_transformers import SentenceTransformer
from transformers import AutoModelForImageClassification, AutoImageProcessor
from gradio_client import Client, handle_file
import openai
from flask import Flask, request, jsonify, Response
from flask_cors import CORS
import base64
import io

# ============================================================
# INITIALIZATION
# ============================================================
print("Initializing Tree & House Story Generator (Flask + Gradio)")

# OpenAI
client = openai.OpenAI(
    base_url="https://api.llm7.io/v1",
    api_key=os.environ.get("LLM7_API_KEY", "unused")
)

# NSFW MODEL
nsfw_processor = AutoImageProcessor.from_pretrained("Falconsai/nsfw_image_detection")
nsfw_model = AutoModelForImageClassification.from_pretrained("Falconsai/nsfw_image_detection")

# CLIP Interrogator
clipi_client = Client("https://fffiloni-clip-interrogator-2.hf.space/")

# SAFETY
UNSAFE_KEYWORDS = [
    "penis", "vagina", "genitals", "testicle",
    "sex", "sexual", "erotic", "porn", "explicit",
    "nipple", "genital", "lewd"
]

def is_safe_description(text):
    text = text.lower()
    return not any(re.search(rf"\b{w}\b", text) for w in UNSAFE_KEYWORDS)

def is_drawing_safe(image):
    try:
        inputs = nsfw_processor(images=image, return_tensors="pt")
        outputs = nsfw_model(**inputs)
        probs = outputs.logits.softmax(dim=1)[0]
        if probs[1] > 0.85:
            return False, "NSFW detected"
        return True, None
    except:
        return True, None

# LOAD TREE & HOUSE MODEL
classifier = None
bert_model = None
tree_stories = []
house_stories = []

try:
    with open("tree_house_model.pkl", "rb") as f:
        model_package = pickle.load(f)

    classifier = model_package["classifier"]
    bert_model = SentenceTransformer(model_package["bert_model_name"])
    tree_stories = model_package["tree_stories"]
    house_stories = model_package["house_stories"]
    model_accuracy = model_package["accuracy"]

    print(f"✓ Model loaded ({model_accuracy*100:.2f}%)")
except:
    print("Tree/House model not loaded")

# CORE FUNCTIONS
def get_image_description(image):
    with tempfile.NamedTemporaryFile(suffix=".jpg", delete=False) as f:
        path = f.name
        image.save(path)
    try:
        desc = clipi_client.predict(
            image=handle_file(path),
            mode="best",
            api_name="/clipi2"
        )
    finally:
        os.remove(path)
    return desc

def predict_category(description):
    if not classifier:
        return None, 0.0, {}
    emb = bert_model.encode([description], normalize_embeddings=True)
    pred = classifier.predict(emb)[0]
    probs = classifier.predict_proba(emb)[0]
    return pred, float(max(probs)), {"tree": float(probs[0]), "house": float(probs[1])}

def get_random_story(category):
    stories = tree_stories if category == "tree" else house_stories
    if not stories:
        return None
    s = random.choice(stories).copy()
    s.pop("category", None)
    return s

def generate_story_openai(description, audience, is_vulgar=False):
    """
    description: user drawing description (string or list)
    audience: string, e.g., 'children aged 5-7'
    is_vulgar: boolean, True if Hugging Face classifier detects vulgar content
    """

    # Handle list input
    if isinstance(description, list):
        first_desc = description[0]
    else:
        first_desc = description

    # Pre-check using local unsafe keywords
    lower_desc = first_desc.lower()
    for word in UNSAFE_KEYWORDS:
        if word in lower_desc:
            return "Your drawing contains inappropriate content and cannot be used to generate a story."

    # If Hugging Face classifier detects vulgar content
    if is_vulgar:
        return "Your drawing contains inappropriate content and cannot be used to generate a story."

    # Prompt with explicit instruction
    prompt = (
        f"Create a short, kid-friendly story for {audience} about: {first_desc}. "
        f"Use simple, cheerful words suitable for children. Include characters, action, "
        f"and make it imaginative. "
        f"Write only 3 paragraphs. "
        f"Also, provide a creative title at the very beginning. "
        f"IMPORTANT: If the input contains any vulgar, sexual, or inappropriate content, "
        f"do NOT generate any story and reply only with: "
        f"'The content is inappropriate and cannot be used for a story.' "
        f"Do NOT add extra questions, suggestions, or prompts at the end."
    )

    try:
        res = client.chat.completions.create(
            model="gpt-5-chat",
            messages=[{"role": "user", "content": prompt}],
            temperature=0.8
        )
        return res.choices[0].message.content
    except:
        return "Story generation failed."


def format_story(story_data):
    """Return story text ONLY – no labels like characters, setting, plot, etc."""

    if isinstance(story_data, str):
        return story_data

    elif isinstance(story_data, dict):
        parts = []

        # Optional title (pwede mo rin alisin kung ayaw mo)
        if "title" in story_data and story_data["title"]:
            parts.append(story_data["title"].strip())

        # All possible story parts (NO LABELS)
        ordered_keys = [
            "characters",
            "setting",
            "plot",
            "problem",
            "ending",
            "lesson",
            "moral",
            "story"
        ]

        for key in ordered_keys:
            if key in story_data and story_data[key]:
                parts.append(story_data[key].strip())

        # Join as clean paragraphs
        return "\n\n".join(parts)

    else:
        return str(story_data)

def process_image_base64(image_base64, audience="Children"):
    image_data = base64.b64decode(image_base64)
    image = Image.open(io.BytesIO(image_data)).convert("RGB")

    # Description mula sa CLIP
    description = get_image_description(image)
    
    # Predict category (tree, house, other)
    category, confidence, probs = predict_category(description)

    # Check if description contains tree/house keywords
    description_lower = description.lower()
    is_tree = "tree" in description_lower or "branch" in description_lower or "leaf" in description_lower or "trunk" in description_lower
    is_house = "house" in description_lower or "building" in description_lower or "home" in description_lower or "roof" in description_lower

    # TREE/HOUSE → always use dataset story, ignore vulgar/NSFW
    if (category == "tree" and is_tree) or (category == "house" and is_house):
        story_data = get_random_story("tree" if category == "tree" else "house")
        story_category = category
        formatted_story = format_story(story_data)
        return {
            "success": True,
            "description": description,
            "category": story_category,
            "confidence": confidence,
            "probabilities": probs,
            "story": formatted_story
        }

    # Ibang drawings → GPT-generated, check NSFW and vulgar keywords
    safe, reason = is_drawing_safe(image)
    if not safe:
        return {"success": False, "error": "Bawal sa bata: NSFW detected", "reason": reason}

    if not is_safe_description(description):
        return {"success": False, "error": "Bawal sa bata: unsafe description detected"}

    formatted_story = generate_story_openai(description, audience)
    story_category = "OPENAI"

    return {
        "success": True,
        "description": description,
        "category": story_category,
        "confidence": None,
        "probabilities": probs,
        "story": formatted_story
    }



# GRADIO FUNCTION
def generate_story_gradio(image, audience):
    buffered = io.BytesIO()
    image.save(buffered, format="PNG")
    img_b64 = base64.b64encode(buffered.getvalue()).decode()
    result = process_image_base64(img_b64, audience)
    if not result["success"]:
        return result.get("error", "Failed"), "", "", ""
    # Story is already formatted as a string
    return result["description"], result["category"], str(result["probabilities"]), result["story"]

# GRADIO UI
with gr.Blocks(title=" Tree & House Story Generator") as demo:
    gr.Markdown("# Tree & House Story Generator")
    gr.Markdown("Upload a drawing and generate a creative story!")
    
    with gr.Row():
        with gr.Column():
            image_input = gr.Image(type="pil", label="Upload a drawing")
            audience_input = gr.Dropdown(["Children", "Teens"], value="Children", label="Audience")
            btn = gr.Button(" Generate Story", variant="primary")
        
        with gr.Column():
            description_output = gr.Textbox(label="Image Description", lines=2)
            category_output = gr.Textbox(label="Category")
            probabilities_output = gr.Textbox(label="Probabilities")
            story_output = gr.Textbox(label="Story", lines=12)
    
    btn.click(
        generate_story_gradio,
        inputs=[image_input, audience_input],
        outputs=[description_output, category_output, probabilities_output, story_output]
    )

# FLASK API
app = Flask(__name__)
CORS(app)

# Create Gradio WSGI app with SSR disabled
gradio_app = gr.routes.App.create_app(demo, ssr_mode=False)

@app.route('/')
def index():
    return Response("Redirecting to Gradio UI...", status=302, headers={"Location": "/gradio"})

@app.route("/api/health", methods=["GET"])
def health_check():
    return jsonify({
        "status": "healthy",
        "model_loaded": classifier is not None,
        "tree_stories": len(tree_stories),
        "house_stories": len(house_stories)
    })

@app.route("/api/generate-story-base64", methods=["POST"])
def generate_story_api():
    data = request.get_json()
    if "image" not in data:
        return jsonify({"error": "No image provided"}), 400
    audience = data.get("audience", "Children")
    result = process_image_base64(data["image"], audience)
    return jsonify(result)

@app.route('/api/test', methods=['GET'])
def api_test():
    """API documentation and test page"""
    html = """
    <!DOCTYPE html>
    <html>
    <head>
        <title>API Test - Story Generator</title>
        <style>
            body { font-family: Arial, sans-serif; max-width: 800px; margin: 50px auto; padding: 20px; }
            h1 { color: #333; }
            .endpoint { background: #f5f5f5; padding: 15px; margin: 20px 0; border-radius: 5px; }
            .method { display: inline-block; padding: 5px 10px; border-radius: 3px; font-weight: bold; }
            .get { background: #61affe; color: white; }
            .post { background: #49cc90; color: white; }
            code { background: #f5f5f5; padding: 2px 5px; border-radius: 3px; }
            .test-section { margin: 20px 0; padding: 20px; border: 1px solid #ddd; border-radius: 5px; }
            button { background: #49cc90; color: white; border: none; padding: 10px 20px; cursor: pointer; border-radius: 5px; }
            button:hover { background: #3da877; }
            #result { margin-top: 15px; padding: 15px; background: #f9f9f9; border-radius: 5px; white-space: pre-wrap; }
        </style>
    </head>
    <body>
        <h1> Story Generator API</h1>
        
        <div class="endpoint">
            <span class="method get">GET</span> <code>/api/health</code>
            <p>Check API status</p>
        </div>
        
        <div class="endpoint">
            <span class="method post">POST</span> <code>/api/generate-story-base64</code>
            <p>Generate story from base64 image</p>
            <strong>Request body:</strong>
            <pre>{
  "image": "base64_encoded_image_string",
  "audience": "Children"
}</pre>
        </div>

        <div class="test-section">
            <h2>Test API</h2>
            <p>Upload an image to test the story generation:</p>
            <input type="file" id="imageInput" accept="image/*">
            <br><br>
            <label>Audience: 
                <select id="audience">
                    <option>Children</option>
                    <option>Teens</option>
                </select>
            </label>
            <br><br>
            <button onclick="testAPI()">Generate Story</button>
            <div id="result"></div>
        </div>

        <script>
            async function testAPI() {
                const fileInput = document.getElementById('imageInput');
                const audience = document.getElementById('audience').value;
                const resultDiv = document.getElementById('result');
                
                if (!fileInput.files[0]) {
                    resultDiv.textContent = 'Please select an image first!';
                    return;
                }
                
                resultDiv.textContent = 'Processing...';
                
                const reader = new FileReader();
                reader.onload = async function(e) {
                    const base64 = e.target.result.split(',')[1];
                    
                    try {
                        const response = await fetch('/api/generate-story-base64', {
                            method: 'POST',
                            headers: {
                                'Content-Type': 'application/json'
                            },
                            body: JSON.stringify({
                                image: base64,
                                audience: audience
                            })
                        });
                        
                        const data = await response.json();
                        resultDiv.textContent = JSON.stringify(data, null, 2);
                    } catch (error) {
                        resultDiv.textContent = 'Error: ' + error.message;
                    }
                };
                
                reader.readAsDataURL(fileInput.files[0]);
            }
        </script>
    </body>
    </html>
    """
    return html

# Mount Gradio at /gradio using WSGI middleware
from werkzeug.middleware.dispatcher import DispatcherMiddleware

app.wsgi_app = DispatcherMiddleware(
    app.wsgi_app,
    {'/gradio': gradio_app}
)

# RUN FLASK APP
if __name__ == "__main__":
    port = int(os.environ.get("PORT", 7860))
    print("=" * 60)
    print(" Flask + Gradio Server Starting")
    print("=" * 60)
    print(f" Gradio UI:           http://localhost:{port}/gradio")
    print(f" Home (Redirect):     http://localhost:{port}/")
    print(f" API Test Page:       http://localhost:{port}/api/test")
    print(f" Health Check:        http://localhost:{port}/api/health")
    print(f" API Endpoint:        http://localhost:{port}/api/generate-story-base64")
    print("=" * 60)
    app.run(host="0.0.0.0", port=port, debug=False)