File size: 16,177 Bytes
7819eb3
395e49f
724ad93
7819eb3
 
 
 
 
 
 
 
67f032c
7819eb3
 
 
 
 
 
 
 
 
e3a17dd
67f032c
7819eb3
67f032c
 
7819eb3
67f032c
 
 
7819eb3
724ad93
 
7819eb3
724ad93
67f032c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
395e49f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
67f032c
 
 
 
 
 
 
395e49f
 
 
 
 
 
 
 
67f032c
 
 
 
395e49f
 
67f032c
 
395e49f
 
 
 
 
 
 
 
 
 
 
 
 
 
724ad93
67f032c
724ad93
395e49f
 
 
 
724ad93
 
395e49f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
724ad93
395e49f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
724ad93
395e49f
67f032c
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
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
import gradio as gr
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch
from huggingface_hub import login
import os
import logging
from datetime import datetime
import json
from typing import List, Dict
import warnings
import spaces
import stat

# Filter out warnings
warnings.filterwarnings('ignore')

# Configure logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

# Environment variables
HF_TOKEN = os.getenv("HUGGING_FACE_TOKEN")
MODEL_NAME = os.getenv("MODEL_NAME", "google/gemma-2-2b-it")

# Create data directory for persistence in Space environment
DATA_DIR = "/home/user/app/data"
os.makedirs(DATA_DIR, exist_ok=True)
# Ensure directory has write permissions
os.chmod(DATA_DIR, stat.S_IRWXU | stat.S_IRWXG | stat.S_IRWXO)
logger.info(f"Using data directory: {DATA_DIR}")

# Cache and history files
CACHE_DIR = os.path.join(DATA_DIR, "cache")
HISTORY_FILE = os.path.join(DATA_DIR, "review_history.json")
os.makedirs(CACHE_DIR, exist_ok=True)
# Ensure cache directory has write permissions
os.chmod(CACHE_DIR, stat.S_IRWXU | stat.S_IRWXG | stat.S_IRWXO)
logger.info(f"Cache directory: {CACHE_DIR}")
logger.info(f"History file: {HISTORY_FILE}")

# Initialize empty history file if it doesn't exist
if not os.path.exists(HISTORY_FILE):
    with open(HISTORY_FILE, 'w') as f:
        json.dump({'history': [], 'metrics': {
            'total_reviews': 0,
            'avg_response_time': 0.0,
            'reviews_today': 0
        }}, f)
    # Ensure history file has write permissions
    os.chmod(HISTORY_FILE, stat.S_IRUSR | stat.S_IWUSR | stat.S_IRGRP | stat.S_IWGRP | stat.S_IROTH | stat.S_IWOTH)
    logger.info("Created empty history file")

class Review:
    def __init__(self, code: str, language: str, suggestions: str):
        self.code = code
        self.language = language
        self.suggestions = suggestions
        self.timestamp = datetime.now().isoformat()
        self.response_time = 0.0
    
    def to_dict(self):
        return {
            'timestamp': self.timestamp,
            'language': self.language,
            'code': self.code,
            'suggestions': self.suggestions,
            'response_time': self.response_time
        }
    
    @classmethod
    def from_dict(cls, data):
        review = cls(data['code'], data['language'], data['suggestions'])
        review.timestamp = data['timestamp']
        review.response_time = data['response_time']
        return review

class CodeReviewer:
    def __init__(self):
        self.model = None
        self.tokenizer = None
        self.device = None
        self.review_history: List[Review] = []
        self.metrics = {
            'total_reviews': 0,
            'avg_response_time': 0.0,
            'reviews_today': 0
        }
        self._initialized = False
        self.load_history()
        
    def load_history(self):
        """Load review history from file."""
        try:
            if os.path.exists(HISTORY_FILE):
                with open(HISTORY_FILE, 'r') as f:
                    data = json.load(f)
                    self.review_history = [Review.from_dict(r) for r in data['history']]
                    self.metrics = data['metrics']
                logger.info(f"Loaded {len(self.review_history)} reviews from history")
        except Exception as e:
            logger.error(f"Error loading history: {e}")
            # Initialize empty history if file doesn't exist or is corrupted
            self.review_history = []
            self.metrics = {
                'total_reviews': 0,
                'avg_response_time': 0.0,
                'reviews_today': 0
            }
    
    def save_history(self):
        """Save review history to file."""
        try:
            data = {
                'history': [r.to_dict() for r in self.review_history],
                'metrics': self.metrics
            }
            # Ensure directory exists with proper permissions
            os.makedirs(os.path.dirname(HISTORY_FILE), exist_ok=True)
            os.chmod(os.path.dirname(HISTORY_FILE), stat.S_IRWXU | stat.S_IRWXG | stat.S_IRWXO)
            
            with open(HISTORY_FILE, 'w') as f:
                json.dump(data, f)
            # Ensure file has proper permissions
            os.chmod(HISTORY_FILE, stat.S_IRUSR | stat.S_IWUSR | stat.S_IRGRP | stat.S_IWGRP | stat.S_IROTH | stat.S_IWOTH)
            logger.info("Saved review history")
        except Exception as e:
            logger.error(f"Error saving history: {e}")
        
    @spaces.GPU
    def ensure_initialized(self):
        """Ensure model is initialized."""
        if not self._initialized:
            self.initialize_model()
            self._initialized = True
        
    def initialize_model(self):
        """Initialize the model and tokenizer."""
        try:
            if HF_TOKEN:
                login(token=HF_TOKEN, add_to_git_credential=True)
            
            logger.info("Loading tokenizer...")
            self.tokenizer = AutoTokenizer.from_pretrained(
                MODEL_NAME,
                token=HF_TOKEN,
                trust_remote_code=True,
                cache_dir=CACHE_DIR
            )
            special_tokens = {
                'pad_token': '[PAD]',
                'eos_token': '</s>',
                'bos_token': '<s>'
            }
            num_added = self.tokenizer.add_special_tokens(special_tokens)
            logger.info(f"Added {num_added} special tokens")
            logger.info("Tokenizer loaded successfully")
            
            logger.info("Loading model...")
            self.model = AutoModelForCausalLM.from_pretrained(
                MODEL_NAME,
                device_map="auto",
                torch_dtype=torch.float16,
                trust_remote_code=True,
                low_cpu_mem_usage=True,
                cache_dir=CACHE_DIR,
                token=HF_TOKEN
            )
            if num_added > 0:
                logger.info("Resizing model embeddings for special tokens")
                self.model.resize_token_embeddings(len(self.tokenizer))
            
            self.device = next(self.model.parameters()).device
            logger.info(f"Model loaded successfully on {self.device}")
            self._initialized = True
            return True
        except Exception as e:
            logger.error(f"Error initializing model: {e}")
            self._initialized = False
            return False

    def create_review_prompt(self, code: str, language: str) -> str:
        """Create a structured prompt for code review."""
        return f"""Review this {language} code. List specific points in these sections:
Issues:
Improvements:
Best Practices:
Security:

Code:
```{language}
{code}
```"""

    @spaces.GPU
    def review_code(self, code: str, language: str) -> str:
        """Perform code review using the model."""
        try:
            if not self._initialized and not self.initialize_model():
                return "Error: Model initialization failed. Please try again later."
            
            start_time = datetime.now()
            prompt = self.create_review_prompt(code, language)
            
            try:
                inputs = self.tokenizer(
                    prompt,
                    return_tensors="pt",
                    truncation=True,
                    max_length=512,
                    padding=True
                )
                if inputs is None:
                    raise ValueError("Failed to tokenize input")
                inputs = inputs.to(self.device)
            except Exception as token_error:
                logger.error(f"Tokenization error: {token_error}")
                return "Error: Failed to process input code. Please try again."
            
            try:
                with torch.no_grad():
                    outputs = self.model.generate(
                        **inputs,
                        max_new_tokens=512,
                        do_sample=True,
                        temperature=0.7,
                        top_p=0.95,
                        num_beams=1,
                        early_stopping=True,
                        pad_token_id=self.tokenizer.pad_token_id,
                        eos_token_id=self.tokenizer.eos_token_id
                    )
            except Exception as gen_error:
                logger.error(f"Generation error: {gen_error}")
                return "Error: Failed to generate review. Please try again."
            
            try:
                response = self.tokenizer.decode(outputs[0], skip_special_tokens=True)
                suggestions = response[len(prompt):].strip()
            except Exception as decode_error:
                logger.error(f"Decoding error: {decode_error}")
                return "Error: Failed to decode model output. Please try again."
            
            # Create and save review
            end_time = datetime.now()
            review = Review(code, language, suggestions)
            review.response_time = (end_time - start_time).total_seconds()
            
            # Update metrics first
            self.metrics['total_reviews'] += 1
            total_time = self.metrics['avg_response_time'] * (self.metrics['total_reviews'] - 1)
            total_time += review.response_time
            self.metrics['avg_response_time'] = total_time / self.metrics['total_reviews']
            
            today = datetime.now().date()
            
            # Add review to history
            self.review_history.append(review)
            
            # Update today's reviews count
            self.metrics['reviews_today'] = sum(
                1 for r in self.review_history 
                if datetime.fromisoformat(r.timestamp).date() == today
            )
            
            # Save to file
            self.save_history()
            
            if self.device and self.device.type == "cuda":
                del inputs, outputs
                torch.cuda.empty_cache()
            
            return suggestions
            
        except Exception as e:
            logger.error(f"Error during code review: {e}")
            return f"Error performing code review: {str(e)}"

    def update_metrics(self, review: Review):
        """Update metrics with new review."""
        self.metrics['total_reviews'] += 1
        
        total_time = self.metrics['avg_response_time'] * (self.metrics['total_reviews'] - 1)
        total_time += review.response_time
        self.metrics['avg_response_time'] = total_time / self.metrics['total_reviews']
        
        today = datetime.now().date()
        self.metrics['reviews_today'] = sum(
            1 for r in self.review_history 
            if datetime.fromisoformat(r.timestamp).date() == today
        )

    def get_history(self) -> List[Dict]:
        """Get formatted review history."""
        return [
            {
                'timestamp': r.timestamp,
                'language': r.language,
                'code': r.code,
                'suggestions': r.suggestions,
                'response_time': f"{r.response_time:.2f}s"
            }
            for r in reversed(self.review_history[-10:])
        ]

    def get_metrics(self) -> Dict:
        """Get current metrics."""
        return {
            'Total Reviews': self.metrics['total_reviews'],
            'Average Response Time': f"{self.metrics['avg_response_time']:.2f}s",
            'Reviews Today': self.metrics['reviews_today'],
            'Device': str(self.device) if self.device else "Not initialized"
        }

# Initialize reviewer
reviewer = CodeReviewer()

# Create Gradio interface
with gr.Blocks(theme=gr.themes.Soft()) as iface:
    gr.Markdown("# Code Review Assistant")
    gr.Markdown("An automated code review system powered by Gemma-2b")
    
    with gr.Tabs():
        with gr.Tab("Review Code"):
            with gr.Row():
                with gr.Column():
                    code_input = gr.Textbox(
                        lines=10,
                        placeholder="Enter your code here...",
                        label="Code"
                    )
                    language_input = gr.Dropdown(
                        choices=["python", "javascript", "java", "cpp", "typescript", "go", "rust"],
                        value="python",
                        label="Language"
                    )
                    submit_btn = gr.Button("Submit for Review", variant="primary")
                with gr.Column():
                    output = gr.Textbox(
                        label="Review Results",
                        lines=10
                    )
        
        with gr.Tab("History"):
            with gr.Row():
                refresh_history = gr.Button("Refresh History", variant="secondary")
            history_output = gr.Textbox(
                label="Review History",
                lines=20,
                value="Click 'Refresh History' to view review history"
            )
        
        with gr.Tab("Metrics"):
            with gr.Row():
                refresh_metrics = gr.Button("Refresh Metrics", variant="secondary")
            metrics_output = gr.JSON(
                label="Performance Metrics"
            )
    
    @spaces.GPU
    def review_code_interface(code: str, language: str) -> str:
        if not code.strip():
            return "Please enter some code to review."
        try:
            reviewer.ensure_initialized()
            result = reviewer.review_code(code, language)
            return result
        except Exception as e:
            logger.error(f"Interface error: {e}")
            return f"Error: {str(e)}"
    
    def get_history_interface() -> str:
        try:
            history = reviewer.get_history()
            if not history:
                return "No reviews yet."
            result = ""
            for review in history:
                result += f"Time: {review['timestamp']}\n"
                result += f"Language: {review['language']}\n"
                result += f"Response Time: {review['response_time']}\n"
                result += "Code:\n```\n" + review['code'] + "\n```\n"
                result += "Suggestions:\n" + review['suggestions'] + "\n"
                result += "-" * 80 + "\n\n"
            return result
        except Exception as e:
            logger.error(f"History error: {e}")
            return "Error retrieving history"
    
    def get_metrics_interface() -> Dict:
        try:
            metrics = reviewer.get_metrics()
            if not metrics:
                return {
                    'Total Reviews': 0,
                    'Average Response Time': '0.00s',
                    'Reviews Today': 0,
                    'Device': str(reviewer.device) if reviewer.device else "Not initialized"
                }
            return metrics
        except Exception as e:
            logger.error(f"Metrics error: {e}")
            return {"error": str(e)}
    
    def update_all_outputs(code: str, language: str) -> tuple:
        """Update all outputs after code review."""
        result = review_code_interface(code, language)
        history = get_history_interface()
        metrics = get_metrics_interface()
        return result, history, metrics
    
    # Connect the interface
    submit_btn.click(
        update_all_outputs,
        inputs=[code_input, language_input],
        outputs=[output, history_output, metrics_output]
    )
    
    refresh_history.click(
        get_history_interface,
        outputs=history_output
    )
    
    refresh_metrics.click(
        get_metrics_interface,
        outputs=metrics_output
    )
    
    # Add example inputs
    gr.Examples(
        examples=[
            ["""def add_numbers(a, b):
    return a + b""", "python"],
            ["""function calculateSum(numbers) {
    let sum = 0;
    for(let i = 0; i < numbers.length; i++) {
        sum += numbers[i];
    }
    return sum;
}""", "javascript"]
        ],
        inputs=[code_input, language_input]
    )

# Launch the app
if __name__ == "__main__":
    iface.launch()