File size: 15,144 Bytes
8f8d0f6 |
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 |
import logging
from typing import List, Optional
import re
from transformers import pipeline, AutoTokenizer
import torch
logger = logging.getLogger(__name__)
class TextSummarizer:
"""Text summarization with chunking for long documents"""
def __init__(self):
self.summarizer = None
self.tokenizer = None
self.max_chunk_length = 1024 # Maximum tokens per chunk
self.max_summary_length = 150
self.min_summary_length = 50
self._initialize_model()
logger.info("TextSummarizer initialized")
def _initialize_model(self):
"""Initialize the summarization model"""
try:
# Try different models in order of preference
model_names = [
"facebook/bart-large-cnn",
"sshleifer/distilbart-cnn-12-6",
"t5-small"
]
for model_name in model_names:
try:
# Use CPU to avoid memory issues on Hugging Face Spaces
device = -1 # CPU only for Hugging Face Spaces
self.summarizer = pipeline(
"summarization",
model=model_name,
tokenizer=model_name,
device=device,
framework="pt"
)
self.tokenizer = AutoTokenizer.from_pretrained(model_name)
logger.info(f"Successfully loaded summarization model: {model_name}")
break
except Exception as e:
logger.warning(f"Failed to load {model_name}: {str(e)}")
continue
if self.summarizer is None:
logger.error("Failed to load any summarization model")
except Exception as e:
logger.error(f"Error initializing summarizer: {str(e)}")
def summarize(self, text: str, max_length: int = None, min_length: int = None) -> str:
"""Summarize text with automatic chunking for long documents"""
if not text or not text.strip():
return ""
if not self.summarizer:
return self._fallback_summarize(text)
try:
# Use provided lengths or defaults
max_len = max_length or self.max_summary_length
min_len = min_length or self.min_summary_length
# Check if text needs chunking
if self._needs_chunking(text):
return self._summarize_long_text(text, max_len, min_len)
else:
return self._summarize_chunk(text, max_len, min_len)
except Exception as e:
logger.error(f"Summarization failed: {str(e)}")
return self._fallback_summarize(text)
def _needs_chunking(self, text: str) -> bool:
"""Check if text needs to be chunked"""
if not self.tokenizer:
return len(text.split()) > 300 # Rough word count threshold
try:
tokens = self.tokenizer.encode(text, add_special_tokens=True)
return len(tokens) > self.max_chunk_length
except:
return len(text.split()) > 300
def _summarize_long_text(self, text: str, max_len: int, min_len: int) -> str:
"""Summarize long text by chunking"""
try:
# Split text into chunks
chunks = self._split_into_chunks(text)
if not chunks:
return self._fallback_summarize(text)
# Summarize each chunk
chunk_summaries = []
for chunk in chunks:
if len(chunk.strip()) > 100: # Only summarize substantial chunks
summary = self._summarize_chunk(
chunk,
max_length=min(max_len // len(chunks) + 20, 100),
min_length=20
)
if summary and summary.strip():
chunk_summaries.append(summary)
if not chunk_summaries:
return self._fallback_summarize(text)
# Combine chunk summaries
combined_summary = " ".join(chunk_summaries)
# If combined summary is still too long, summarize again
if self._needs_chunking(combined_summary) and len(chunk_summaries) > 1:
final_summary = self._summarize_chunk(combined_summary, max_len, min_len)
return final_summary if final_summary else combined_summary
return combined_summary
except Exception as e:
logger.error(f"Long text summarization failed: {str(e)}")
return self._fallback_summarize(text)
def _summarize_chunk(self, text: str, max_length: int, min_length: int) -> str:
"""Summarize a single chunk of text"""
try:
if not text or len(text.strip()) < 50:
return text
# Clean text
cleaned_text = self._clean_text_for_summarization(text)
if not cleaned_text:
return text[:200] + "..." if len(text) > 200 else text
# Generate summary
result = self.summarizer(
cleaned_text,
max_length=max_length,
min_length=min_length,
do_sample=False,
truncation=True
)
if result and len(result) > 0 and 'summary_text' in result[0]:
summary = result[0]['summary_text'].strip()
# Post-process summary
summary = self._post_process_summary(summary)
return summary if summary else cleaned_text[:200] + "..."
return cleaned_text[:200] + "..."
except Exception as e:
logger.error(f"Chunk summarization failed: {str(e)}")
return text[:200] + "..." if len(text) > 200 else text
def _split_into_chunks(self, text: str) -> List[str]:
"""Split text into manageable chunks"""
try:
# Split by paragraphs first
paragraphs = [p.strip() for p in text.split('\n\n') if p.strip()]
if not paragraphs:
paragraphs = [text]
chunks = []
current_chunk = ""
current_length = 0
for paragraph in paragraphs:
paragraph_length = len(paragraph.split())
# If adding this paragraph would exceed chunk size, start new chunk
if current_length + paragraph_length > 250 and current_chunk:
chunks.append(current_chunk.strip())
current_chunk = paragraph
current_length = paragraph_length
else:
if current_chunk:
current_chunk += "\n\n" + paragraph
else:
current_chunk = paragraph
current_length += paragraph_length
# Add remaining chunk
if current_chunk.strip():
chunks.append(current_chunk.strip())
# If no proper chunks, split by sentences
if not chunks or len(chunks) == 1 and len(chunks[0].split()) > 400:
return self._split_by_sentences(text)
return chunks
except Exception as e:
logger.error(f"Text splitting failed: {str(e)}")
return [text]
def _split_by_sentences(self, text: str) -> List[str]:
"""Split text by sentences as fallback"""
try:
sentences = re.split(r'[.!?]+\s+', text)
chunks = []
current_chunk = ""
for sentence in sentences:
if len((current_chunk + " " + sentence).split()) > 200:
if current_chunk:
chunks.append(current_chunk.strip())
current_chunk = sentence
else:
if current_chunk:
current_chunk += ". " + sentence
else:
current_chunk = sentence
if current_chunk.strip():
chunks.append(current_chunk.strip())
return chunks if chunks else [text]
except Exception as e:
logger.error(f"Sentence splitting failed: {str(e)}")
return [text]
def _clean_text_for_summarization(self, text: str) -> str:
"""Clean text for better summarization"""
if not text:
return ""
# Remove URLs
text = re.sub(r'http[s]?://(?:[a-zA-Z]|[0-9]|[$-_@.&+]|[!*\\(\\),]|(?:%[0-9a-fA-F][0-9a-fA-F]))+', '', text)
# Remove email addresses
text = re.sub(r'\b[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Z|a-z]{2,}\b', '', text)
# Remove excessive whitespace
text = re.sub(r'\s+', ' ', text)
# Remove common news artifacts
artifacts = [
r'\(Reuters\)', r'\(AP\)', r'\(Bloomberg\)', r'\(CNN\)',
r'-- .*$', r'Photo:.*$', r'Image:.*$', r'Video:.*$',
r'Subscribe.*$', r'Follow us.*$'
]
for artifact in artifacts:
text = re.sub(artifact, '', text, flags=re.IGNORECASE | re.MULTILINE)
return text.strip()
def _post_process_summary(self, summary: str) -> str:
"""Post-process generated summary"""
if not summary:
return ""
# Remove incomplete sentences at the end
sentences = re.split(r'[.!?]+', summary)
if len(sentences) > 1 and len(sentences[-1].strip()) < 10:
summary = '.'.join(sentences[:-1]) + '.'
# Capitalize first letter
summary = summary[0].upper() + summary[1:] if len(summary) > 1 else summary.upper()
# Ensure summary ends with punctuation
if summary and summary[-1] not in '.!?':
summary += '.'
return summary.strip()
def _fallback_summarize(self, text: str) -> str:
"""Fallback summarization using simple extraction"""
try:
if not text or len(text.strip()) < 50:
return text
# Split into sentences
sentences = re.split(r'[.!?]+', text)
sentences = [s.strip() for s in sentences if s.strip() and len(s.split()) > 5]
if not sentences:
return text[:200] + "..." if len(text) > 200 else text
# Take first few sentences (extractive summary)
num_sentences = min(3, len(sentences))
summary_sentences = sentences[:num_sentences]
summary = '. '.join(summary_sentences)
if not summary.endswith('.'):
summary += '.'
# If summary is too long, truncate
if len(summary) > 300:
words = summary.split()
summary = ' '.join(words[:40]) + '...'
return summary
except Exception as e:
logger.error(f"Fallback summarization failed: {str(e)}")
return text[:200] + "..." if len(text) > 200 else text
def batch_summarize(self, texts: List[str], **kwargs) -> List[str]:
"""Summarize multiple texts"""
summaries = []
for text in texts:
try:
summary = self.summarize(text, **kwargs)
summaries.append(summary)
except Exception as e:
logger.error(f"Batch summarization failed for one text: {str(e)}")
summaries.append(self._fallback_summarize(text))
return summaries
def get_summary_stats(self, original_text: str, summary: str) -> dict:
"""Get statistics about the summarization"""
try:
original_words = len(original_text.split())
summary_words = len(summary.split())
compression_ratio = summary_words / original_words if original_words > 0 else 0
return {
'original_length': original_words,
'summary_length': summary_words,
'compression_ratio': compression_ratio,
'compression_percentage': (1 - compression_ratio) * 100
}
except Exception as e:
logger.error(f"Error calculating summary stats: {str(e)}")
return {
'original_length': 0,
'summary_length': 0,
'compression_ratio': 0,
'compression_percentage': 0
}
# Utility functions
def extract_key_sentences(text: str, num_sentences: int = 3) -> List[str]:
"""Extract key sentences using simple heuristics"""
try:
sentences = re.split(r'[.!?]+', text)
sentences = [s.strip() for s in sentences if s.strip() and len(s.split()) > 5]
if not sentences:
return []
# Score sentences based on position and keyword density
scored_sentences = []
for i, sentence in enumerate(sentences):
score = 0
# Position bonus (earlier sentences get higher scores)
if i < len(sentences) * 0.3:
score += 3
elif i < len(sentences) * 0.6:
score += 2
else:
score += 1
# Length bonus (medium-length sentences preferred)
words = len(sentence.split())
if 10 <= words <= 25:
score += 2
elif 5 <= words <= 35:
score += 1
# Keyword bonus (sentences with common business/finance terms)
keywords = [
'company', 'business', 'revenue', 'profit', 'growth', 'market',
'financial', 'earnings', 'investment', 'stock', 'shares', 'economy'
]
sentence_lower = sentence.lower()
keyword_count = sum(1 for keyword in keywords if keyword in sentence_lower)
score += keyword_count
scored_sentences.append((sentence, score))
# Sort by score and return top sentences
scored_sentences.sort(key=lambda x: x[1], reverse=True)
return [sent[0] for sent in scored_sentences[:num_sentences]]
except Exception as e:
logger.error(f"Key sentence extraction failed: {str(e)}")
return [] |