Spaces:
Running
Running
File size: 16,762 Bytes
ee7d7b9 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 | """
🗣️ NLP ENGINE - DataVision Natural Language Processing
=======================================================
Enterprise NLP capabilities:
- Named Entity Recognition (business entities)
- Intent Classification with confidence
- Sentiment Analysis for feedback data
- Keyword Extraction
- Text Summarization
FREE: Uses Groq/Gemini LLMs + local spaCy when available
"""
import re
import logging
from typing import Dict, List, Optional, Any, Tuple
from dataclasses import dataclass, field
from datetime import datetime
from enum import Enum
import json
from core.llm import chat
logger = logging.getLogger(__name__)
# Try to import spaCy for better NER
SPACY_AVAILABLE = False
nlp = None
try:
import spacy
nlp = spacy.load("en_core_web_sm")
SPACY_AVAILABLE = True
logger.info("✅ spaCy loaded for NLP")
except:
logger.info("⚠️ spaCy not available, using LLM-based NLP")
class EntityType(Enum):
"""Types of entities we care about in business data"""
COMPANY = "company"
PRODUCT = "product"
PERSON = "person"
METRIC = "metric"
DATE = "date"
MONEY = "money"
PERCENTAGE = "percentage"
QUANTITY = "quantity"
LOCATION = "location"
COLUMN = "column" # Data column reference
@dataclass
class Entity:
"""An extracted entity"""
text: str
type: EntityType
start: int
end: int
confidence: float = 1.0
@dataclass
class Sentiment:
"""Sentiment analysis result"""
label: str # positive, negative, neutral
score: float # -1 to 1
confidence: float
aspects: Dict[str, str] = field(default_factory=dict) # aspect -> sentiment
@dataclass
class NLPResult:
"""Complete NLP analysis result"""
original_text: str
entities: List[Entity]
sentiment: Optional[Sentiment]
keywords: List[str]
summary: Optional[str]
processing_time_ms: int
class NLPEngine:
"""
🗣️ DataVision NLP Engine
Provides:
- Entity extraction (companies, products, metrics, dates, etc.)
- Sentiment analysis for feedback/reviews
- Keyword extraction
- Text summarization
"""
def __init__(self):
self.entity_patterns = self._build_entity_patterns()
def _build_entity_patterns(self) -> Dict[EntityType, List[str]]:
"""Build regex patterns for entity extraction"""
return {
EntityType.MONEY: [
r'\$[\d,]+\.?\d*',
r'₹[\d,]+\.?\d*',
r'€[\d,]+\.?\d*',
r'£[\d,]+\.?\d*',
r'\b\d+(?:,\d{3})*(?:\.\d+)?\s*(?:dollars?|USD|rupees?|INR|euros?|EUR)\b',
],
EntityType.PERCENTAGE: [
r'\b\d+(?:\.\d+)?%',
r'\b\d+(?:\.\d+)?\s*percent\b',
],
EntityType.DATE: [
r'\b\d{4}-\d{2}-\d{2}\b',
r'\b\d{2}/\d{2}/\d{4}\b',
r'\b(?:Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)[a-z]*\s+\d{1,2},?\s*\d{4}\b',
r'\bQ[1-4]\s*\d{4}\b',
r'\b(?:last|next|this)\s+(?:week|month|year|quarter)\b',
],
EntityType.QUANTITY: [
r'\b\d+(?:,\d{3})*(?:\.\d+)?\s*(?:units?|items?|pieces?|orders?)\b',
],
}
async def analyze(
self,
text: str,
extract_sentiment: bool = True,
extract_keywords: bool = True,
summarize: bool = False,
max_summary_length: int = 150
) -> NLPResult:
"""
Full NLP analysis of text
Args:
text: Text to analyze
extract_sentiment: Whether to analyze sentiment
extract_keywords: Whether to extract keywords
summarize: Whether to generate summary
max_summary_length: Max length of summary
Returns:
Complete NLP result
"""
start_time = datetime.now()
# 1. Extract entities
entities = await self.extract_entities(text)
# 2. Sentiment analysis
sentiment = None
if extract_sentiment:
sentiment = await self.analyze_sentiment(text)
# 3. Keyword extraction
keywords = []
if extract_keywords:
keywords = await self.extract_keywords(text)
# 4. Summarization
summary = None
if summarize and len(text) > max_summary_length * 2:
summary = await self.summarize(text, max_summary_length)
duration_ms = int((datetime.now() - start_time).total_seconds() * 1000)
return NLPResult(
original_text=text,
entities=entities,
sentiment=sentiment,
keywords=keywords,
summary=summary,
processing_time_ms=duration_ms
)
async def extract_entities(
self,
text: str,
data_columns: Optional[List[str]] = None
) -> List[Entity]:
"""Extract named entities from text"""
entities = []
# 1. Pattern-based extraction (fast, reliable)
entities.extend(self._extract_pattern_entities(text))
# 2. Extract column references if schema provided
if data_columns:
entities.extend(self._extract_column_references(text, data_columns))
# 3. Use spaCy if available
if SPACY_AVAILABLE and nlp:
entities.extend(self._extract_spacy_entities(text))
else:
# Use LLM for complex entity extraction
llm_entities = await self._extract_llm_entities(text)
entities.extend(llm_entities)
# Deduplicate and sort
seen = set()
unique_entities = []
for e in sorted(entities, key=lambda x: x.start):
key = (e.text.lower(), e.type)
if key not in seen:
seen.add(key)
unique_entities.append(e)
return unique_entities
def _extract_pattern_entities(self, text: str) -> List[Entity]:
"""Extract entities using regex patterns"""
entities = []
for entity_type, patterns in self.entity_patterns.items():
for pattern in patterns:
for match in re.finditer(pattern, text, re.IGNORECASE):
entities.append(Entity(
text=match.group(),
type=entity_type,
start=match.start(),
end=match.end(),
confidence=0.9
))
return entities
def _extract_column_references(
self,
text: str,
columns: List[str]
) -> List[Entity]:
"""Extract references to data columns"""
entities = []
text_lower = text.lower()
for col in columns:
col_lower = col.lower()
# Look for exact match or with underscores replaced
patterns = [
col_lower,
col_lower.replace('_', ' '),
col_lower.replace('_', ''),
]
for pattern in patterns:
idx = text_lower.find(pattern)
if idx >= 0:
entities.append(Entity(
text=col,
type=EntityType.COLUMN,
start=idx,
end=idx + len(pattern),
confidence=0.95
))
break
return entities
def _extract_spacy_entities(self, text: str) -> List[Entity]:
"""Extract entities using spaCy"""
entities = []
if not nlp:
return entities
doc = nlp(text)
type_mapping = {
'ORG': EntityType.COMPANY,
'PRODUCT': EntityType.PRODUCT,
'PERSON': EntityType.PERSON,
'DATE': EntityType.DATE,
'MONEY': EntityType.MONEY,
'PERCENT': EntityType.PERCENTAGE,
'QUANTITY': EntityType.QUANTITY,
'GPE': EntityType.LOCATION,
'LOC': EntityType.LOCATION,
}
for ent in doc.ents:
if ent.label_ in type_mapping:
entities.append(Entity(
text=ent.text,
type=type_mapping[ent.label_],
start=ent.start_char,
end=ent.end_char,
confidence=0.85
))
return entities
async def _extract_llm_entities(self, text: str) -> List[Entity]:
"""Extract entities using LLM"""
entities = []
if len(text) > 2000:
text = text[:2000]
prompt = f"""Extract named entities from this text. Return JSON array.
Text: "{text}"
Extract these types: COMPANY, PRODUCT, PERSON, METRIC, DATE
Format:
[{{"text": "entity text", "type": "ENTITY_TYPE"}}]
Return only the JSON array, nothing else."""
try:
response = chat(messages=prompt, temperature=0.1, max_tokens=500)
# Parse JSON from response
json_match = response[response.find('['):response.rfind(']')+1]
if json_match:
llm_entities = json.loads(json_match)
for item in llm_entities:
entity_type = EntityType.COMPANY # Default
type_str = item.get("type", "").upper()
type_map = {
'COMPANY': EntityType.COMPANY,
'PRODUCT': EntityType.PRODUCT,
'PERSON': EntityType.PERSON,
'METRIC': EntityType.METRIC,
'DATE': EntityType.DATE,
}
if type_str in type_map:
entity_type = type_map[type_str]
text_val = item.get("text", "")
idx = text.find(text_val)
entities.append(Entity(
text=text_val,
type=entity_type,
start=idx if idx >= 0 else 0,
end=idx + len(text_val) if idx >= 0 else len(text_val),
confidence=0.75
))
except Exception as e:
logger.warning(f"LLM entity extraction failed: {e}")
return entities
async def analyze_sentiment(
self,
text: str,
aspects: Optional[List[str]] = None
) -> Sentiment:
"""
Analyze sentiment of text
Args:
text: Text to analyze
aspects: Optional specific aspects to analyze (e.g., ['price', 'quality'])
Returns:
Sentiment analysis result
"""
if len(text) > 1000:
text = text[:1000]
prompt = f"""Analyze the sentiment of this text.
Text: "{text}"
Provide:
1. Overall sentiment: positive, negative, or neutral
2. Score: -1.0 (very negative) to 1.0 (very positive)
3. Confidence: 0.0 to 1.0
Format your response exactly as:
SENTIMENT: [positive/negative/neutral]
SCORE: [number]
CONFIDENCE: [number]"""
try:
response = chat(messages=prompt, temperature=0.1, max_tokens=100)
sentiment_match = re.search(r'SENTIMENT:\s*(\w+)', response, re.IGNORECASE)
score_match = re.search(r'SCORE:\s*([-\d.]+)', response, re.IGNORECASE)
conf_match = re.search(r'CONFIDENCE:\s*([\d.]+)', response, re.IGNORECASE)
label = sentiment_match.group(1).lower() if sentiment_match else "neutral"
score = float(score_match.group(1)) if score_match else 0.0
confidence = float(conf_match.group(1)) if conf_match else 0.7
# Clamp values
score = max(-1.0, min(1.0, score))
confidence = max(0.0, min(1.0, confidence))
return Sentiment(
label=label,
score=score,
confidence=confidence,
aspects={}
)
except Exception as e:
logger.warning(f"Sentiment analysis failed: {e}")
return Sentiment(label="neutral", score=0.0, confidence=0.5)
async def extract_keywords(
self,
text: str,
max_keywords: int = 10
) -> List[str]:
"""Extract important keywords from text"""
if len(text) > 1000:
text = text[:1000]
prompt = f"""Extract the {max_keywords} most important keywords from this text.
Text: "{text}"
Return only a comma-separated list of keywords, nothing else."""
try:
response = chat(messages=prompt, temperature=0.1, max_tokens=100)
# Parse comma-separated keywords
keywords = [k.strip() for k in response.split(',')]
return [k for k in keywords if k and len(k) > 2][:max_keywords]
except Exception as e:
logger.warning(f"Keyword extraction failed: {e}")
return []
async def summarize(
self,
text: str,
max_length: int = 150
) -> str:
"""Summarize long text"""
if len(text) <= max_length:
return text
prompt = f"""Summarize this text in {max_length} characters or less. Be concise but capture key points.
Text: "{text[:3000]}"
Summary:"""
try:
response = chat(messages=prompt, temperature=0.3, max_tokens=200)
return response.strip()[:max_length*2] # Allow some flexibility
except Exception as e:
logger.warning(f"Summarization failed: {e}")
return text[:max_length] + "..."
async def analyze_reviews(
self,
reviews: List[str]
) -> Dict[str, Any]:
"""
Analyze a batch of reviews/feedback
Returns aggregate sentiment, common themes, and issues
"""
sentiments = []
all_keywords = []
for review in reviews[:50]: # Limit to 50 reviews
sentiment = await self.analyze_sentiment(review)
sentiments.append(sentiment)
keywords = await self.extract_keywords(review, max_keywords=5)
all_keywords.extend(keywords)
# Aggregate results
avg_score = sum(s.score for s in sentiments) / len(sentiments) if sentiments else 0
positive_count = sum(1 for s in sentiments if s.label == "positive")
negative_count = sum(1 for s in sentiments if s.label == "negative")
neutral_count = sum(1 for s in sentiments if s.label == "neutral")
# Count keyword frequency
keyword_freq = {}
for kw in all_keywords:
kw_lower = kw.lower()
keyword_freq[kw_lower] = keyword_freq.get(kw_lower, 0) + 1
top_keywords = sorted(keyword_freq.items(), key=lambda x: x[1], reverse=True)[:10]
return {
"total_reviews": len(reviews),
"analyzed": len(sentiments),
"average_sentiment_score": round(avg_score, 2),
"sentiment_distribution": {
"positive": positive_count,
"negative": negative_count,
"neutral": neutral_count
},
"top_keywords": [{"keyword": k, "count": c} for k, c in top_keywords],
"overall_label": "positive" if avg_score > 0.2 else "negative" if avg_score < -0.2 else "neutral"
}
# Global instance
_nlp_engine: Optional[NLPEngine] = None
def get_nlp_engine() -> NLPEngine:
"""Get or create the global NLP engine"""
global _nlp_engine
if _nlp_engine is None:
_nlp_engine = NLPEngine()
return _nlp_engine
async def extract_entities(
text: str,
data_columns: Optional[List[str]] = None
) -> List[Dict]:
"""Quick function to extract entities"""
engine = get_nlp_engine()
entities = await engine.extract_entities(text, data_columns)
return [
{"text": e.text, "type": e.type.value, "confidence": e.confidence}
for e in entities
]
async def analyze_sentiment(text: str) -> Dict:
"""Quick function for sentiment analysis"""
engine = get_nlp_engine()
result = await engine.analyze_sentiment(text)
return {
"label": result.label,
"score": result.score,
"confidence": result.confidence
}
|