File size: 10,503 Bytes
b30f068 | 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 | """
Parameter Extractor
Extracts parameters (city, location, etc.) from natural language queries
"""
import re
import spacy
from typing import Optional, Dict, List
# Load spaCy model (en_core_web_sm)
try:
nlp = spacy.load("en_core_web_sm")
except OSError:
print("⚠️ spaCy model not found. Run: python -m spacy download en_core_web_sm")
nlp = None
def extract_city_from_query(query: str) -> Optional[str]:
"""
Extract city name from a natural language query
Args:
query: User's natural language question
Returns:
City name if found, None otherwise
Examples:
"What's the weather in London?" -> "London"
"Show me weather for New York" -> "New York"
"Temperature in Tokyo" -> "Tokyo"
"""
if nlp is None:
# Fallback to regex if spaCy not available
return _extract_city_regex(query)
# Method 1: Try spaCy NER (Named Entity Recognition)
doc = nlp(query)
for ent in doc.ents:
if ent.label_ in ["GPE", "LOC"]: # Geopolitical Entity or Location
return ent.text
# Method 2: Try regex patterns
city = _extract_city_regex(query)
if city:
return city
return None
def _extract_city_regex(query: str) -> Optional[str]:
"""
Extract city using regex patterns
Args:
query: User query
Returns:
City name if found
"""
# Common patterns for city extraction
patterns = [
r"(?:weather|temperature|forecast|climate)\s+(?:in|for|at|near)\s+([A-Z][a-zA-Z\s]+?)(?:\?|$|,)",
r"(?:in|for|at|near)\s+([A-Z][a-zA-Z\s]+?)(?:\?|$|,)",
r"([A-Z][a-zA-Z\s]+?)\s+(?:weather|temperature|forecast)",
]
for pattern in patterns:
match = re.search(pattern, query)
if match:
city = match.group(1).strip()
# Clean up common words
exclude_words = {"weather", "temperature", "the", "show", "me", "get"}
if city.lower() not in exclude_words:
return city
return None
def extract_location_from_soil_query(query: str, conversation_context: list = None) -> Optional[Dict]:
"""
Extract location from soil-related queries, using conversation context for follow-ups
Args:
query: User's natural language question
conversation_context: Optional list of previous messages for context
Format: [{"role": "user/assistant", "content": "..."}, ...]
Returns:
Dict with location info, or None
{"location": "California"} or {"lat": 40.7, "lon": -74.0}
Examples:
"soil data for Iowa" -> {"location": "Iowa"}
"Show me soil in California" -> {"location": "California"}
Previous: "soil data for ames" (failed), Current: "Ames iowa" -> {"location": "Ames, Iowa"}
"""
# First, try to extract from current query
location_info = None
if nlp is None:
location_info = _extract_location_regex(query)
else:
# Try spaCy NER
doc = nlp(query)
for ent in doc.ents:
if ent.label_ in ["GPE", "LOC"]:
location_info = {"location": ent.text}
break
# Fallback to regex if NER didn't find anything
if not location_info:
location_info = _extract_location_regex(query)
# If location found in current query, return it
if location_info:
return location_info
# If no location in current query, check conversation context
if conversation_context:
# Look for location mentions in previous messages
# Start from most recent and work backwards
for msg in reversed(conversation_context):
content = msg.get("content", "")
if not content:
continue
# Check if previous message was about soil
content_lower = content.lower()
is_soil_related = any(kw in content_lower for kw in ["soil", "dirt", "ground", "earth"])
# Try to extract location from previous message
if nlp is None:
prev_location = _extract_location_regex(content)
else:
doc = nlp(content)
prev_location = None
for ent in doc.ents:
if ent.label_ in ["GPE", "LOC"]:
prev_location = {"location": ent.text}
break
if not prev_location:
prev_location = _extract_location_regex(content)
# If we found a location in a soil-related message, use it
if prev_location and is_soil_related:
# Combine with current query if it has additional location info
# Example: Previous had "Ames", current has "Iowa" -> "Ames, Iowa"
current_location = _extract_location_regex(query)
if current_location:
# Combine locations
prev_loc = prev_location.get("location", "")
curr_loc = current_location.get("location", "")
if prev_loc and curr_loc and prev_loc.lower() != curr_loc.lower():
combined = f"{prev_loc}, {curr_loc}"
return {"location": combined}
elif curr_loc:
return current_location
return prev_location
# Also check if current query is just a location (follow-up)
# Example: Previous: "soil data for ames" (failed), Current: "Ames iowa"
if prev_location and len(query.split()) <= 3:
# Current query looks like a follow-up location
# Try to extract location from current query
current_location = _extract_location_regex(query)
if current_location:
# Combine with previous location
prev_loc = prev_location.get("location", "")
curr_loc = current_location.get("location", "")
if prev_loc and curr_loc:
# Check if they're different (e.g., "Ames" + "Iowa")
if prev_loc.lower() not in curr_loc.lower() and curr_loc.lower() not in prev_loc.lower():
combined = f"{prev_loc}, {curr_loc}"
return {"location": combined}
else:
# One contains the other, use the more complete one
return current_location if len(curr_loc) > len(prev_loc) else prev_location
return current_location
else:
# Current query might be the location itself
# Check if it looks like a location (capitalized words)
words = query.split()
if len(words) <= 3 and all(w[0].isupper() if w else False for w in words):
return {"location": query}
return None
def _extract_location_regex(query: str) -> Optional[Dict]:
"""Extract location using regex patterns"""
patterns = [
r"soil\s+(?:in|for|at|near)\s+([A-Z][a-zA-Z\s]+?)(?:\?|$|,)",
r"(?:in|for|at|near)\s+([A-Z][a-zA-Z\s]+?)(?:\?|$|,)",
r"([A-Z][a-zA-Z\s]+?)\s+soil",
]
for pattern in patterns:
match = re.search(pattern, query)
if match:
location = match.group(1).strip()
return {"location": location}
# Additional pattern: Check if query is just capitalized words (likely a location)
# Example: "Ames iowa", "New York", "Los Angeles"
words = query.strip().split()
if len(words) >= 1 and len(words) <= 3:
# Check if all words start with capital letters (likely a location)
if all(word and word[0].isupper() for word in words):
# Filter out common non-location words
exclude_words = {"The", "A", "An", "And", "Or", "But", "For", "With"}
location_words = [w for w in words if w not in exclude_words]
if location_words:
return {"location": " ".join(location_words)}
return None
def detect_temperature_unit(query: str) -> str:
"""
Detect if user wants Celsius or Fahrenheit
Args:
query: User query
Returns:
"metric" for Celsius, "imperial" for Fahrenheit
"""
query_lower = query.lower()
# Check for explicit mentions
if any(word in query_lower for word in ["fahrenheit", "°f", "f"]):
return "imperial"
# Default to metric (Celsius)
return "metric"
def extract_keywords_from_query(query: str) -> List[str]:
"""
Extract important keywords from query
Args:
query: User query
Returns:
List of keywords
"""
if nlp is None:
# Fallback: simple word splitting
words = re.findall(r'\b\w+\b', query.lower())
# Filter out common stop words
stop_words = {'the', 'a', 'an', 'is', 'are', 'was', 'were', 'in', 'on', 'at', 'to', 'for', 'of', 'and', 'or'}
return [w for w in words if w not in stop_words]
doc = nlp(query)
# Extract meaningful keywords (nouns, verbs, adjectives)
keywords = []
for token in doc:
if token.pos_ in ["NOUN", "VERB", "ADJ", "PROPN"]:
if not token.is_stop and len(token.text) > 2:
keywords.append(token.lemma_.lower())
return keywords
# Test function
if __name__ == "__main__":
print("Testing Parameter Extractor...")
print("-" * 50)
test_queries = [
"What's the weather in London?",
"Show me temperature in New York",
"Is it raining in Tokyo?",
"Weather in Paris in Fahrenheit",
"How's the weather in San Francisco?",
"soil data for California",
"Show me soil in Iowa"
]
for query in test_queries:
print(f"\nQuery: {query}")
# Extract city/location
city = extract_city_from_query(query)
print(f" City: {city}")
# Extract keywords
keywords = extract_keywords_from_query(query)
print(f" Keywords: {keywords}")
# Detect units
units = detect_temperature_unit(query)
print(f" Units: {units}")
|