File size: 2,904 Bytes
eff9965 |
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 |
import pickle
import json
import numpy as np
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.metrics.pairwise import cosine_similarity
class BuildingMaterialsInference:
def __init__(self):
# Load model data
with open('building_materials_model.pkl', 'rb') as f:
model_data = pickle.load(f)
self.knowledge_base = model_data['knowledge_base']
self.vectorizer = model_data['vectorizer']
self.query_vectors = model_data['query_vectors']
self.queries = model_data['queries']
self.responses = model_data['responses']
def __call__(self, inputs):
"""Hugging Face inference API compatible method"""
if isinstance(inputs, dict):
query = inputs.get('query', inputs.get('inputs', ''))
else:
query = str(inputs)
return self.search_materials(query)
def search_materials(self, user_query, top_k=3):
"""Search for building materials based on user query"""
user_query_lower = user_query.lower()
# Direct match first
if user_query_lower in self.knowledge_base:
response_data = json.loads(self.knowledge_base[user_query_lower])
return {
"results": response_data.get('results', []),
"confidence": 1.0,
"query_matched": user_query
}
# Semantic similarity search
user_vector = self.vectorizer.transform([user_query_lower])
similarities = cosine_similarity(user_vector, self.query_vectors)[0]
# Get top matches
top_indices = np.argsort(similarities)[::-1][:top_k]
# Combine results from top matches
all_results = []
seen_suppliers = set()
for idx in top_indices:
if similarities[idx] > 0.1: # Minimum similarity threshold
response_data = json.loads(self.responses[idx])
for result in response_data.get('results', []):
supplier = result.get('supplier', '')
if supplier not in seen_suppliers:
all_results.append(result)
seen_suppliers.add(supplier)
# Sort by price if available
def extract_price(result):
price_str = result.get('price', '£0')
try:
return float(price_str.replace('£', '').replace(',', ''))
except:
return 999999
all_results.sort(key=extract_price)
return {
"results": all_results[:5], # Return top 5
"confidence": float(max(similarities)) if len(similarities) > 0 else 0.0,
"query_matched": user_query
}
# Create model instance for Hugging Face
model = BuildingMaterialsInference()
|