File size: 2,885 Bytes
c57a856 db131d7 df11f29 694dd34 c55d69f df11f29 c55d69f 72bb033 c55d69f 0991258 c55d69f 0991258 694dd34 0991258 7ff91e3 52cc43b 7ff91e3 17eaeb4 694dd34 |
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 |
import streamlit as st
import tl_calamancy_lg
import os
import pandas as pd
import json
from sklearn.metrics.pairwise import cosine_similarity
# Define the cache decorator for loading the spaCy model
@st.cache_resource()
def load_nlp_model():
return tl_calamancy_lg.load()
# Load the spaCy model using the cached function
nlp = load_nlp_model()
# Define the cache decorator for loading the DataFrame
@st.cache_data
def load_data(file_path):
# Open the JSON file
with open(file_path, 'r') as file:
# Load the JSON data
data = json.load(file)
# Extract patterns and responses into separate lists
patterns_data = []
responses_data = []
for intent in data["intents"]:
tag = intent["tag"]
patterns = intent.get("patterns", [])
responses = intent.get("responses", [])
for pattern in patterns:
patterns_data.append({"tag": tag, "pattern": pattern})
for response in responses:
responses_data.append({"tag": tag, "response": response})
# Create and return DataFrames
patterns_df = pd.DataFrame(patterns_data)
responses_df = pd.DataFrame(responses_data)
return patterns_df, responses_df
# Get the absolute path of the script directory
cwd = os.getcwd()
# Read the CSV file
file_path = os.path.join(cwd, "dataset_v2.json")
# Load the DataFrames using the cached function
patterns_df, responses_df = load_data(file_path)
# Define the cache decorator for the similarity function
@st.cache_data
def get_most_similar_tag(user_query, dataframe):
# Process user query and existing queries with spaCy
all_queries = list(dataframe['pattern']) + [user_query]
processed_queries = [nlp(query) for query in all_queries]
# Get word vectors for each query
vectors = [query.vector for query in processed_queries]
# Calculate cosine similarity
similarity_matrix = cosine_similarity(vectors, vectors)
# Extract similarity scores for the user query
user_similarity_scores = similarity_matrix[-1, :-1]
# Find the index of the tag with the highest similarity score
most_similar_index = user_similarity_scores.argmax()
# Get the most similar tag
most_similar_tag = dataframe['tag'].iloc[most_similar_index]
# Return the most similar tag and its similarity score
return most_similar_tag, user_similarity_scores[most_similar_index]
def main():
# StreamLit Title
st.title("TagaCare")
# React to user input
if prompt := st.chat_input("Magtanong ng lunas sa sakit"):
# Use the cached function to get the most similar tag
returned_tag, returned_score = get_most_similar_tag(prompt, patterns_df)
st.success(returned_tag + str(returned_score))
st.success(responses_df[responses_df['tag']==returned_tag]['response'])
if __name__ == "__main__":
main()
|