TomData commited on
Commit
e7ac557
·
1 Parent(s): 963621e

revert keywordearch to static

Browse files
Files changed (1) hide show
  1. src/chatbot.py +2 -5
src/chatbot.py CHANGED
@@ -1,7 +1,6 @@
1
  from langchain_core.prompts import ChatPromptTemplate
2
  from langchain_community.llms.huggingface_hub import HuggingFaceHub
3
  from langchain_community.embeddings import HuggingFaceEmbeddings
4
- from langchain_community.vectorstores import FAISS
5
 
6
  from src.vectordatabase import RAG, get_vectorstore
7
  import pandas as pd
@@ -12,8 +11,6 @@ import pandas as pd
12
 
13
  # Define important variables
14
  embeddings = HuggingFaceEmbeddings(model_name="paraphrase-multilingual-MiniLM-L12-v2") # Remove embedding input parameter from functions?
15
- db_all = FAISS.load_local(folder_path="./src/FAISS", index_name="speeches_1949_09_12",
16
- embeddings=embeddings, allow_dangerous_deserialization=True)
17
  llm = HuggingFaceHub(
18
  # ToDo: Try different models here
19
  repo_id="mistralai/Mixtral-8x7B-Instruct-v0.1",
@@ -112,7 +109,7 @@ def chatbot(message, history, db_inputs, prompt_language, llm=llm):
112
  return response
113
 
114
 
115
- def keyword_search(query, db=db_all, n=10, embeddings=embeddings, method='ss', party_filter='All'):
116
  """
117
  Retrieve speech contents based on keywords using a specified method.
118
 
@@ -154,7 +151,7 @@ def keyword_search(query, db=db_all, n=10, embeddings=embeddings, method='ss', p
154
  - The `party_filter` parameter is a filter for retrieving speech contents by party affiliation. Specify 'All' to retrieve
155
  speeches from all parties (default is 'All').
156
  """
157
-
158
  query_embedding = embeddings.embed_query(query)
159
 
160
  # Maximal Marginal Relevance
 
1
  from langchain_core.prompts import ChatPromptTemplate
2
  from langchain_community.llms.huggingface_hub import HuggingFaceHub
3
  from langchain_community.embeddings import HuggingFaceEmbeddings
 
4
 
5
  from src.vectordatabase import RAG, get_vectorstore
6
  import pandas as pd
 
11
 
12
  # Define important variables
13
  embeddings = HuggingFaceEmbeddings(model_name="paraphrase-multilingual-MiniLM-L12-v2") # Remove embedding input parameter from functions?
 
 
14
  llm = HuggingFaceHub(
15
  # ToDo: Try different models here
16
  repo_id="mistralai/Mixtral-8x7B-Instruct-v0.1",
 
109
  return response
110
 
111
 
112
+ def keyword_search(query, n=10, embeddings=embeddings, method='ss', party_filter='All'):
113
  """
114
  Retrieve speech contents based on keywords using a specified method.
115
 
 
151
  - The `party_filter` parameter is a filter for retrieving speech contents by party affiliation. Specify 'All' to retrieve
152
  speeches from all parties (default is 'All').
153
  """
154
+ db = get_vectorstore(inputs="All", embeddings=embeddings)
155
  query_embedding = embeddings.embed_query(query)
156
 
157
  # Maximal Marginal Relevance