Technocoloredgeek commited on
Commit
d0a25e9
·
verified ·
1 Parent(s): bb21ab8

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +10 -16
app.py CHANGED
@@ -10,7 +10,7 @@ from langchain_core.runnables import RunnablePassthrough
10
  from qdrant_client import QdrantClient
11
  from qdrant_client.http.models import Distance, VectorParams
12
  from operator import itemgetter
13
- from sentence_transformers import SentenceTransformer
14
 
15
  # Set up API keys
16
  os.environ["OPENAI_API_KEY"] = st.secrets["OPENAI_API_KEY"]
@@ -21,16 +21,6 @@ pdf_links = [
21
  "https://www.whitehouse.gov/wp-content/uploads/2022/10/Blueprint-for-an-AI-Bill-of-Rights.pdf"
22
  ]
23
 
24
- class CustomHuggingFaceEmbeddings:
25
- def __init__(self, model_name):
26
- self.model = SentenceTransformer(model_name)
27
-
28
- def embed_documents(self, texts):
29
- return self.model.encode(texts).tolist()
30
-
31
- def embed_query(self, text):
32
- return self.model.encode(text).tolist()
33
-
34
  @st.cache_resource
35
  def load_and_process_pdfs(pdf_links):
36
  documents = []
@@ -51,22 +41,26 @@ def load_and_process_pdfs(pdf_links):
51
  def setup_vectorstore():
52
  LOCATION = ":memory:"
53
  COLLECTION_NAME = "AI_Ethics_Framework"
54
- VECTOR_SIZE = 768 # Adjust this if your model's output size is different
55
-
56
  qdrant_client = QdrantClient(location=LOCATION)
57
 
 
 
 
 
 
 
58
  # Create the collection
59
  qdrant_client.create_collection(
60
  collection_name=COLLECTION_NAME,
61
  vectors_config=VectorParams(size=VECTOR_SIZE, distance=Distance.COSINE),
62
  )
63
 
64
- # Create the vector store with your custom embeddings
65
- custom_embeddings = CustomHuggingFaceEmbeddings("Technocoloredgeek/midterm-finetuned-embedding")
66
  qdrant_vector_store = QdrantVectorStore(
67
  client=qdrant_client,
68
  collection_name=COLLECTION_NAME,
69
- embedding=custom_embeddings
70
  )
71
 
72
  # Load and add documents
 
10
  from qdrant_client import QdrantClient
11
  from qdrant_client.http.models import Distance, VectorParams
12
  from operator import itemgetter
13
+ from langchain_community.embeddings import HuggingFaceEmbeddings
14
 
15
  # Set up API keys
16
  os.environ["OPENAI_API_KEY"] = st.secrets["OPENAI_API_KEY"]
 
21
  "https://www.whitehouse.gov/wp-content/uploads/2022/10/Blueprint-for-an-AI-Bill-of-Rights.pdf"
22
  ]
23
 
 
 
 
 
 
 
 
 
 
 
24
  @st.cache_resource
25
  def load_and_process_pdfs(pdf_links):
26
  documents = []
 
41
  def setup_vectorstore():
42
  LOCATION = ":memory:"
43
  COLLECTION_NAME = "AI_Ethics_Framework"
44
+
 
45
  qdrant_client = QdrantClient(location=LOCATION)
46
 
47
+ # Create the embeddings
48
+ embeddings = HuggingFaceEmbeddings(model_name="Technocoloredgeek/midterm-finetuned-embedding")
49
+
50
+ # Get the vector size from the embeddings
51
+ VECTOR_SIZE = len(embeddings.embed_query("test"))
52
+
53
  # Create the collection
54
  qdrant_client.create_collection(
55
  collection_name=COLLECTION_NAME,
56
  vectors_config=VectorParams(size=VECTOR_SIZE, distance=Distance.COSINE),
57
  )
58
 
59
+ # Create the vector store
 
60
  qdrant_vector_store = QdrantVectorStore(
61
  client=qdrant_client,
62
  collection_name=COLLECTION_NAME,
63
+ embeddings=embeddings
64
  )
65
 
66
  # Load and add documents