Spaces:
Runtime error
Runtime error
Commit
·
e198633
1
Parent(s):
20dfb46
Update app.py
Browse files
app.py
CHANGED
@@ -8,6 +8,14 @@ from langchain.embeddings import HuggingFaceEmbeddings, SentenceTransformerEmbed
|
|
8 |
from langchain.vectorstores import FAISS
|
9 |
|
10 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
11 |
|
12 |
def read_pdf(file):
|
13 |
with tempfile.NamedTemporaryFile(delete=False) as temp:
|
@@ -20,24 +28,15 @@ st.title('PDF Text Extractor')
|
|
20 |
|
21 |
uploaded_file = st.file_uploader("Choose a PDF file", type="pdf")
|
22 |
if uploaded_file is not None:
|
23 |
-
raw_documents = read_pdf(uploaded_file)
|
24 |
-
|
25 |
-
|
26 |
-
|
27 |
-
|
28 |
-
)
|
29 |
-
|
30 |
-
|
31 |
-
|
32 |
-
|
33 |
-
|
34 |
-
|
35 |
-
# embeddings_text = embeddings_fun.embed_documents(documents)
|
36 |
-
|
37 |
-
faiss_db = FAISS.from_documents(documents, embeddings_fun)
|
38 |
-
|
39 |
-
query = st.text_input("Entrer une question")
|
40 |
-
docs = faiss_db.similarity_search(query)
|
41 |
-
|
42 |
-
st.text('La reponse à votre question:')
|
43 |
-
st.write(docs[0].page_content)
|
|
|
8 |
from langchain.vectorstores import FAISS
|
9 |
|
10 |
|
11 |
+
splitter = SentenceTransformersTokenTextSplitter(model_name='dangvantuan/sentence-camembert-large',
|
12 |
+
chunk_overlap=50
|
13 |
+
)
|
14 |
+
embeddings_fun = HuggingFaceEmbeddings(model_name='dangvantuan/sentence-camembert-large')
|
15 |
+
|
16 |
+
query = st.text_input("Entrer une question")
|
17 |
+
|
18 |
+
st.text('La reponse à votre question:')
|
19 |
|
20 |
def read_pdf(file):
|
21 |
with tempfile.NamedTemporaryFile(delete=False) as temp:
|
|
|
28 |
|
29 |
uploaded_file = st.file_uploader("Choose a PDF file", type="pdf")
|
30 |
if uploaded_file is not None:
|
31 |
+
raw_documents = read_pdf(uploaded_file)
|
32 |
+
documents = splitter.split_documents(raw_documents)
|
33 |
+
|
34 |
+
# embeddings_text = embeddings_fun.embed_documents(documents)
|
35 |
+
|
36 |
+
faiss_db = FAISS.from_documents(documents, embeddings_fun)
|
37 |
+
|
38 |
+
docs = faiss_db.similarity_search(query)
|
39 |
+
|
40 |
+
st.write(docs[0].page_content)
|
41 |
+
else:
|
42 |
+
st.write("file not uploaded correctly")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|