kawadou
commited on
Commit
•
79a7b9c
1
Parent(s):
b74ef0e
Add application file
Browse files- app.py +14 -4
- requirements.txt +1 -0
app.py
CHANGED
@@ -3,11 +3,21 @@ import fitz # PyMuPDF
|
|
3 |
from sentence_transformers import SentenceTransformer, util
|
4 |
import faiss
|
5 |
from transformers import pipeline
|
|
|
|
|
6 |
|
7 |
st.title("Évaluation Stagiaire Data Scientist")
|
8 |
|
9 |
uploaded_file = st.file_uploader("Choisissez un fichier PDF", type="pdf")
|
10 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
11 |
def extract_text_from_pdf(pdf_path):
|
12 |
text = ""
|
13 |
pdf_document = fitz.open(pdf_path)
|
@@ -55,10 +65,10 @@ def generate_training_plan(scores, threshold=0.7):
|
|
55 |
return plan
|
56 |
|
57 |
if uploaded_file is not None:
|
58 |
-
|
59 |
-
|
60 |
-
|
61 |
-
|
62 |
st.write("Texte extrait du document PDF:")
|
63 |
st.write(document_text[:1000]) # Affiche les 1000 premiers caractères du texte extrait
|
64 |
|
|
|
3 |
from sentence_transformers import SentenceTransformer, util
|
4 |
import faiss
|
5 |
from transformers import pipeline
|
6 |
+
import os
|
7 |
+
from pathlib import Path
|
8 |
|
9 |
st.title("Évaluation Stagiaire Data Scientist")
|
10 |
|
11 |
uploaded_file = st.file_uploader("Choisissez un fichier PDF", type="pdf")
|
12 |
|
13 |
+
def save_uploaded_file(uploaded_file, directory):
|
14 |
+
directory = Path(directory)
|
15 |
+
directory.mkdir(parents=True, exist_ok=True)
|
16 |
+
file_path = directory / uploaded_file.name
|
17 |
+
with open(file_path, "wb") as f:
|
18 |
+
f.write(uploaded_file.getbuffer())
|
19 |
+
return file_path
|
20 |
+
|
21 |
def extract_text_from_pdf(pdf_path):
|
22 |
text = ""
|
23 |
pdf_document = fitz.open(pdf_path)
|
|
|
65 |
return plan
|
66 |
|
67 |
if uploaded_file is not None:
|
68 |
+
file_path = save_uploaded_file(uploaded_file, "uploaded_documents")
|
69 |
+
st.write(f"Fichier téléchargé et sauvegardé sous : {file_path}")
|
70 |
+
|
71 |
+
document_text = extract_text_from_pdf(file_path)
|
72 |
st.write("Texte extrait du document PDF:")
|
73 |
st.write(document_text[:1000]) # Affiche les 1000 premiers caractères du texte extrait
|
74 |
|
requirements.txt
CHANGED
@@ -4,3 +4,4 @@ sentence-transformers
|
|
4 |
faiss-cpu
|
5 |
transformers
|
6 |
torch
|
|
|
|
4 |
faiss-cpu
|
5 |
transformers
|
6 |
torch
|
7 |
+
sentencepiece
|