Spaces:
Sleeping
Sleeping
Update cv_quality.py
Browse files- cv_quality.py +33 -12
cv_quality.py
CHANGED
@@ -1,31 +1,52 @@
|
|
1 |
import os
|
2 |
-
|
|
|
|
|
3 |
from dotenv import load_dotenv
|
4 |
from yolo_text_extraction import pdf_to_text
|
5 |
|
6 |
load_dotenv()
|
7 |
|
8 |
class CV:
|
9 |
-
def __init__(self,
|
10 |
-
self.
|
11 |
self.doc_loader = {
|
12 |
-
".
|
13 |
-
".
|
14 |
-
".
|
15 |
}
|
16 |
|
17 |
def get_cv_text(self):
|
18 |
-
|
19 |
-
|
20 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
21 |
data = loader.load()
|
|
|
|
|
|
|
|
|
22 |
if data:
|
23 |
text = "\n".join([page.page_content for page in data])
|
24 |
-
return text if text.strip() else pdf_to_text(
|
25 |
else:
|
26 |
-
return pdf_to_text(
|
27 |
else:
|
28 |
-
return pdf_to_text(
|
29 |
|
30 |
def analyse_cv_quality(self):
|
31 |
from cv_analyzer import analyze_cv
|
|
|
1 |
import os
|
2 |
+
import requests
|
3 |
+
from io import BytesIO
|
4 |
+
from langchain_community.document_loaders import PyPDFLoader, Docx2txtLoader, UnstructuredFileLoader
|
5 |
from dotenv import load_dotenv
|
6 |
from yolo_text_extraction import pdf_to_text
|
7 |
|
8 |
load_dotenv()
|
9 |
|
10 |
class CV:
|
11 |
+
def __init__(self, file_url):
|
12 |
+
self.file_url = file_url
|
13 |
self.doc_loader = {
|
14 |
+
".pdf": PyPDFLoader,
|
15 |
+
".docx": Docx2txtLoader,
|
16 |
+
".txt": UnstructuredFileLoader
|
17 |
}
|
18 |
|
19 |
def get_cv_text(self):
|
20 |
+
# Download the file from Supabase
|
21 |
+
response = requests.get(self.file_url)
|
22 |
+
file_content = BytesIO(response.content)
|
23 |
+
|
24 |
+
# Determine file extension
|
25 |
+
_, ext = os.path.splitext(self.file_url.split("?")[0]) # Remove query parameters
|
26 |
+
|
27 |
+
if ext.lower() in self.doc_loader:
|
28 |
+
if ext.lower() == '.pdf':
|
29 |
+
loader = self.doc_loader[ext.lower()](file_content)
|
30 |
+
else:
|
31 |
+
# For non-PDF files, save temporarily and use the appropriate loader
|
32 |
+
with tempfile.NamedTemporaryFile(delete=False, suffix=ext) as temp_file:
|
33 |
+
temp_file.write(file_content.getvalue())
|
34 |
+
temp_file_path = temp_file.name
|
35 |
+
|
36 |
+
loader = self.doc_loader[ext.lower()](temp_file_path)
|
37 |
+
|
38 |
data = loader.load()
|
39 |
+
|
40 |
+
if ext.lower() != '.pdf':
|
41 |
+
os.unlink(temp_file_path) # Delete the temporary file
|
42 |
+
|
43 |
if data:
|
44 |
text = "\n".join([page.page_content for page in data])
|
45 |
+
return text if text.strip() else pdf_to_text(file_content)
|
46 |
else:
|
47 |
+
return pdf_to_text(file_content)
|
48 |
else:
|
49 |
+
return pdf_to_text(file_content)
|
50 |
|
51 |
def analyse_cv_quality(self):
|
52 |
from cv_analyzer import analyze_cv
|