Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -24,5 +24,18 @@ st.write("""
|
|
24 |
""")
|
25 |
|
26 |
st.header('Input')
|
27 |
-
job_desc_pdf()
|
28 |
-
resume_pdf()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
24 |
""")
|
25 |
|
26 |
st.header('Input')
|
27 |
+
jobs_data= job_desc_pdf()
|
28 |
+
resume_df= resume_pdf()
|
29 |
+
|
30 |
+
|
31 |
+
setup_nltk_resources()
|
32 |
+
|
33 |
+
# Unzip wordnet
|
34 |
+
corpora_path = "/kaggle/working/nltk_data/corpora"
|
35 |
+
wordnet_zip = os.path.join(corpora_path, "wordnet.zip")
|
36 |
+
unzip_nltk_resource(wordnet_zip, corpora_path)
|
37 |
+
|
38 |
+
# Apply preprocessing
|
39 |
+
jobs_data['processed_description'] = jobs_data['description'].apply(preprocess_text)
|
40 |
+
jobs_data_cleaned = drop_duplicates(jobs_data, column_name='description')
|
41 |
+
|