Spaces:
Sleeping
Sleeping
Create utils.py
Browse files
utils.py
ADDED
@@ -0,0 +1,135 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import PyPDF2
|
2 |
+
from docx import Document
|
3 |
+
from pptx import Presentation
|
4 |
+
from nlp import get_average_similarity_scores
|
5 |
+
import numpy as np
|
6 |
+
import plotly.graph_objects as go
|
7 |
+
import os
|
8 |
+
import tempfile
|
9 |
+
import shutil
|
10 |
+
|
11 |
+
# Langchain document loaders
|
12 |
+
from langchain.document_loaders import PyPDFLoader #for pdf files
|
13 |
+
from langchain.document_loaders import TextLoader #for text files
|
14 |
+
from langchain.document_loaders import Docx2txtLoader #for docx files
|
15 |
+
from langchain.document_loaders import UnstructuredPowerPointLoader #for pptx files
|
16 |
+
|
17 |
+
from constants import StreamlitException
|
18 |
+
from PyPDF2.errors import PdfReadError
|
19 |
+
from zipfile import BadZipFile
|
20 |
+
|
21 |
+
def load_file(st, uploaded_file):
|
22 |
+
# uploaded_file is the output of st.sidebar.file_uploader
|
23 |
+
file_type = uploaded_file.type
|
24 |
+
try:
|
25 |
+
os.mkdir('downloaded_files')
|
26 |
+
except FileExistsError:
|
27 |
+
pass
|
28 |
+
download_path = os.path.join('downloaded_files', uploaded_file.name)
|
29 |
+
with tempfile.NamedTemporaryFile(delete=False) as tmp_file:
|
30 |
+
# Write the contents of the uploaded file to the temporary file
|
31 |
+
tmp_file.write(uploaded_file.read())
|
32 |
+
tmp_file.flush()
|
33 |
+
shutil.copy(tmp_file.name, download_path)
|
34 |
+
try:
|
35 |
+
if file_type == "application/pdf":
|
36 |
+
resume_text_raw = extract_pdf_text(uploaded_file)
|
37 |
+
lang_loader = PyPDFLoader(download_path)
|
38 |
+
elif file_type == "application/vnd.openxmlformats-officedocument.wordprocessingml.document":
|
39 |
+
resume_text_raw = extract_word_text(uploaded_file)
|
40 |
+
lang_loader = Docx2txtLoader(download_path)
|
41 |
+
elif file_type == "application/vnd.ms-powerpoint" or file_type == "application/vnd.openxmlformats-officedocument.presentationml.presentation":
|
42 |
+
resume_text_raw = extract_ppt_text(uploaded_file)
|
43 |
+
lang_loader = UnstructuredPowerPointLoader(download_path)
|
44 |
+
else:
|
45 |
+
return StreamlitException("**Error**: Invalid file format. Please upload a Word, PDF, or PowerPoint file.")
|
46 |
+
except (PdfReadError, BadZipFile):
|
47 |
+
return StreamlitException("**Error**: Invalid file content. Please upload a valid Word, PDF, or PowerPoint file.")
|
48 |
+
|
49 |
+
return resume_text_raw, lang_loader
|
50 |
+
|
51 |
+
|
52 |
+
# Function to extract text from a PDF file
|
53 |
+
def extract_pdf_text(file):
|
54 |
+
pdf_reader = PyPDF2.PdfReader(file)
|
55 |
+
text = ""
|
56 |
+
for page in pdf_reader.pages:
|
57 |
+
lines = page.extract_text().split('\n')
|
58 |
+
for line in lines:
|
59 |
+
text += line.strip() + ".\n"
|
60 |
+
return text
|
61 |
+
|
62 |
+
|
63 |
+
# Function to extract text from a Word file
|
64 |
+
def extract_word_text(file):
|
65 |
+
doc = Document(file)
|
66 |
+
text = ''
|
67 |
+
p_iter = iter(doc.paragraphs)
|
68 |
+
t_iter = iter(doc.tables)
|
69 |
+
while True:
|
70 |
+
try:
|
71 |
+
paragraph = next(p_iter)
|
72 |
+
text += paragraph.text + '.\n'
|
73 |
+
except StopIteration:
|
74 |
+
break
|
75 |
+
try:
|
76 |
+
table = next(t_iter)
|
77 |
+
for row in table.rows:
|
78 |
+
for cell in row.cells:
|
79 |
+
text += cell.text + '\t'
|
80 |
+
text += '\n'
|
81 |
+
except StopIteration:
|
82 |
+
pass
|
83 |
+
return text
|
84 |
+
|
85 |
+
|
86 |
+
# Function to extract text from a PowerPoint file
|
87 |
+
def extract_ppt_text(file):
|
88 |
+
prs = Presentation(file)
|
89 |
+
text = ""
|
90 |
+
for slide in prs.slides:
|
91 |
+
for shape in slide.shapes:
|
92 |
+
if shape.has_text_frame:
|
93 |
+
text += shape.text_frame.text
|
94 |
+
return text
|
95 |
+
|
96 |
+
# Function to plot the average similarity score for each job description phrase
|
97 |
+
def plot_similarity_scores(job_description_phrases, resume_phrases):
|
98 |
+
avg_similarity_scores = get_average_similarity_scores(job_description_phrases, resume_phrases)
|
99 |
+
sorted_scores = sorted(enumerate(avg_similarity_scores), key=lambda x: x[1], reverse=True)[:10]
|
100 |
+
indices = [i[0] for i in sorted_scores]
|
101 |
+
sorted_scores = [i[1] for i in sorted_scores]
|
102 |
+
|
103 |
+
y_pos = list(range(len(indices)))
|
104 |
+
|
105 |
+
fig = go.Figure()
|
106 |
+
fig.add_trace(go.Bar(
|
107 |
+
y=y_pos,
|
108 |
+
x=sorted_scores,
|
109 |
+
orientation='h'
|
110 |
+
))
|
111 |
+
|
112 |
+
fig.update_layout(
|
113 |
+
yaxis=dict(
|
114 |
+
tickmode="array",
|
115 |
+
tickvals=y_pos,
|
116 |
+
ticktext=[s[:100].ljust(100) + '...' if len(s) > 100 else s.ljust(75) for s in np.array(job_description_phrases)[indices]],
|
117 |
+
tickfont=dict(size=14),
|
118 |
+
autorange="reversed",
|
119 |
+
side="right",
|
120 |
+
automargin=True
|
121 |
+
),
|
122 |
+
xaxis=dict(
|
123 |
+
tickmode="array",
|
124 |
+
tickvals=np.round(np.arange(0, 1.2, 0.2), 2),
|
125 |
+
ticktext=np.round(np.arange(0, 1.2, 0.2), 2),
|
126 |
+
tickfont=dict(size=14),
|
127 |
+
range=[0, 1.05]
|
128 |
+
),
|
129 |
+
showlegend=False,
|
130 |
+
margin=dict(t=0)
|
131 |
+
)
|
132 |
+
|
133 |
+
fig.update_xaxes(title="Average Similarity Score", title_font=dict(size=14))
|
134 |
+
|
135 |
+
return fig
|