Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -1,33 +1,74 @@
|
|
1 |
import os
|
2 |
import tempfile
|
|
|
3 |
import streamlit as st
|
4 |
from transformers import pipeline
|
5 |
import docx
|
6 |
import textract
|
7 |
|
|
|
|
|
8 |
#####################################
|
9 |
-
#
|
10 |
#####################################
|
11 |
@st.cache_resource(show_spinner=False)
|
12 |
-
def
|
13 |
try:
|
14 |
-
#
|
15 |
-
|
16 |
-
|
17 |
-
"
|
18 |
-
model="llava-hf/llava-interleave-qwen-0.5b-hf",
|
19 |
trust_remote_code=True
|
20 |
)
|
21 |
-
return
|
22 |
except Exception as e:
|
23 |
-
st.error(f"Error loading
|
24 |
st.stop()
|
25 |
|
26 |
-
|
27 |
-
st.write("
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
28 |
|
29 |
#####################################
|
30 |
-
# Function
|
31 |
#####################################
|
32 |
def extract_text_from_file(file_obj):
|
33 |
"""
|
@@ -38,7 +79,6 @@ def extract_text_from_file(file_obj):
|
|
38 |
text = ""
|
39 |
|
40 |
if ext == ".txt":
|
41 |
-
# For text files, decode the byte stream into a string.
|
42 |
try:
|
43 |
text = file_obj.read().decode("utf-8")
|
44 |
except Exception as e:
|
@@ -46,14 +86,12 @@ def extract_text_from_file(file_obj):
|
|
46 |
|
47 |
elif ext == ".docx":
|
48 |
try:
|
49 |
-
# Use python-docx to read .docx files.
|
50 |
document = docx.Document(file_obj)
|
51 |
text = "\n".join([para.text for para in document.paragraphs])
|
52 |
except Exception as e:
|
53 |
text = f"Error processing DOCX file: {e}"
|
54 |
|
55 |
elif ext == ".doc":
|
56 |
-
# For .doc files, use textract. textract expects a filename, so save temporarily.
|
57 |
try:
|
58 |
with tempfile.NamedTemporaryFile(delete=False, suffix=".doc") as tmp:
|
59 |
tmp.write(file_obj.read())
|
@@ -73,55 +111,54 @@ def extract_text_from_file(file_obj):
|
|
73 |
return text
|
74 |
|
75 |
#####################################
|
76 |
-
# Function
|
77 |
-
#####################################
|
78 |
-
def summarize_text(text):
|
79 |
-
"""
|
80 |
-
Summarize the given text using the summarization pipeline.
|
81 |
-
Adjust max_length and min_length as needed.
|
82 |
-
"""
|
83 |
-
if not text.strip():
|
84 |
-
return "No text available to summarize."
|
85 |
-
|
86 |
-
try:
|
87 |
-
# The summarization pipeline might have limitations on text length.
|
88 |
-
# For long documents, consider splitting the text into smaller chunks.
|
89 |
-
summary = summarizer(text, max_length=150, min_length=40, do_sample=False)
|
90 |
-
return summary[0]["summary_text"]
|
91 |
-
except Exception as e:
|
92 |
-
return f"Error during summarization: {e}"
|
93 |
-
|
94 |
-
#####################################
|
95 |
-
# Main Processing Logic
|
96 |
#####################################
|
97 |
def process_resume(file_obj):
|
98 |
if file_obj is None:
|
99 |
return None, None
|
100 |
|
|
|
101 |
resume_text = extract_text_from_file(file_obj)
|
102 |
-
|
103 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
104 |
|
105 |
#####################################
|
106 |
# Streamlit Interface
|
107 |
#####################################
|
108 |
-
st.title("Resume
|
109 |
st.markdown(
|
110 |
"""
|
111 |
Upload your resume file — supported formats: **.doc**, **.docx**, and **.txt**.
|
112 |
-
The app will extract the text content from your resume
|
|
|
113 |
"""
|
114 |
)
|
115 |
|
116 |
uploaded_file = st.file_uploader("Upload Resume", type=["doc", "docx", "txt"])
|
117 |
|
118 |
-
if st.button("
|
119 |
if uploaded_file is None:
|
120 |
st.error("Please upload a file first.")
|
121 |
else:
|
122 |
with st.spinner("Processing..."):
|
123 |
-
resume_text,
|
124 |
st.subheader("Extracted Resume Text")
|
125 |
st.text_area("", resume_text, height=250)
|
126 |
-
st.subheader("
|
127 |
-
st.text_area("",
|
|
|
1 |
import os
|
2 |
import tempfile
|
3 |
+
import textwrap
|
4 |
import streamlit as st
|
5 |
from transformers import pipeline
|
6 |
import docx
|
7 |
import textract
|
8 |
|
9 |
+
from PIL import Image, ImageDraw, ImageFont
|
10 |
+
|
11 |
#####################################
|
12 |
+
# Model Loading: Image-Text to Text
|
13 |
#####################################
|
14 |
@st.cache_resource(show_spinner=False)
|
15 |
+
def load_image_to_text_pipeline():
|
16 |
try:
|
17 |
+
# Load the image-text to text model.
|
18 |
+
model_pipeline = pipeline(
|
19 |
+
"image-to-text",
|
20 |
+
model="deepseek-ai/deepseek-vl2-tiny",
|
|
|
21 |
trust_remote_code=True
|
22 |
)
|
23 |
+
return model_pipeline
|
24 |
except Exception as e:
|
25 |
+
st.error(f"Error loading image-to-text model: {e}")
|
26 |
st.stop()
|
27 |
|
28 |
+
model_pipeline = load_image_to_text_pipeline()
|
29 |
+
st.write("Image-text to text model loaded successfully!")
|
30 |
+
|
31 |
+
#####################################
|
32 |
+
# Function: Convert Text to an Image
|
33 |
+
#####################################
|
34 |
+
def text_to_image(text, img_width=800, bg_color="white", text_color="black", font_size=20):
|
35 |
+
"""
|
36 |
+
Convert a long text string into a PIL Image.
|
37 |
+
The function wraps text so that it fits within the desired width.
|
38 |
+
"""
|
39 |
+
# Load a default font.
|
40 |
+
try:
|
41 |
+
font = ImageFont.truetype("arial.ttf", font_size)
|
42 |
+
except IOError:
|
43 |
+
# Fallback to default PIL font if arial is not found.
|
44 |
+
font = ImageFont.load_default()
|
45 |
+
|
46 |
+
# Wrap the text into lines.
|
47 |
+
wrapper = textwrap.TextWrapper(width=80)
|
48 |
+
lines = wrapper.wrap(text=text)
|
49 |
+
if not lines:
|
50 |
+
lines = [" "]
|
51 |
+
|
52 |
+
# Calculate the required image height.
|
53 |
+
line_height = font.getsize("A")[1]
|
54 |
+
img_height = line_height * (len(lines) + 2)
|
55 |
+
|
56 |
+
# Create a new image with white background.
|
57 |
+
img = Image.new("RGB", (img_width, img_height), color=bg_color)
|
58 |
+
draw = ImageDraw.Draw(img)
|
59 |
+
|
60 |
+
# Draw each line of text
|
61 |
+
y_text = 10
|
62 |
+
for line in lines:
|
63 |
+
# Center text horizontally.
|
64 |
+
text_width, _ = draw.textsize(line, font=font)
|
65 |
+
x_text = (img_width - text_width) / 2
|
66 |
+
draw.text((x_text, y_text), line, font=font, fill=text_color)
|
67 |
+
y_text += line_height
|
68 |
+
return img
|
69 |
|
70 |
#####################################
|
71 |
+
# Function: Extract Text from File
|
72 |
#####################################
|
73 |
def extract_text_from_file(file_obj):
|
74 |
"""
|
|
|
79 |
text = ""
|
80 |
|
81 |
if ext == ".txt":
|
|
|
82 |
try:
|
83 |
text = file_obj.read().decode("utf-8")
|
84 |
except Exception as e:
|
|
|
86 |
|
87 |
elif ext == ".docx":
|
88 |
try:
|
|
|
89 |
document = docx.Document(file_obj)
|
90 |
text = "\n".join([para.text for para in document.paragraphs])
|
91 |
except Exception as e:
|
92 |
text = f"Error processing DOCX file: {e}"
|
93 |
|
94 |
elif ext == ".doc":
|
|
|
95 |
try:
|
96 |
with tempfile.NamedTemporaryFile(delete=False, suffix=".doc") as tmp:
|
97 |
tmp.write(file_obj.read())
|
|
|
111 |
return text
|
112 |
|
113 |
#####################################
|
114 |
+
# Function: Process Resume Using the Model
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
115 |
#####################################
|
116 |
def process_resume(file_obj):
|
117 |
if file_obj is None:
|
118 |
return None, None
|
119 |
|
120 |
+
# Extract text from file.
|
121 |
resume_text = extract_text_from_file(file_obj)
|
122 |
+
if not resume_text.strip():
|
123 |
+
return resume_text, "No text available to process."
|
124 |
+
|
125 |
+
# Convert the extracted text to an image.
|
126 |
+
text_image = text_to_image(resume_text)
|
127 |
+
|
128 |
+
try:
|
129 |
+
# Pass the generated image to the image-to-text model.
|
130 |
+
result = model_pipeline(text_image)
|
131 |
+
# The expected output is a list of dictionaries with key "generated_text".
|
132 |
+
if isinstance(result, list) and "generated_text" in result[0]:
|
133 |
+
processed_text = result[0]["generated_text"]
|
134 |
+
else:
|
135 |
+
processed_text = "Unexpected model output format."
|
136 |
+
except Exception as e:
|
137 |
+
processed_text = f"Error during model inference: {e}"
|
138 |
+
|
139 |
+
return resume_text, processed_text
|
140 |
|
141 |
#####################################
|
142 |
# Streamlit Interface
|
143 |
#####################################
|
144 |
+
st.title("Resume Processing App")
|
145 |
st.markdown(
|
146 |
"""
|
147 |
Upload your resume file — supported formats: **.doc**, **.docx**, and **.txt**.
|
148 |
+
The app will extract the text content from your resume, convert it to an image,
|
149 |
+
and then use the image-text to text model to process it.
|
150 |
"""
|
151 |
)
|
152 |
|
153 |
uploaded_file = st.file_uploader("Upload Resume", type=["doc", "docx", "txt"])
|
154 |
|
155 |
+
if st.button("Process Resume"):
|
156 |
if uploaded_file is None:
|
157 |
st.error("Please upload a file first.")
|
158 |
else:
|
159 |
with st.spinner("Processing..."):
|
160 |
+
resume_text, processed_text = process_resume(uploaded_file)
|
161 |
st.subheader("Extracted Resume Text")
|
162 |
st.text_area("", resume_text, height=250)
|
163 |
+
st.subheader("Model Output")
|
164 |
+
st.text_area("", processed_text, height=150)
|