Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -4,43 +4,29 @@ import streamlit as st
|
|
4 |
import docx
|
5 |
import textract
|
6 |
from transformers import pipeline
|
7 |
-
|
8 |
-
|
|
|
9 |
|
10 |
#####################################
|
11 |
-
#
|
12 |
#####################################
|
13 |
-
@st.cache_resource(show_spinner=
|
14 |
def load_models():
|
15 |
-
"""
|
16 |
-
|
17 |
-
|
18 |
-
|
19 |
-
|
20 |
-
|
21 |
-
models['
|
22 |
-
|
23 |
-
|
24 |
-
|
25 |
-
|
26 |
-
device=0 if st.session_state.get('use_gpu', False) else -1)
|
27 |
-
|
28 |
-
# Start threads to load models in parallel
|
29 |
-
threads = [
|
30 |
-
threading.Thread(target=load_summarizer_thread),
|
31 |
-
threading.Thread(target=load_similarity_thread)
|
32 |
-
]
|
33 |
-
|
34 |
-
for thread in threads:
|
35 |
-
thread.start()
|
36 |
-
|
37 |
-
for thread in threads:
|
38 |
-
thread.join()
|
39 |
-
|
40 |
-
return models
|
41 |
|
42 |
#####################################
|
43 |
-
# Function: Extract Text from File
|
44 |
#####################################
|
45 |
def extract_text_from_file(file_obj):
|
46 |
"""
|
@@ -54,18 +40,15 @@ def extract_text_from_file(file_obj):
|
|
54 |
if ext == ".docx":
|
55 |
try:
|
56 |
document = docx.Document(file_obj)
|
57 |
-
# Use a list comprehension and join for better performance
|
58 |
text = "\n".join(para.text for para in document.paragraphs if para.text.strip())
|
59 |
except Exception as e:
|
60 |
text = f"Error processing DOCX file: {e}"
|
61 |
elif ext == ".doc":
|
62 |
try:
|
63 |
-
# Use a context manager for better file handling
|
64 |
with tempfile.NamedTemporaryFile(delete=False, suffix=".doc") as tmp:
|
65 |
tmp.write(file_obj.read())
|
66 |
tmp_filename = tmp.name
|
67 |
text = textract.process(tmp_filename).decode("utf-8")
|
68 |
-
# Clean up the temporary file immediately
|
69 |
os.unlink(tmp_filename)
|
70 |
except Exception as e:
|
71 |
text = f"Error processing DOC file: {e}"
|
@@ -74,20 +57,19 @@ def extract_text_from_file(file_obj):
|
|
74 |
return text
|
75 |
|
76 |
#####################################
|
77 |
-
# Function: Summarize Resume Text
|
78 |
#####################################
|
79 |
def summarize_resume_text(resume_text, models):
|
80 |
"""
|
81 |
-
Generates a concise summary of the resume text using the
|
82 |
"""
|
83 |
summarizer = models['summarizer']
|
84 |
-
|
85 |
-
#
|
86 |
-
# Break text into chunks and summarize important parts
|
87 |
max_input_length = 1024 # PEGASUS-XSUM limit
|
88 |
|
89 |
if len(resume_text) > max_input_length:
|
90 |
-
#
|
91 |
chunks = [resume_text[i:i+max_input_length] for i in range(0, min(len(resume_text), 3*max_input_length), max_input_length)]
|
92 |
summaries = []
|
93 |
|
@@ -96,7 +78,6 @@ def summarize_resume_text(resume_text, models):
|
|
96 |
summaries.append(chunk_summary)
|
97 |
|
98 |
candidate_summary = " ".join(summaries)
|
99 |
-
# Summarize again if combined summary is too long
|
100 |
if len(candidate_summary) > max_input_length:
|
101 |
candidate_summary = summarizer(candidate_summary[:max_input_length], max_length=150, min_length=40, do_sample=False)[0]['summary_text']
|
102 |
else:
|
@@ -105,11 +86,11 @@ def summarize_resume_text(resume_text, models):
|
|
105 |
return candidate_summary
|
106 |
|
107 |
#####################################
|
108 |
-
# Function: Compare Candidate Summary to Company Prompt
|
109 |
#####################################
|
110 |
def compute_suitability(candidate_summary, company_prompt, models):
|
111 |
"""
|
112 |
-
Compute the similarity between candidate summary and company prompt
|
113 |
Returns a score in the range [0, 1].
|
114 |
"""
|
115 |
similarity_pipeline = models['similarity']
|
@@ -125,130 +106,92 @@ def compute_suitability(candidate_summary, company_prompt, models):
|
|
125 |
return score
|
126 |
|
127 |
#####################################
|
128 |
-
#
|
129 |
#####################################
|
130 |
-
|
131 |
-
|
132 |
-
Extracts text from the uploaded file and then generates a summary
|
133 |
-
using a text summarization model.
|
134 |
"""
|
135 |
-
|
136 |
-
|
137 |
-
|
138 |
-
|
139 |
-
|
140 |
-
|
141 |
-
status.update(label="Error: No text could be extracted", state="error")
|
142 |
-
return ""
|
143 |
-
|
144 |
-
status.update(label=f"Extracted {len(resume_text)} characters. Generating summary...")
|
145 |
-
|
146 |
-
candidate_summary = summarize_resume_text(resume_text, models)
|
147 |
-
status.update(label="Processing complete!", state="complete")
|
148 |
-
|
149 |
-
return candidate_summary
|
150 |
|
151 |
-
|
152 |
-
|
153 |
-
|
154 |
-
|
155 |
-
|
|
|
156 |
|
157 |
-
|
158 |
-
|
159 |
-
st.session_state.use_gpu = False
|
160 |
|
161 |
-
#
|
162 |
-
|
163 |
-
|
164 |
-
|
165 |
-
st.session_state.use_gpu = True
|
166 |
else:
|
167 |
-
st.
|
168 |
-
|
169 |
-
|
170 |
-
|
171 |
-
|
172 |
-
|
173 |
-
|
174 |
-
|
175 |
-
|
176 |
-
st.markdown(
|
177 |
-
"""
|
178 |
-
Upload your resume file in **.doc** or **.docx** format. The app performs the following tasks:
|
179 |
-
1. Extracts text from the resume.
|
180 |
-
2. Uses a transformer-based model to generate a concise candidate summary.
|
181 |
-
3. Compares the candidate summary with a company profile to produce a suitability score.
|
182 |
-
"""
|
183 |
-
)
|
184 |
-
|
185 |
-
# Use columns for better layout
|
186 |
-
col1, col2 = st.columns([1, 1])
|
187 |
-
|
188 |
-
with col1:
|
189 |
-
# File uploader for resume
|
190 |
-
uploaded_file = st.file_uploader("Upload Resume", type=["doc", "docx"])
|
191 |
-
|
192 |
-
# Button to process the resume
|
193 |
-
if st.button("Process Resume", type="primary", use_container_width=True):
|
194 |
-
if uploaded_file is None:
|
195 |
-
st.error("Please upload a resume file first.")
|
196 |
-
else:
|
197 |
-
candidate_summary = process_resume(uploaded_file, models)
|
198 |
-
if candidate_summary: # only if summary is generated
|
199 |
st.session_state["candidate_summary"] = candidate_summary
|
200 |
-
|
201 |
-
# Display candidate summary if available
|
202 |
-
if "candidate_summary" in st.session_state:
|
203 |
-
st.subheader("Candidate Summary")
|
204 |
-
st.markdown(st.session_state["candidate_summary"])
|
205 |
|
206 |
-
|
207 |
-
|
208 |
-
|
209 |
-
|
210 |
-
"artificial intelligence, and software development. As part of Alphabet Inc., Google seeks candidates with strong "
|
211 |
-
"problem-solving skills, adaptability, and collaboration abilities. Technical roles require proficiency in programming "
|
212 |
-
"languages such as Python, Java, C++, Go, or JavaScript, with expertise in data structures, algorithms, and system design. "
|
213 |
-
"Additionally, skills in AI, cybersecurity, UX/UI design, and digital marketing are highly valued. Google fosters a culture "
|
214 |
-
"of innovation, expecting candidates to demonstrate creativity, analytical thinking, and a passion for cutting-edge technology."
|
215 |
-
)
|
216 |
|
217 |
-
|
218 |
-
|
219 |
-
|
220 |
-
|
221 |
-
|
222 |
-
|
|
|
|
|
|
|
|
|
223 |
|
224 |
-
|
225 |
-
|
226 |
-
|
227 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
228 |
else:
|
229 |
-
|
230 |
-
|
231 |
-
|
232 |
-
|
233 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
234 |
else:
|
235 |
-
|
236 |
-
score = compute_suitability(candidate_summary, company_prompt, models)
|
237 |
-
|
238 |
-
# Display score with a progress bar for visual feedback
|
239 |
-
st.success(f"Suitability Score: {score:.2f} (range 0 to 1)")
|
240 |
-
st.progress(score)
|
241 |
-
|
242 |
-
# Add interpretation of score
|
243 |
-
if score > 0.75:
|
244 |
-
st.info("Excellent match! Your profile appears very well suited for this company.")
|
245 |
-
elif score > 0.5:
|
246 |
-
st.info("Good match. Your profile aligns with many aspects of the company's requirements.")
|
247 |
-
elif score > 0.3:
|
248 |
-
st.info("Moderate match. Consider highlighting more relevant skills or experience.")
|
249 |
-
else:
|
250 |
-
st.info("Low match. Your profile may need significant adjustments to better align with this company.")
|
251 |
-
|
252 |
-
|
253 |
-
if __name__ == "__main__":
|
254 |
-
main()
|
|
|
4 |
import docx
|
5 |
import textract
|
6 |
from transformers import pipeline
|
7 |
+
|
8 |
+
# Set page title
|
9 |
+
st.set_page_config(page_title="Resume Analyzer and Company Suitability Checker")
|
10 |
|
11 |
#####################################
|
12 |
+
# Preload Models
|
13 |
#####################################
|
14 |
+
@st.cache_resource(show_spinner=True)
|
15 |
def load_models():
|
16 |
+
"""Load all models at startup"""
|
17 |
+
with st.spinner("Loading AI models... This may take a minute on first run."):
|
18 |
+
models = {}
|
19 |
+
# Load summarization model
|
20 |
+
models['summarizer'] = pipeline("summarization", model="google/pegasus-xsum")
|
21 |
+
# Load similarity model
|
22 |
+
models['similarity'] = pipeline("sentence-similarity", model="sentence-transformers/all-MiniLM-L6-v2")
|
23 |
+
return models
|
24 |
+
|
25 |
+
# Preload models immediately when app starts
|
26 |
+
models = load_models()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
27 |
|
28 |
#####################################
|
29 |
+
# Function: Extract Text from File
|
30 |
#####################################
|
31 |
def extract_text_from_file(file_obj):
|
32 |
"""
|
|
|
40 |
if ext == ".docx":
|
41 |
try:
|
42 |
document = docx.Document(file_obj)
|
|
|
43 |
text = "\n".join(para.text for para in document.paragraphs if para.text.strip())
|
44 |
except Exception as e:
|
45 |
text = f"Error processing DOCX file: {e}"
|
46 |
elif ext == ".doc":
|
47 |
try:
|
|
|
48 |
with tempfile.NamedTemporaryFile(delete=False, suffix=".doc") as tmp:
|
49 |
tmp.write(file_obj.read())
|
50 |
tmp_filename = tmp.name
|
51 |
text = textract.process(tmp_filename).decode("utf-8")
|
|
|
52 |
os.unlink(tmp_filename)
|
53 |
except Exception as e:
|
54 |
text = f"Error processing DOC file: {e}"
|
|
|
57 |
return text
|
58 |
|
59 |
#####################################
|
60 |
+
# Function: Summarize Resume Text
|
61 |
#####################################
|
62 |
def summarize_resume_text(resume_text, models):
|
63 |
"""
|
64 |
+
Generates a concise summary of the resume text using the summarization model.
|
65 |
"""
|
66 |
summarizer = models['summarizer']
|
67 |
+
|
68 |
+
# Handle long text
|
|
|
69 |
max_input_length = 1024 # PEGASUS-XSUM limit
|
70 |
|
71 |
if len(resume_text) > max_input_length:
|
72 |
+
# Process in chunks if text is too long
|
73 |
chunks = [resume_text[i:i+max_input_length] for i in range(0, min(len(resume_text), 3*max_input_length), max_input_length)]
|
74 |
summaries = []
|
75 |
|
|
|
78 |
summaries.append(chunk_summary)
|
79 |
|
80 |
candidate_summary = " ".join(summaries)
|
|
|
81 |
if len(candidate_summary) > max_input_length:
|
82 |
candidate_summary = summarizer(candidate_summary[:max_input_length], max_length=150, min_length=40, do_sample=False)[0]['summary_text']
|
83 |
else:
|
|
|
86 |
return candidate_summary
|
87 |
|
88 |
#####################################
|
89 |
+
# Function: Compare Candidate Summary to Company Prompt
|
90 |
#####################################
|
91 |
def compute_suitability(candidate_summary, company_prompt, models):
|
92 |
"""
|
93 |
+
Compute the similarity between candidate summary and company prompt.
|
94 |
Returns a score in the range [0, 1].
|
95 |
"""
|
96 |
similarity_pipeline = models['similarity']
|
|
|
106 |
return score
|
107 |
|
108 |
#####################################
|
109 |
+
# Streamlit Interface
|
110 |
#####################################
|
111 |
+
st.title("Resume Analyzer and Company Suitability Checker")
|
112 |
+
st.markdown(
|
|
|
|
|
113 |
"""
|
114 |
+
Upload your resume file in **.doc** or **.docx** format. The app performs the following tasks:
|
115 |
+
1. Extracts text from the resume.
|
116 |
+
2. Uses a transformer-based model to generate a concise candidate summary.
|
117 |
+
3. Compares the candidate summary with a company profile to produce a suitability score.
|
118 |
+
"""
|
119 |
+
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
120 |
|
121 |
+
# Use two columns with equal width
|
122 |
+
col1, col2 = st.columns(2)
|
123 |
+
|
124 |
+
with col1:
|
125 |
+
# File uploader for resume
|
126 |
+
uploaded_file = st.file_uploader("Upload Resume", type=["doc", "docx"])
|
127 |
|
128 |
+
if uploaded_file is not None:
|
129 |
+
st.write(f"{uploaded_file.name} {uploaded_file.size/1024:.1f}KB")
|
|
|
130 |
|
131 |
+
# Button to process the resume
|
132 |
+
if st.button("Process Resume", type="primary", use_container_width=True):
|
133 |
+
if uploaded_file is None:
|
134 |
+
st.error("Please upload a resume file first.")
|
|
|
135 |
else:
|
136 |
+
with st.status("Processing resume...") as status:
|
137 |
+
status.update(label="Extracting text from resume...")
|
138 |
+
resume_text = extract_text_from_file(uploaded_file)
|
139 |
+
|
140 |
+
if not resume_text or resume_text.strip() == "":
|
141 |
+
status.update(label="Error: No text could be extracted", state="error")
|
142 |
+
else:
|
143 |
+
status.update(label=f"Extracted {len(resume_text)} characters. Generating summary...")
|
144 |
+
candidate_summary = summarize_resume_text(resume_text, models)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
145 |
st.session_state["candidate_summary"] = candidate_summary
|
146 |
+
status.update(label="Processing complete!", state="complete")
|
|
|
|
|
|
|
|
|
147 |
|
148 |
+
# Display candidate summary if available
|
149 |
+
if "candidate_summary" in st.session_state:
|
150 |
+
st.subheader("Candidate Summary")
|
151 |
+
st.markdown(st.session_state["candidate_summary"])
|
|
|
|
|
|
|
|
|
|
|
|
|
152 |
|
153 |
+
with col2:
|
154 |
+
# Pre-defined company prompt for Google LLC.
|
155 |
+
default_company_prompt = (
|
156 |
+
"Google LLC, a global leader in technology and innovation, specializes in internet services, cloud computing, "
|
157 |
+
"artificial intelligence, and software development. As part of Alphabet Inc., Google seeks candidates with strong "
|
158 |
+
"problem-solving skills, adaptability, and collaboration abilities. Technical roles require proficiency in programming "
|
159 |
+
"languages such as Python, Java, C++, Go, or JavaScript, with expertise in data structures, algorithms, and system design. "
|
160 |
+
"Additionally, skills in AI, cybersecurity, UX/UI design, and digital marketing are highly valued. Google fosters a culture "
|
161 |
+
"of innovation, expecting candidates to demonstrate creativity, analytical thinking, and a passion for cutting-edge technology."
|
162 |
+
)
|
163 |
|
164 |
+
# Company prompt text area.
|
165 |
+
company_prompt = st.text_area(
|
166 |
+
"Enter company details:",
|
167 |
+
value=default_company_prompt,
|
168 |
+
height=150,
|
169 |
+
)
|
170 |
+
|
171 |
+
# Button to compute the suitability score.
|
172 |
+
if st.button("Compute Suitability Score", type="primary", use_container_width=True):
|
173 |
+
if "candidate_summary" not in st.session_state:
|
174 |
+
st.error("Please process the resume first!")
|
175 |
+
else:
|
176 |
+
candidate_summary = st.session_state["candidate_summary"]
|
177 |
+
if candidate_summary.strip() == "":
|
178 |
+
st.error("Candidate summary is empty; please check your resume file.")
|
179 |
+
elif company_prompt.strip() == "":
|
180 |
+
st.error("Please enter the company information.")
|
181 |
else:
|
182 |
+
with st.spinner("Computing suitability score..."):
|
183 |
+
score = compute_suitability(candidate_summary, company_prompt, models)
|
184 |
+
|
185 |
+
# Display score with a progress bar for visual feedback
|
186 |
+
st.success(f"Suitability Score: {score:.2f} (range 0 to 1)")
|
187 |
+
st.progress(score)
|
188 |
+
|
189 |
+
# Add interpretation of score
|
190 |
+
if score > 0.75:
|
191 |
+
st.info("Excellent match! Your profile appears very well suited for this company.")
|
192 |
+
elif score > 0.5:
|
193 |
+
st.info("Good match. Your profile aligns with many aspects of the company's requirements.")
|
194 |
+
elif score > 0.3:
|
195 |
+
st.info("Moderate match. Consider highlighting more relevant skills or experience.")
|
196 |
else:
|
197 |
+
st.info("Low match. Your profile may need significant adjustments to better align with this company.")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|