root commited on
Commit
72d33a9
·
1 Parent(s): eaa3094
Files changed (2) hide show
  1. app.py +384 -570
  2. requirements.txt +18 -20
app.py CHANGED
@@ -1,89 +1,131 @@
1
  import streamlit as st
2
  import pdfplumber
3
- import io
4
- import spacy
5
- from transformers import pipeline, AutoModelForCausalLM, AutoTokenizer
6
- import subprocess
7
- import sys
8
- import torch
9
  import re
10
  import pandas as pd
11
- import numpy as np
 
 
12
  import plotly.express as px
13
  import plotly.graph_objects as go
14
- from datetime import datetime
15
- import dateparser
16
- from sentence_transformers import SentenceTransformer
17
- import nltk
18
- from nltk.tokenize import word_tokenize
19
- from nltk.corpus import stopwords
20
- from sklearn.metrics.pairwise import cosine_similarity
21
- import faiss
22
- import requests
23
- from bs4 import BeautifulSoup
24
- import networkx as nx
25
- import Levenshtein
26
- import json
27
- import matplotlib.pyplot as plt
28
- from io import BytesIO
29
- import base64
30
- from sentence_transformers import util
31
-
32
- # Download NLTK resources
33
- @st.cache_resource
34
- def download_nltk_resources():
35
- nltk.download('punkt')
36
- nltk.download('stopwords')
37
- nltk.download('wordnet')
38
- nltk.download('averaged_perceptron_tagger')
39
-
40
- download_nltk_resources()
41
 
 
42
  st.set_page_config(
43
  page_title="Resume Screener & Skill Extractor",
44
  page_icon="📄",
45
  layout="wide"
46
  )
47
 
48
- # Download spaCy model if not already downloaded
49
- @st.cache_resource
50
- def download_spacy_model():
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
51
  try:
52
- nlp = spacy.load("en_core_web_sm")
53
- except OSError:
54
- subprocess.check_call([sys.executable, "-m", "spacy", "download", "en_core_web_sm"])
55
- nlp = spacy.load("en_core_web_sm")
56
- return nlp
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
57
 
58
- # Load the NLP models
59
  @st.cache_resource
60
  def load_models():
61
- summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
62
- nlp = download_spacy_model()
63
 
64
- # Load sentence transformer model for semantic matching
65
- sentence_model = SentenceTransformer('paraphrase-MiniLM-L6-v2')
 
 
 
 
 
 
 
 
 
 
 
 
 
66
 
67
- # Load Qwen3-8B model for career advice
68
- try:
69
- device = "cuda" if torch.cuda.is_available() else "cpu"
70
- qwen_tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen3-8B")
71
- qwen_model = AutoModelForCausalLM.from_pretrained(
72
- "Qwen/Qwen3-8B",
73
- torch_dtype=torch.float16 if device == "cuda" else torch.float32,
74
- device_map="auto"
75
- )
76
- except Exception as e:
77
- st.error(f"Failed to load Qwen3-8B model: {str(e)}")
78
- qwen_tokenizer = None
79
- qwen_model = None
80
 
81
- return summarizer, nlp, qwen_tokenizer, qwen_model, sentence_model
82
-
83
- # Initialize models
84
- summarizer, nlp, qwen_tokenizer, qwen_model, sentence_model = load_models()
 
 
 
 
 
 
 
85
 
86
- # Job descriptions and required skills
87
  job_descriptions = {
88
  "Software Engineer": {
89
  "skills": ["python", "java", "javascript", "sql", "algorithms", "data structures",
@@ -120,137 +162,272 @@ job_descriptions = {
120
  "Mid-level": "3-5 years of experience, model development, feature engineering",
121
  "Senior": "6+ years of experience, advanced ML techniques, research experience"
122
  }
123
- },
124
- "Product Manager": {
125
- "skills": ["product strategy", "roadmap planning", "user stories", "agile", "market research",
126
- "stakeholder management", "analytics", "user experience", "a/b testing", "prioritization"],
127
- "description": "Seeking product managers who can drive product vision, strategy, and execution.",
128
- "must_have": ["product strategy", "roadmap planning", "stakeholder management"],
129
- "nice_to_have": ["agile", "analytics", "a/b testing"],
130
- "seniority_levels": {
131
- "Junior": "0-2 years of experience, assisting with feature definition and user stories",
132
- "Mid-level": "3-5 years of experience, owning products/features, market research",
133
- "Senior": "6+ years of experience, defining product vision, managing teams, strategic planning"
134
- }
135
- },
136
- "DevOps Engineer": {
137
- "skills": ["linux", "aws", "docker", "kubernetes", "ci/cd", "terraform",
138
- "ansible", "monitoring", "scripting", "automation", "security"],
139
- "description": "Looking for DevOps engineers to build and maintain infrastructure and deployment pipelines.",
140
- "must_have": ["linux", "docker", "ci/cd"],
141
- "nice_to_have": ["kubernetes", "terraform", "aws"],
142
- "seniority_levels": {
143
- "Junior": "0-2 years of experience, basic system administration, scripting",
144
- "Mid-level": "3-5 years of experience, container orchestration, infrastructure as code",
145
- "Senior": "6+ years of experience, architecture design, security, team leadership"
146
- }
147
  }
148
  }
149
 
 
150
  def extract_text_from_pdf(pdf_file):
 
151
  text = ""
152
- with pdfplumber.open(pdf_file) as pdf:
153
- for page in pdf.pages:
154
- text += page.extract_text() or ""
 
 
 
155
  return text
156
 
157
- def analyze_resume(text, job_title):
158
- # Extract relevant skills
159
- doc = nlp(text.lower())
160
  found_skills = []
161
  required_skills = job_descriptions[job_title]["skills"]
162
 
 
163
  for skill in required_skills:
164
- if skill in text.lower():
165
  found_skills.append(skill)
166
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
167
  # Generate summary
168
- chunks = [text[i:i + 1000] for i in range(0, len(text), 1000)]
169
- summaries = []
170
- for chunk in chunks[:3]: # Process first 3000 characters to avoid token limits
171
- summary = summarizer(chunk, max_length=150, min_length=50, do_sample=False)[0]["summary_text"]
172
- summaries.append(summary)
 
 
 
173
 
174
- # Extract experience timeline
175
  experiences = extract_experience(text)
176
 
177
  # Calculate semantic match score
178
- match_score = semantic_matching(text, job_title)
179
-
180
- # Estimate seniority
181
- seniority, years_experience, leadership_count, must_have_percentage = estimate_seniority(experiences, found_skills, job_title)
 
 
 
 
 
 
 
 
182
 
183
- # Extract skill levels
184
- skill_levels = extract_skill_levels(text, found_skills)
185
 
186
- # Check for timeline inconsistencies
187
- inconsistencies = check_timeline_inconsistencies(experiences)
 
 
 
 
188
 
189
- # Verify companies
190
- company_verification = verify_companies(experiences)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
191
 
192
- # Predict career trajectory
193
- career_prediction = predict_career_trajectory(experiences, seniority, job_title)
194
 
195
  return {
196
  'found_skills': found_skills,
197
- 'summary': " ".join(summaries),
 
198
  'experiences': experiences,
199
  'match_score': match_score,
200
  'seniority': seniority,
201
- 'years_experience': years_experience,
202
- 'skill_levels': skill_levels,
203
  'inconsistencies': inconsistencies,
204
- 'company_verification': company_verification,
205
  'career_prediction': career_prediction
206
  }
207
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
208
  def generate_career_advice(resume_text, job_title, found_skills, missing_skills):
209
- if qwen_model is None or qwen_tokenizer is None:
210
- return "Career advice model not available. Please check the model installation."
211
-
212
- # Create a prompt for the model
213
- prompt = f"""
214
- You are a professional career advisor. Based on the resume and the target job position,
215
- provide personalized advice on skills to develop and suggest projects that would help the candidate
216
- become a better fit for the position.
217
 
218
- Resume summary: {resume_text[:1000]}...
219
 
220
- Target position: {job_title}
221
 
222
- Job requirements: {job_descriptions[job_title]['description']}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
223
 
224
- Skills the candidate has: {', '.join(found_skills)}
225
 
226
- Skills the candidate needs to develop: {', '.join(missing_skills)}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
227
 
228
- Provide the following:
229
- 1. Specific advice on how to develop the missing skills
230
- 2. 3-5 project ideas that would showcase these skills
231
- 3. Resources for learning (courses, books, websites)
232
  """
 
 
233
 
234
- # Generate advice using Qwen3-8B
235
- try:
236
- inputs = qwen_tokenizer(prompt, return_tensors="pt").to(qwen_model.device)
237
- with torch.no_grad():
238
- outputs = qwen_model.generate(
239
- **inputs,
240
- max_new_tokens=1024,
241
- temperature=0.7,
242
- top_p=0.9,
243
- do_sample=True
244
- )
245
- advice = qwen_tokenizer.decode(outputs[0][inputs.input_ids.shape[1]:], skip_special_tokens=True)
246
- return advice
247
- except Exception as e:
248
- return f"Failed to generate career advice: {str(e)}"
249
 
250
- # Streamlit UI
251
- st.title("📄 Resume Screener & Skill Extractor")
252
 
253
- # Add description
254
  st.markdown("""
255
  This app helps recruiters analyze resumes by:
256
  - Extracting relevant skills for specific job positions
@@ -283,38 +460,36 @@ if uploaded_file and job_title:
283
  text = extract_text_from_pdf(uploaded_file)
284
 
285
  # Analyze resume
286
- resume_data = analyze_resume(text, job_title)
287
 
288
  # Calculate missing skills
289
  missing_skills = [skill for skill in job_descriptions[job_title]["skills"]
290
- if skill not in resume_data['found_skills']]
291
 
292
  # Display results in tabs
293
- tab1, tab2, tab3, tab4, tab5, tab6 = st.tabs([
294
  "📊 Skills Match",
295
  "📝 Resume Summary",
296
  "🎯 Skills Gap",
297
- "👨‍💼 Career Path",
298
- "🔍 Authentication",
299
  "🚀 Career Advice"
300
  ])
301
 
302
  with tab1:
303
- # First create columns for skill match percentage and semantic match
304
  col1, col2 = st.columns(2)
305
 
306
  with col1:
307
  # Display matched skills
308
  st.subheader("🎯 Matched Skills")
309
- if resume_data['found_skills']:
310
- for skill in resume_data['found_skills']:
311
  # Show skill with proficiency level
312
- level = resume_data['skill_levels'].get(skill, 'intermediate')
313
  level_emoji = "🟢" if level == 'advanced' else "🟡" if level == 'intermediate' else "🟠"
314
  st.success(f"{level_emoji} {skill.title()} ({level.title()})")
315
 
316
  # Calculate match percentage
317
- match_percentage = len(resume_data['found_skills']) / len(job_descriptions[job_title]["skills"]) * 100
318
  st.metric("Skills Match", f"{match_percentage:.1f}%")
319
  else:
320
  st.warning("No direct skill matches found.")
@@ -322,11 +497,11 @@ if uploaded_file and job_title:
322
  with col2:
323
  # Display semantic match score
324
  st.subheader("💡 Semantic Match")
325
- st.metric("Overall Match Score", f"{resume_data['match_score']:.1f}%")
326
 
327
  # Display must-have skills match
328
  must_have_skills = job_descriptions[job_title]["must_have"]
329
- must_have_count = sum(1 for skill in must_have_skills if skill in resume_data['found_skills'])
330
  must_have_percentage = (must_have_count / len(must_have_skills)) * 100
331
 
332
  st.write("Must-have skills:")
@@ -335,20 +510,20 @@ if uploaded_file and job_title:
335
 
336
  # Professional level assessment
337
  st.subheader("🧠 Seniority Assessment")
338
- st.info(f"**{resume_data['seniority']}** ({resume_data['years_experience']:.1f} years equivalent experience)")
339
- st.write(job_descriptions[job_title]["seniority_levels"][resume_data['seniority']])
340
 
341
  with tab2:
342
  # Display resume summary
343
  st.subheader("📝 Resume Summary")
344
- st.write(resume_data['summary'])
345
 
346
  # Display experience timeline
347
  st.subheader("⏳ Experience Timeline")
348
- if resume_data['experiences']:
349
  # Convert experiences to dataframe for display
350
  exp_data = []
351
- for exp in resume_data['experiences']:
352
  if 'start_date' in exp and 'end_date' in exp:
353
  exp_data.append({
354
  'Company': exp['company'],
@@ -369,33 +544,36 @@ if uploaded_file and job_title:
369
  st.dataframe(exp_df)
370
 
371
  # Create a timeline visualization if dates are available
372
- timeline_data = [exp for exp in resume_data['experiences'] if 'start_date' in exp and 'end_date' in exp]
373
- if timeline_data:
374
- # Sort by start date
375
- timeline_data = sorted(timeline_data, key=lambda x: x['start_date'])
376
-
377
- # Create figure
378
- fig = go.Figure()
379
-
380
- for i, exp in enumerate(timeline_data):
381
- fig.add_trace(go.Bar(
382
- x=[(exp['end_date'] - exp['start_date']).days / 30], # Duration in months
383
- y=[exp['company']],
384
- orientation='h',
385
- name=exp['role'],
386
- hovertext=f"{exp['role']} at {exp['company']}<br>{exp['start_date'].strftime('%b %Y')} - {exp['end_date'].strftime('%b %Y') if exp['end_date'] != datetime.now() else 'Present'}<br>Duration: {exp.get('duration_months', 0)} months",
387
- marker=dict(color=px.colors.qualitative.Plotly[i % len(px.colors.qualitative.Plotly)])
388
- ))
389
-
390
- fig.update_layout(
391
- title="Career Timeline",
392
- xaxis_title="Duration (months)",
393
- yaxis_title="Company",
394
- height=400,
395
- margin=dict(l=0, r=0, b=0, t=30)
396
- )
397
-
398
- st.plotly_chart(fig, use_container_width=True)
 
 
 
399
  else:
400
  st.warning("No work experience data could be extracted.")
401
 
@@ -420,7 +598,7 @@ if uploaded_file and job_title:
420
 
421
  # Show must-have skills that are missing
422
  missing_must_have = [skill for skill in job_descriptions[job_title]["must_have"]
423
- if skill not in resume_data['found_skills']]
424
 
425
  if missing_must_have:
426
  st.error("**Critical Skills Missing:**")
@@ -433,7 +611,7 @@ if uploaded_file and job_title:
433
 
434
  # Show nice-to-have skills gap
435
  missing_nice_to_have = [skill for skill in job_descriptions[job_title]["nice_to_have"]
436
- if skill not in resume_data['found_skills']]
437
 
438
  if missing_nice_to_have:
439
  st.warning("**Nice-to-Have Skills Missing:**")
@@ -441,388 +619,24 @@ if uploaded_file and job_title:
441
  st.write(f"- {skill.title()}")
442
  else:
443
  st.success("Candidate has all the nice-to-have skills!")
444
-
445
- with tab4:
446
- # Display career path insights
447
- st.subheader("👨‍💼 Career Trajectory")
448
-
449
- # Show career prediction
450
- st.info(resume_data['career_prediction'])
451
-
452
- # Show experience trends
453
- st.subheader("📈 Experience Analysis")
454
-
455
- # Check for job hopping
456
- if len(resume_data['experiences']) >= 3:
457
- # Calculate average job duration
458
- durations = [exp.get('duration_months', 0) for exp in resume_data['experiences']
459
- if 'duration_months' in exp]
460
-
461
- if durations:
462
- avg_duration = sum(durations) / len(durations)
463
 
464
- if avg_duration < 12:
465
- st.warning(f"🚩 **Frequent Job Changes**: Average job duration is only {avg_duration:.1f} months")
466
- elif avg_duration < 24:
467
- st.warning(f"⚠️ **Moderate Job Hopping**: Average job duration is {avg_duration:.1f} months")
468
- else:
469
- st.success(f"✅ **Stable Employment**: Average job duration is {avg_duration:.1f} months")
470
-
471
- # Show inconsistencies if any
472
- if resume_data['inconsistencies']:
473
- st.subheader("⚠️ Timeline Inconsistencies")
474
- for issue in resume_data['inconsistencies']:
475
- if issue['type'] == 'overlap':
476
- st.warning(issue['description'])
477
- elif issue['type'] == 'gap':
478
- st.info(issue['description'])
479
-
480
- with tab5:
481
- # Display authentication signals
482
- st.subheader("🔍 Resume Authentication")
483
-
484
- # Company verification results
485
- st.write("**Company Verification Results:**")
486
-
487
- if resume_data['company_verification']:
488
- # Count suspicious companies
489
- suspicious_count = sum(1 for v in resume_data['company_verification']
490
- if v['status'] == 'suspicious')
491
-
492
- if suspicious_count == 0:
493
- st.success("✅ All companies mentioned in the resume passed basic verification")
494
- else:
495
- st.warning(f"⚠️ {suspicious_count} companies require further verification")
496
-
497
- # Display verification details
498
- verification_data = [{
499
- 'Company': v['company'],
500
- 'Status': v['status'].title(),
501
- 'Notes': v['reason']
502
- } for v in resume_data['company_verification']]
503
-
504
- st.dataframe(pd.DataFrame(verification_data))
505
- else:
506
- st.info("No company information found for verification.")
507
-
508
- # Timeline consistency check
509
- st.write("**Timeline Consistency Check:**")
510
-
511
- if not resume_data['inconsistencies']:
512
- st.success("✅ No timeline inconsistencies detected")
513
- else:
514
- st.warning(f"⚠️ {len(resume_data['inconsistencies'])} timeline inconsistencies found")
515
- for issue in resume_data['inconsistencies']:
516
- st.write(f"- {issue['description']}")
517
 
518
- with tab6:
519
  # Display career advice
520
  st.subheader("🚀 Career Advice and Project Recommendations")
521
 
522
  if st.button("Generate Career Advice"):
523
  with st.spinner("Generating personalized career advice..."):
524
- advice = generate_career_advice(text, job_title, resume_data['found_skills'], missing_skills)
525
  st.markdown(advice)
526
 
527
  except Exception as e:
528
  st.error(f"An error occurred while processing the resume: {str(e)}")
 
529
 
530
  # Add footer
531
  st.markdown("---")
532
- st.markdown("Made with ❤️ using Streamlit and Hugging Face")
533
-
534
- # Semantic matching between resume and job description
535
- def semantic_matching(resume_text, job_title):
536
- job_desc = job_descriptions[job_title]["description"]
537
-
538
- # Encode texts using sentence transformers
539
- resume_embedding = sentence_model.encode(resume_text, convert_to_tensor=True)
540
- job_embedding = sentence_model.encode(job_desc, convert_to_tensor=True)
541
-
542
- # Calculate cosine similarity
543
- cos_sim = cosine_similarity(
544
- resume_embedding.cpu().numpy().reshape(1, -1),
545
- job_embedding.cpu().numpy().reshape(1, -1)
546
- )[0][0]
547
-
548
- return cos_sim * 100 # Convert to percentage
549
-
550
- # Extract experience timeline from resume
551
- def extract_experience(text):
552
- # Pattern to find work experience entries
553
- # Look for patterns like "Company Name | Role | Jan 2020 - Present"
554
- exp_pattern = r"(?i)(.*?(?:inc|llc|ltd|company|corp|corporation|group)?)\s*(?:[|•-]\s*)?(.*?)(?:[|•-]\s*)((?:jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)[\w\s,]*\d{4}\s*(?:-|to|–)\s*(?:(?:jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)[\w\s,]*\d{4}|present))"
555
-
556
- experiences = []
557
- for match in re.finditer(exp_pattern, text, re.IGNORECASE):
558
- company = match.group(1).strip()
559
- role = match.group(2).strip()
560
- duration = match.group(3).strip()
561
-
562
- # Parse dates
563
- try:
564
- date_range = duration.split('-') if '-' in duration else duration.split('to') if 'to' in duration else duration.split('–')
565
- start_date = dateparser.parse(date_range[0].strip())
566
-
567
- if 'present' in date_range[1].lower():
568
- end_date = datetime.now()
569
- else:
570
- end_date = dateparser.parse(date_range[1].strip())
571
-
572
- if start_date and end_date:
573
- # Calculate duration in months
574
- months = (end_date.year - start_date.year) * 12 + (end_date.month - start_date.month)
575
-
576
- experiences.append({
577
- 'company': company,
578
- 'role': role,
579
- 'start_date': start_date,
580
- 'end_date': end_date,
581
- 'duration_months': months
582
- })
583
- except:
584
- # If date parsing fails, still include the experience without dates
585
- experiences.append({
586
- 'company': company,
587
- 'role': role,
588
- 'duration': duration
589
- })
590
-
591
- return experiences
592
-
593
- # Estimate seniority based on experience and skills
594
- def estimate_seniority(experiences, found_skills, job_title):
595
- # Calculate total experience in years
596
- total_months = sum(exp.get('duration_months', 0) for exp in experiences if 'duration_months' in exp)
597
- total_years = total_months / 12
598
-
599
- # Count leadership keywords in roles
600
- leadership_keywords = ['lead', 'senior', 'manager', 'head', 'principal', 'architect', 'director']
601
- leadership_count = 0
602
-
603
- for exp in experiences:
604
- role = exp.get('role', '').lower()
605
- for keyword in leadership_keywords:
606
- if keyword in role:
607
- leadership_count += 1
608
- break
609
-
610
- # Calculate skill match percentage for must-have skills
611
- must_have_skills = job_descriptions[job_title]["must_have"]
612
- must_have_count = sum(1 for skill in must_have_skills if skill in [s.lower() for s in found_skills])
613
- must_have_percentage = (must_have_count / len(must_have_skills)) * 100 if must_have_skills else 0
614
-
615
- # Determine seniority level
616
- if total_years < 3:
617
- seniority = "Junior"
618
- elif total_years < 6:
619
- seniority = "Mid-level"
620
- else:
621
- seniority = "Senior"
622
-
623
- # Adjust based on leadership roles and skill match
624
- if leadership_count >= 2 and seniority != "Senior":
625
- seniority = "Senior" if total_years >= 4 else seniority
626
- if must_have_percentage < 50 and seniority == "Senior":
627
- seniority = "Mid-level"
628
-
629
- return seniority, total_years, leadership_count, must_have_percentage
630
-
631
- # Check for timeline inconsistencies
632
- def check_timeline_inconsistencies(experiences):
633
- if not experiences:
634
- return []
635
-
636
- inconsistencies = []
637
- sorted_experiences = sorted(
638
- [exp for exp in experiences if 'start_date' in exp and 'end_date' in exp],
639
- key=lambda x: x['start_date']
640
- )
641
-
642
- for i in range(len(sorted_experiences) - 1):
643
- current = sorted_experiences[i]
644
- next_exp = sorted_experiences[i + 1]
645
-
646
- # Check for overlapping full-time roles
647
- if current['end_date'] > next_exp['start_date']:
648
- overlap_months = (current['end_date'].year - next_exp['start_date'].year) * 12 + \
649
- (current['end_date'].month - next_exp['start_date'].month)
650
-
651
- if overlap_months > 1: # Allow 1 month overlap for transitions
652
- inconsistencies.append({
653
- 'type': 'overlap',
654
- 'description': f"Overlapping roles: {current['company']} and {next_exp['company']} " +
655
- f"overlap by {overlap_months} months"
656
- })
657
-
658
- # Check for gaps in employment
659
- for i in range(len(sorted_experiences) - 1):
660
- current = sorted_experiences[i]
661
- next_exp = sorted_experiences[i + 1]
662
-
663
- gap_months = (next_exp['start_date'].year - current['end_date'].year) * 12 + \
664
- (next_exp['start_date'].month - current['end_date'].month)
665
-
666
- if gap_months > 3: # Flag gaps longer than 3 months
667
- inconsistencies.append({
668
- 'type': 'gap',
669
- 'description': f"Employment gap of {gap_months} months between " +
670
- f"{current['company']} and {next_exp['company']}"
671
- })
672
-
673
- return inconsistencies
674
-
675
- # Verify company existence (simplified version)
676
- def verify_companies(experiences):
677
- verification_results = []
678
-
679
- for exp in experiences:
680
- company = exp.get('company', '')
681
- if not company:
682
- continue
683
-
684
- # Simple heuristic - companies less than 3 characters are suspicious
685
- if len(company) < 3:
686
- verification_results.append({
687
- 'company': company,
688
- 'status': 'suspicious',
689
- 'reason': 'Company name too short'
690
- })
691
- continue
692
-
693
- # Check if company matches common fake patterns
694
- fake_patterns = ['abc company', 'xyz corp', 'my company', 'personal project']
695
- if any(pattern in company.lower() for pattern in fake_patterns):
696
- verification_results.append({
697
- 'company': company,
698
- 'status': 'suspicious',
699
- 'reason': 'Matches pattern of fake company names'
700
- })
701
- continue
702
-
703
- # In a real implementation, you'd call an API to check if the company exists
704
- # For this demo, we'll just mark all others as verified
705
- verification_results.append({
706
- 'company': company,
707
- 'status': 'verified',
708
- 'reason': 'Passed basic verification checks'
709
- })
710
-
711
- return verification_results
712
-
713
- # Extract skill levels from text
714
- def extract_skill_levels(text, skills):
715
- skill_levels = {}
716
- proficiency_indicators = {
717
- 'basic': ['basic', 'familiar', 'beginner', 'fundamentals', 'exposure'],
718
- 'intermediate': ['intermediate', 'proficient', 'experienced', 'competent', 'skilled'],
719
- 'advanced': ['advanced', 'expert', 'mastery', 'specialist', 'lead', 'senior']
720
- }
721
-
722
- for skill in skills:
723
- # Look for sentences containing the skill
724
- sentences = re.findall(r'[^.!?]*%s[^.!?]*[.!?]' % re.escape(skill), text.lower())
725
-
726
- # Default level
727
- level = 'intermediate'
728
-
729
- # Check for years of experience indicators
730
- years_pattern = re.compile(r'(\d+)\s*(?:\+)?\s*years?(?:\s+of)?\s+(?:experience|exp)?\s+(?:with|in|using)?\s+%s' % re.escape(skill), re.IGNORECASE)
731
- for sentence in sentences:
732
- years_match = years_pattern.search(sentence)
733
- if years_match:
734
- years = int(years_match.group(1))
735
- if years < 2:
736
- level = 'basic'
737
- elif years < 5:
738
- level = 'intermediate'
739
- else:
740
- level = 'advanced'
741
- break
742
-
743
- # Check for proficiency indicators
744
- if level == 'intermediate': # Only override if not already set by years
745
- for level_name, indicators in proficiency_indicators.items():
746
- for indicator in indicators:
747
- pattern = re.compile(r'%s\s+(?:\w+\s+){0,3}%s' % (indicator, re.escape(skill)), re.IGNORECASE)
748
- if any(pattern.search(sentence) for sentence in sentences):
749
- level = level_name
750
- break
751
- if level != 'intermediate':
752
- break
753
-
754
- skill_levels[skill] = level
755
-
756
- return skill_levels
757
-
758
- # Generate career trajectory prediction
759
- def predict_career_trajectory(experiences, seniority, job_title):
760
- if not experiences:
761
- return "Unable to predict trajectory due to insufficient experience data."
762
-
763
- # Extract roles in chronological order
764
- roles = [exp.get('role', '').lower() for exp in experiences if 'role' in exp]
765
-
766
- # If less than 2 roles, not enough data for prediction
767
- if len(roles) < 2:
768
- if seniority == "Junior":
769
- next_role = "Mid-level " + job_title
770
- elif seniority == "Mid-level":
771
- next_role = "Senior " + job_title
772
- else: # Senior
773
- leadership_titles = {
774
- "Software Engineer": "Technical Lead or Engineering Manager",
775
- "Data Scientist": "Lead Data Scientist or Data Science Manager",
776
- "Interaction Designer": "Design Lead or UX Director",
777
- "Product Manager": "Senior Product Manager or Director of Product",
778
- "DevOps Engineer": "DevOps Lead or Infrastructure Architect"
779
- }
780
- next_role = leadership_titles.get(job_title, f"Director of {job_title}")
781
-
782
- return f"Based on current seniority level, the next logical role could be: {next_role}"
783
-
784
- # Check for upward mobility patterns
785
- progression_indicators = ['junior', 'senior', 'lead', 'manager', 'director', 'vp', 'head', 'chief']
786
- current_level = -1
787
-
788
- for role in roles:
789
- for i, indicator in enumerate(progression_indicators):
790
- if indicator in role:
791
- if i > current_level:
792
- current_level = i
793
-
794
- # Predict next role based on current level
795
- if current_level < len(progression_indicators) - 1:
796
- next_level = progression_indicators[current_level + 1]
797
-
798
- # Map to specific job titles
799
- if next_level == 'senior' and 'senior' not in roles[-1].lower():
800
- next_role = f"Senior {job_title}"
801
- elif next_level == 'lead':
802
- next_role = f"{job_title} Lead"
803
- elif next_level == 'manager':
804
- if job_title == "Software Engineer":
805
- next_role = "Engineering Manager"
806
- else:
807
- next_role = f"{job_title} Manager"
808
- elif next_level == 'director':
809
- next_role = f"Director of {job_title}s"
810
- elif next_level == 'vp':
811
- next_role = f"VP of {job_title}s"
812
- elif next_level == 'head':
813
- next_role = f"Head of {job_title}"
814
- elif next_level == 'chief':
815
- if job_title == "Software Engineer":
816
- next_role = "CTO (Chief Technology Officer)"
817
- elif job_title == "Data Scientist":
818
- next_role = "Chief Data Officer"
819
- elif job_title == "Product Manager":
820
- next_role = "Chief Product Officer"
821
- else:
822
- next_role = f"Chief {job_title} Officer"
823
- else:
824
- next_role = f"{next_level.title()} {job_title}"
825
- else:
826
- next_role = "Executive Leadership or Strategic Advisory roles"
827
-
828
- return f"Based on career progression, the next logical role could be: {next_role}"
 
1
  import streamlit as st
2
  import pdfplumber
 
 
 
 
 
 
3
  import re
4
  import pandas as pd
5
+ import matplotlib.pyplot as plt
6
+ import torch
7
+ from datetime import datetime
8
  import plotly.express as px
9
  import plotly.graph_objects as go
10
+ import numpy as np
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
11
 
12
+ # Display startup message
13
  st.set_page_config(
14
  page_title="Resume Screener & Skill Extractor",
15
  page_icon="📄",
16
  layout="wide"
17
  )
18
 
19
+ st.title("📄 Resume Screener & Skill Extractor")
20
+ startup_message = st.empty()
21
+ startup_message.info("Loading dependencies and models... This may take a minute on first run.")
22
+
23
+ # Import dependencies with fallbacks
24
+ try:
25
+ import spacy
26
+ spacy_available = True
27
+ except ImportError:
28
+ spacy_available = False
29
+ st.warning("spaCy is not available. Some features will be limited.")
30
+
31
+ try:
32
+ from transformers import pipeline
33
+ transformers_available = True
34
+ except ImportError:
35
+ transformers_available = False
36
+ st.warning("Transformers is not available. Summary generation will be limited.")
37
+
38
+ try:
39
+ import nltk
40
+ from nltk.tokenize import word_tokenize
41
+ nltk_available = True
42
+
43
+ # Download required NLTK resources
44
  try:
45
+ nltk.data.find('tokenizers/punkt')
46
+ except LookupError:
47
+ nltk.download('punkt')
48
+ except ImportError:
49
+ nltk_available = False
50
+ st.warning("NLTK is not available. Some text processing features will be limited.")
51
+
52
+ # Custom sentence-transformers fallback
53
+ try:
54
+ from sentence_transformers import SentenceTransformer
55
+ try:
56
+ from sentence_transformers import util as st_util
57
+ sentence_transformers_available = True
58
+ except ImportError:
59
+ # Define our own utility functions
60
+ class CustomSTUtil:
61
+ @staticmethod
62
+ def pytorch_cos_sim(a, b):
63
+ if not isinstance(a, torch.Tensor):
64
+ a = torch.tensor(a)
65
+ if not isinstance(b, torch.Tensor):
66
+ b = torch.tensor(b)
67
+
68
+ if len(a.shape) == 1:
69
+ a = a.unsqueeze(0)
70
+ if len(b.shape) == 1:
71
+ b = b.unsqueeze(0)
72
+
73
+ a_norm = torch.nn.functional.normalize(a, p=2, dim=1)
74
+ b_norm = torch.nn.functional.normalize(b, p=2, dim=1)
75
+ return torch.mm(a_norm, b_norm.transpose(0, 1))
76
+
77
+ st_util = CustomSTUtil()
78
+ sentence_transformers_available = True
79
+ except ImportError:
80
+ sentence_transformers_available = False
81
+ st.warning("Sentence Transformers is not available. Semantic matching will be disabled.")
82
 
83
+ # Load models with exception handling
84
  @st.cache_resource
85
  def load_models():
86
+ models = {}
 
87
 
88
+ # Load spaCy if available
89
+ if spacy_available:
90
+ try:
91
+ models['nlp'] = spacy.load("en_core_web_sm")
92
+ except OSError:
93
+ try:
94
+ import subprocess
95
+ import sys
96
+ subprocess.check_call([sys.executable, "-m", "spacy", "download", "en_core_web_sm"])
97
+ models['nlp'] = spacy.load("en_core_web_sm")
98
+ except Exception as e:
99
+ st.warning(f"Could not load spaCy model: {e}")
100
+ models['nlp'] = None
101
+ else:
102
+ models['nlp'] = None
103
 
104
+ # Load summarizer if transformers available
105
+ if transformers_available:
106
+ try:
107
+ models['summarizer'] = pipeline("summarization", model="facebook/bart-large-cnn")
108
+ except Exception as e:
109
+ st.warning(f"Could not load summarizer model: {e}")
110
+ # Simple fallback summarizer
111
+ models['summarizer'] = lambda text, **kwargs: [{"summary_text": ". ".join(text.split(". ")[:5]) + "."}]
112
+ else:
113
+ # Simple fallback summarizer
114
+ models['summarizer'] = lambda text, **kwargs: [{"summary_text": ". ".join(text.split(". ")[:5]) + "."}]
 
 
115
 
116
+ # Load sentence transformer if available
117
+ if sentence_transformers_available:
118
+ try:
119
+ models['sentence_model'] = SentenceTransformer('paraphrase-MiniLM-L6-v2')
120
+ except Exception as e:
121
+ st.warning(f"Could not load sentence transformer model: {e}")
122
+ models['sentence_model'] = None
123
+ else:
124
+ models['sentence_model'] = None
125
+
126
+ return models
127
 
128
+ # Job descriptions dictionary
129
  job_descriptions = {
130
  "Software Engineer": {
131
  "skills": ["python", "java", "javascript", "sql", "algorithms", "data structures",
 
162
  "Mid-level": "3-5 years of experience, model development, feature engineering",
163
  "Senior": "6+ years of experience, advanced ML techniques, research experience"
164
  }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
165
  }
166
  }
167
 
168
+ # Core functionality
169
  def extract_text_from_pdf(pdf_file):
170
+ """Extract text from PDF file."""
171
  text = ""
172
+ try:
173
+ with pdfplumber.open(pdf_file) as pdf:
174
+ for page in pdf.pages:
175
+ text += page.extract_text() or ""
176
+ except Exception as e:
177
+ st.error(f"Error extracting text from PDF: {e}")
178
  return text
179
 
180
+ def extract_skills(text, job_title, nlp=None):
181
+ """Extract skills from resume text."""
 
182
  found_skills = []
183
  required_skills = job_descriptions[job_title]["skills"]
184
 
185
+ # Simple keyword matching (no NLP needed)
186
  for skill in required_skills:
187
+ if skill.lower() in text.lower():
188
  found_skills.append(skill)
189
 
190
+ return found_skills
191
+
192
+ def extract_experience(text):
193
+ """Extract work experience from resume text."""
194
+ experiences = []
195
+
196
+ # Define regex pattern for experiences
197
+ experience_pattern = r"(?i)(\w+[\w\s&,.']+)\s*(?:[-|•]|\bat\b)\s*([A-Za-z][\w\s&,.']+)\s*(?:[-|•]|\bfrom\b)\s*(\d{4}(?:\s*[-–]\s*(?:\d{4}|present|current)))"
198
+
199
+ matches = re.finditer(experience_pattern, text)
200
+ for match in matches:
201
+ company = match.group(1).strip()
202
+ role = match.group(2).strip()
203
+ duration = match.group(3).strip()
204
+
205
+ # Process dates
206
+ try:
207
+ date_parts = re.split(r'[-–]', duration)
208
+ start_year = int(date_parts[0].strip())
209
+
210
+ if len(date_parts) > 1 and 'present' not in date_parts[1].lower() and 'current' not in date_parts[1].lower():
211
+ end_year = int(date_parts[1].strip())
212
+ end_date = datetime(end_year, 12, 31)
213
+ else:
214
+ end_year = datetime.now().year
215
+ end_date = datetime.now()
216
+
217
+ start_date = datetime(start_year, 1, 1)
218
+ duration_months = (end_date.year - start_date.year) * 12 + (end_date.month - start_date.month)
219
+
220
+ experiences.append({
221
+ 'company': company,
222
+ 'role': role,
223
+ 'start_date': start_date,
224
+ 'end_date': end_date,
225
+ 'duration_months': duration_months
226
+ })
227
+ except:
228
+ experiences.append({
229
+ 'company': company,
230
+ 'role': role,
231
+ 'duration': duration
232
+ })
233
+
234
+ return experiences
235
+
236
+ def analyze_resume(text, job_title, models):
237
+ """Analyze resume text."""
238
+ # Extract skills
239
+ found_skills = extract_skills(text, job_title, models.get('nlp'))
240
+
241
  # Generate summary
242
+ if models.get('summarizer'):
243
+ try:
244
+ summary = models['summarizer'](text[:3000], max_length=150, min_length=50, do_sample=False)[0]["summary_text"]
245
+ except Exception as e:
246
+ st.warning(f"Error generating summary: {e}")
247
+ summary = text[:500] + "..."
248
+ else:
249
+ summary = text[:500] + "..."
250
 
251
+ # Extract work experience
252
  experiences = extract_experience(text)
253
 
254
  # Calculate semantic match score
255
+ match_score = 0
256
+ if models.get('sentence_model') and sentence_transformers_available:
257
+ try:
258
+ resume_embedding = models['sentence_model'].encode(text[:5000], convert_to_tensor=True)
259
+ job_embedding = models['sentence_model'].encode(job_descriptions[job_title]["description"], convert_to_tensor=True)
260
+
261
+ match_score = float(st_util.pytorch_cos_sim(resume_embedding, job_embedding)[0][0]) * 100
262
+ except Exception as e:
263
+ st.warning(f"Error calculating semantic match: {e}")
264
+ else:
265
+ # Fallback to keyword-based score
266
+ match_score = (len(found_skills) / len(job_descriptions[job_title]["skills"])) * 100
267
 
268
+ # Calculate seniority level
269
+ years_exp = sum(exp.get('duration_months', 0) for exp in experiences if 'duration_months' in exp) / 12
270
 
271
+ if years_exp < 3:
272
+ seniority = "Junior"
273
+ elif years_exp < 6:
274
+ seniority = "Mid-level"
275
+ else:
276
+ seniority = "Senior"
277
 
278
+ # Detect skill levels
279
+ skill_levels = {}
280
+ for skill in found_skills:
281
+ # Default level
282
+ skill_levels[skill] = "intermediate"
283
+
284
+ # Look for advanced indicators
285
+ advanced_patterns = [
286
+ f"expert in {skill}",
287
+ f"advanced {skill}",
288
+ f"extensive experience with {skill}"
289
+ ]
290
+ if any(pattern in text.lower() for pattern in advanced_patterns):
291
+ skill_levels[skill] = "advanced"
292
+
293
+ # Look for basic indicators
294
+ basic_patterns = [
295
+ f"familiar with {skill}",
296
+ f"basic knowledge of {skill}",
297
+ f"introduced to {skill}"
298
+ ]
299
+ if any(pattern in text.lower() for pattern in basic_patterns):
300
+ skill_levels[skill] = "basic"
301
+
302
+ # Check for inconsistencies in timeline
303
+ inconsistencies = []
304
+ if len(experiences) >= 2:
305
+ # Sort experiences by start date
306
+ sorted_exps = sorted(
307
+ [exp for exp in experiences if 'start_date' in exp],
308
+ key=lambda x: x['start_date']
309
+ )
310
+
311
+ # Check for overlaps
312
+ for i in range(len(sorted_exps) - 1):
313
+ current = sorted_exps[i]
314
+ next_exp = sorted_exps[i+1]
315
+
316
+ if current['end_date'] > next_exp['start_date']:
317
+ inconsistencies.append({
318
+ 'type': 'overlap',
319
+ 'description': f"Overlapping roles at {current['company']} and {next_exp['company']}"
320
+ })
321
 
322
+ # Generate a simple career prediction
323
+ career_prediction = predict_career_path(seniority, job_title)
324
 
325
  return {
326
  'found_skills': found_skills,
327
+ 'skill_levels': skill_levels,
328
+ 'summary': summary,
329
  'experiences': experiences,
330
  'match_score': match_score,
331
  'seniority': seniority,
332
+ 'years_experience': years_exp,
 
333
  'inconsistencies': inconsistencies,
 
334
  'career_prediction': career_prediction
335
  }
336
 
337
+ def predict_career_path(seniority, job_title):
338
+ """Generate a simple career prediction."""
339
+ if seniority == "Junior":
340
+ return f"Next potential role: Senior {job_title}"
341
+ elif seniority == "Mid-level":
342
+ roles = {
343
+ "Software Engineer": "Team Lead, Technical Lead, or Engineering Manager",
344
+ "Data Scientist": "Senior Data Scientist or Data Science Lead",
345
+ "Interaction Designer": "Senior Designer or UX Lead"
346
+ }
347
+ return f"Next potential roles: {roles.get(job_title, f'Senior {job_title}')}"
348
+ else: # Senior
349
+ roles = {
350
+ "Software Engineer": "Engineering Manager, Software Architect, or CTO",
351
+ "Data Scientist": "Head of Data Science, ML Engineering Manager, or Chief Data Officer",
352
+ "Interaction Designer": "Design Director, Head of UX, or VP of Design"
353
+ }
354
+ return f"Next potential roles: {roles.get(job_title, f'Director of {job_title}')}"
355
+
356
  def generate_career_advice(resume_text, job_title, found_skills, missing_skills):
357
+ """Generate career advice based on resume analysis."""
358
+ advice = f"""## Career Development Plan for {job_title}
 
 
 
 
 
 
359
 
360
+ ### Skills to Develop
361
 
362
+ The following skills would strengthen your profile for this position:
363
 
364
+ """
365
+
366
+ for skill in missing_skills:
367
+ advice += f"- **{skill.title()}**: "
368
+
369
+ if skill == "python":
370
+ advice += "Take online courses like Coursera's Python for Everybody or follow tutorials on Real Python."
371
+ elif skill == "java":
372
+ advice += "Complete the Oracle Java Certification or contribute to open-source Java projects."
373
+ elif skill == "javascript":
374
+ advice += "Build interactive web applications using modern frameworks like React or Vue."
375
+ elif skill == "cloud":
376
+ advice += "Get hands-on experience with AWS, Azure, or GCP through their free tier offerings."
377
+ elif "algorithm" in skill or "data structure" in skill:
378
+ advice += "Practice on platforms like LeetCode or HackerRank and study algorithm design principles."
379
+ elif "ui" in skill or "ux" in skill:
380
+ advice += "Create a portfolio of design work and study interaction design principles."
381
+ elif "machine learning" in skill:
382
+ advice += "Take Andrew Ng's Machine Learning course on Coursera and work on ML projects with real datasets."
383
+ else:
384
+ advice += f"Research and practice this skill through online courses, tutorials, and hands-on projects."
385
+
386
+ advice += "\n\n"
387
+
388
+ advice += f"""
389
+ ### Project Ideas
390
 
391
+ Consider these projects to showcase your skills for a {job_title} position:
392
 
393
+ """
394
+
395
+ if job_title == "Software Engineer":
396
+ advice += """
397
+ 1. **Full-Stack Web Application**: Build a complete web app with frontend, backend, and database
398
+ 2. **API Service**: Create a RESTful or GraphQL API with proper authentication and documentation
399
+ 3. **Open Source Contribution**: Contribute to relevant open-source projects in your area of interest
400
+ """
401
+ elif job_title == "Data Scientist":
402
+ advice += """
403
+ 1. **Predictive Model**: Build and deploy a machine learning model that solves a real-world problem
404
+ 2. **Data Dashboard**: Create an interactive visualization dashboard for complex datasets
405
+ 3. **Natural Language Processing**: Develop a text classification or sentiment analysis project
406
+ """
407
+ elif job_title == "Interaction Designer":
408
+ advice += """
409
+ 1. **Design System**: Create a comprehensive design system with components and usage guidelines
410
+ 2. **UX Case Study**: Document your design process for a real or fictional product improvement
411
+ 3. **Interactive Prototype**: Design a fully functional prototype that demonstrates your interaction design skills
412
+ """
413
+
414
+ advice += """
415
+ ### Learning Resources
416
 
417
+ - **Online Platforms**: Coursera, Udemy, Pluralsight, LinkedIn Learning
418
+ - **Practice Sites**: GitHub, HackerRank, LeetCode, Kaggle
419
+ - **Communities**: Stack Overflow, Reddit programming communities, relevant Discord servers
 
420
  """
421
+
422
+ return advice
423
 
424
+ # Load models
425
+ models = load_models()
 
 
 
 
 
 
 
 
 
 
 
 
 
426
 
427
+ # Clear startup message
428
+ startup_message.empty()
429
 
430
+ # App description
431
  st.markdown("""
432
  This app helps recruiters analyze resumes by:
433
  - Extracting relevant skills for specific job positions
 
460
  text = extract_text_from_pdf(uploaded_file)
461
 
462
  # Analyze resume
463
+ analysis_results = analyze_resume(text, job_title, models)
464
 
465
  # Calculate missing skills
466
  missing_skills = [skill for skill in job_descriptions[job_title]["skills"]
467
+ if skill not in analysis_results['found_skills']]
468
 
469
  # Display results in tabs
470
+ tab1, tab2, tab3, tab4 = st.tabs([
471
  "📊 Skills Match",
472
  "📝 Resume Summary",
473
  "🎯 Skills Gap",
 
 
474
  "🚀 Career Advice"
475
  ])
476
 
477
  with tab1:
478
+ # Create two columns
479
  col1, col2 = st.columns(2)
480
 
481
  with col1:
482
  # Display matched skills
483
  st.subheader("🎯 Matched Skills")
484
+ if analysis_results['found_skills']:
485
+ for skill in analysis_results['found_skills']:
486
  # Show skill with proficiency level
487
+ level = analysis_results['skill_levels'].get(skill, 'intermediate')
488
  level_emoji = "🟢" if level == 'advanced' else "🟡" if level == 'intermediate' else "🟠"
489
  st.success(f"{level_emoji} {skill.title()} ({level.title()})")
490
 
491
  # Calculate match percentage
492
+ match_percentage = len(analysis_results['found_skills']) / len(job_descriptions[job_title]["skills"]) * 100
493
  st.metric("Skills Match", f"{match_percentage:.1f}%")
494
  else:
495
  st.warning("No direct skill matches found.")
 
497
  with col2:
498
  # Display semantic match score
499
  st.subheader("💡 Semantic Match")
500
+ st.metric("Overall Match Score", f"{analysis_results['match_score']:.1f}%")
501
 
502
  # Display must-have skills match
503
  must_have_skills = job_descriptions[job_title]["must_have"]
504
+ must_have_count = sum(1 for skill in must_have_skills if skill in analysis_results['found_skills'])
505
  must_have_percentage = (must_have_count / len(must_have_skills)) * 100
506
 
507
  st.write("Must-have skills:")
 
510
 
511
  # Professional level assessment
512
  st.subheader("🧠 Seniority Assessment")
513
+ st.info(f"**{analysis_results['seniority']}** ({analysis_results['years_experience']:.1f} years equivalent experience)")
514
+ st.write(job_descriptions[job_title]["seniority_levels"][analysis_results['seniority']])
515
 
516
  with tab2:
517
  # Display resume summary
518
  st.subheader("📝 Resume Summary")
519
+ st.write(analysis_results['summary'])
520
 
521
  # Display experience timeline
522
  st.subheader("⏳ Experience Timeline")
523
+ if analysis_results['experiences']:
524
  # Convert experiences to dataframe for display
525
  exp_data = []
526
+ for exp in analysis_results['experiences']:
527
  if 'start_date' in exp and 'end_date' in exp:
528
  exp_data.append({
529
  'Company': exp['company'],
 
544
  st.dataframe(exp_df)
545
 
546
  # Create a timeline visualization if dates are available
547
+ timeline_data = [exp for exp in analysis_results['experiences'] if 'start_date' in exp and 'end_date' in exp]
548
+ if timeline_data and len(timeline_data) > 0:
549
+ try:
550
+ # Sort by start date
551
+ timeline_data = sorted(timeline_data, key=lambda x: x['start_date'])
552
+
553
+ # Create figure
554
+ fig = go.Figure()
555
+
556
+ for i, exp in enumerate(timeline_data):
557
+ fig.add_trace(go.Bar(
558
+ x=[(exp['end_date'] - exp['start_date']).days / 30], # Duration in months
559
+ y=[exp['company']],
560
+ orientation='h',
561
+ name=exp['role'],
562
+ hovertext=f"{exp['role']} at {exp['company']}",
563
+ marker=dict(color=px.colors.qualitative.Plotly[i % len(px.colors.qualitative.Plotly)])
564
+ ))
565
+
566
+ fig.update_layout(
567
+ title="Career Timeline",
568
+ xaxis_title="Duration (months)",
569
+ yaxis_title="Company",
570
+ height=400,
571
+ margin=dict(l=0, r=0, b=0, t=30)
572
+ )
573
+
574
+ st.plotly_chart(fig, use_container_width=True)
575
+ except Exception as e:
576
+ st.warning(f"Could not create timeline visualization: {e}")
577
  else:
578
  st.warning("No work experience data could be extracted.")
579
 
 
598
 
599
  # Show must-have skills that are missing
600
  missing_must_have = [skill for skill in job_descriptions[job_title]["must_have"]
601
+ if skill not in analysis_results['found_skills']]
602
 
603
  if missing_must_have:
604
  st.error("**Critical Skills Missing:**")
 
611
 
612
  # Show nice-to-have skills gap
613
  missing_nice_to_have = [skill for skill in job_descriptions[job_title]["nice_to_have"]
614
+ if skill not in analysis_results['found_skills']]
615
 
616
  if missing_nice_to_have:
617
  st.warning("**Nice-to-Have Skills Missing:**")
 
619
  st.write(f"- {skill.title()}")
620
  else:
621
  st.success("Candidate has all the nice-to-have skills!")
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
622
 
623
+ # Display career trajectory
624
+ st.subheader("👨‍💼 Career Trajectory")
625
+ st.info(analysis_results['career_prediction'])
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
626
 
627
+ with tab4:
628
  # Display career advice
629
  st.subheader("🚀 Career Advice and Project Recommendations")
630
 
631
  if st.button("Generate Career Advice"):
632
  with st.spinner("Generating personalized career advice..."):
633
+ advice = generate_career_advice(text, job_title, analysis_results['found_skills'], missing_skills)
634
  st.markdown(advice)
635
 
636
  except Exception as e:
637
  st.error(f"An error occurred while processing the resume: {str(e)}")
638
+ st.exception(e)
639
 
640
  # Add footer
641
  st.markdown("---")
642
+ st.markdown("Made with ❤️ using Streamlit and Hugging Face")
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
requirements.txt CHANGED
@@ -1,24 +1,22 @@
1
- streamlit==1.22.0
2
- pdfplumber==0.9.0
3
- spacy==3.5.3
4
- transformers==4.28.1
5
- torch==1.13.1
6
- huggingface-hub==0.14.1
7
  sentence-transformers==2.2.2
8
- nltk==3.8.1
9
- plotly==5.14.1
 
 
 
 
 
 
 
 
10
  pandas==1.5.3
11
  numpy==1.24.3
12
  matplotlib==3.7.1
13
- pydantic==1.10.8
14
- protobuf<4.0.0
15
- tqdm>=4.27
16
- regex>=2022.1.18
17
- scikit-learn==1.0.2
18
- scipy==1.8.1
19
- dateparser==1.1.8
20
- python-Levenshtein==0.21.1
21
- networkx==2.8.8
22
- faiss-cpu==1.7.4
23
- beautifulsoup4==4.12.2
24
- requests==2.31.0
 
1
+ # Core dependencies - order matters!
2
+ pydantic==1.10.8
3
+ spacy==3.5.0
 
 
 
4
  sentence-transformers==2.2.2
5
+ torch==1.13.1
6
+ transformers==4.28.1
7
+
8
+ # PDF processing
9
+ pdfplumber==0.9.0
10
+
11
+ # Web UI
12
+ streamlit==1.22.0
13
+
14
+ # Data processing
15
  pandas==1.5.3
16
  numpy==1.24.3
17
  matplotlib==3.7.1
18
+ plotly==5.14.1
19
+
20
+ # Utilities
21
+ nltk==3.8.1
22
+ scikit-learn==1.0.2