Spaces:
Running
Running
File size: 12,283 Bytes
07273d8 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 |
"""Skill extraction from CV text using LLM."""
from __future__ import annotations
from typing import Any, Dict, List
from LLM.llm_models import cv_analyzer_model # Using CV-specific model (same as orchestrator)
import json
import re
def _print_terminal_log(action: str, details: str = ""):
"""Print formatted log to terminal."""
timestamp = __import__('datetime').datetime.now().strftime("%H:%M:%S")
if details:
print(f"[{timestamp}] [CV ANALYZER] {action} :: {details}")
else:
print(f"[{timestamp}] [CV ANALYZER] {action}")
def extract_skills_from_cv_text(cv_text: str, log_callback=None) -> Dict[str, Any]:
"""
Extract skills and relevant information from CV text using LLM.
Args:
cv_text: The extracted text content from a CV
Returns:
Dictionary containing extracted skills and candidate information
"""
prompt = f"""Analyze the following CV/Resume text and extract ALL relevant information in a structured format.
CV TEXT:
{cv_text}
Please extract and organize the following information:
1. TECHNICAL SKILLS: Programming languages, frameworks, tools, technologies
2. SOFT SKILLS: Communication, leadership, teamwork, problem-solving, etc.
3. PROFESSIONAL EXPERIENCE: Years of experience, job titles, companies
4. EDUCATION: Degrees, certifications, institutions
5. DOMAIN EXPERTISE: Industries, specific domains (e.g., Finance, Healthcare, AI/ML)
Return your analysis in the following JSON-like structure:
{{
"technical_skills": ["skill1", "skill2", ...],
"soft_skills": ["skill1", "skill2", ...],
"experience_years": <number or "unknown">,
"recent_roles": ["role1", "role2", ...],
"education": ["degree1", "degree2", ...],
"certifications": ["cert1", "cert2", ...],
"domain_expertise": ["domain1", "domain2", ...],
"summary": "A brief 2-3 sentence summary of the candidate's profile"
}}
Be thorough and extract as many relevant skills as possible. If information is not available, use empty arrays or "unknown"."""
try:
_print_terminal_log("Starting AI skill extraction from CV text")
if log_callback:
log_callback("AI Skill Extraction", {"status": "Initializing LLM model..."})
# Use the CV analyzer model (same provider as orchestrator - HF/Gemini)
model = cv_analyzer_model
_print_terminal_log("LLM Initialized", f"Model ready, CV length: {len(cv_text)} chars")
if log_callback:
log_callback("AI Analysis", {"status": "Sending CV to AI for analysis", "cv_length": len(cv_text)})
messages = [
{
"role": "system",
"content": "You are an expert HR analyst specializing in CV/Resume analysis and skill extraction. Extract information accurately and comprehensively."
},
{
"role": "user",
"content": prompt
}
]
if log_callback:
log_callback("LLM Request", {"message_count": len(messages), "model": "cv_analyzer_model"})
_print_terminal_log("Sending request to AI", "Waiting for skill extraction...")
response = model.generate(messages=messages)
# Handle ChatMessage object - convert to string
if hasattr(response, 'content'):
response_text = response.content
else:
response_text = str(response)
_print_terminal_log("AI Response Received", f"Response length: {len(response_text) if response_text else 0} chars")
if log_callback:
log_callback("AI Response Received", {"response_length": len(response_text) if response_text else 0})
# Extract JSON from response (handle markdown code blocks)
if log_callback:
log_callback("Parsing AI Response", {"status": "Extracting structured data from AI response"})
json_match = re.search(r'\{[\s\S]*\}', response_text)
if json_match:
if log_callback:
log_callback("JSON Extraction", {"status": "Found JSON in response, parsing..."})
_print_terminal_log("Parsing JSON response", "Extracting structured skill data...")
skills_data = json.loads(json_match.group())
tech_count = len(skills_data.get("technical_skills", []))
soft_count = len(skills_data.get("soft_skills", []))
_print_terminal_log("Skills Extracted Successfully",
f"Technical: {tech_count}, Soft: {soft_count}, Total: {tech_count + soft_count}")
if log_callback:
log_callback("Skills Parsed Successfully", {
"technical_skills": tech_count,
"soft_skills": soft_count,
"total_skills": tech_count + soft_count
})
else:
if log_callback:
log_callback("JSON Extraction Failed", {"status": "No JSON found, using fallback structure"})
# Fallback: return a basic structure with the raw response
skills_data = {
"technical_skills": [],
"soft_skills": [],
"experience_years": "unknown",
"recent_roles": [],
"education": [],
"certifications": [],
"domain_expertise": [],
"summary": response_text[:500] # First 500 chars
}
_print_terminal_log("✅ CV Analysis Complete", "All skills successfully extracted and structured")
if log_callback:
log_callback("✅ Extraction Complete", {"status": "CV processing finished successfully"})
return skills_data
except Exception as e:
error_msg = str(e)
_print_terminal_log(f"❌ ERROR: {type(e).__name__}", error_msg)
if log_callback:
log_callback("❌ AI Extraction Error", {"error": error_msg, "type": type(e).__name__})
# Return error information
return {
"error": error_msg,
"technical_skills": [],
"soft_skills": [],
"experience_years": "unknown",
"recent_roles": [],
"education": [],
"certifications": [],
"domain_expertise": [],
"summary": f"Failed to extract skills: {error_msg}"
}
def format_skills_for_display(skills_data: Dict[str, Any]) -> str:
"""
Format extracted skills data into HTML for display in Gradio.
Args:
skills_data: Dictionary containing extracted skills
Returns:
HTML string for display
"""
if "error" in skills_data:
return f"""
<div style="padding: 20px; background: var(--bg-card); border: 1px solid var(--arc-red); border-radius: 4px;">
<h3 style="color: var(--arc-red); margin-top: 0;">⚠️ Error Extracting Skills</h3>
<p style="color: var(--text-dim);">{skills_data.get('summary', 'Unknown error')}</p>
</div>
"""
html_parts = [
'<div style="padding: 24px; background: var(--bg-card); border: 1px solid var(--border-bright); border-radius: 4px; margin-top: 20px;">',
'<h2 style="color: var(--arc-orange); margin-top: 0; display: flex; align-items: center; gap: 12px;">',
'<span style="font-size: 32px;">📄</span> CV ANALYSIS COMPLETE',
'</h2>',
]
# Summary
if skills_data.get("summary"):
html_parts.append(f'<div style="background: var(--bg-panel); padding: 16px; border-left: 4px solid var(--arc-cyan); margin-bottom: 24px;">')
html_parts.append(f'<p style="color: var(--text-main); margin: 0; line-height: 1.6;">{skills_data["summary"]}</p>')
html_parts.append('</div>')
# Technical Skills
if skills_data.get("technical_skills"):
html_parts.append('<div style="margin-bottom: 20px;">')
html_parts.append('<h3 style="color: var(--arc-yellow); margin-bottom: 12px;">💻 TECHNICAL SKILLS</h3>')
html_parts.append('<div style="display: flex; flex-wrap: wrap; gap: 8px;">')
for skill in skills_data["technical_skills"]:
html_parts.append(
f'<span style="background: var(--bg-panel); border: 1px solid var(--border-dim); '
f'padding: 6px 12px; border-radius: 4px; color: var(--arc-green); font-weight: 600; '
f'font-size: 13px;">{skill}</span>'
)
html_parts.append('</div></div>')
# Soft Skills
if skills_data.get("soft_skills"):
html_parts.append('<div style="margin-bottom: 20px;">')
html_parts.append('<h3 style="color: var(--arc-yellow); margin-bottom: 12px;">🤝 SOFT SKILLS</h3>')
html_parts.append('<div style="display: flex; flex-wrap: wrap; gap: 8px;">')
for skill in skills_data["soft_skills"]:
html_parts.append(
f'<span style="background: var(--bg-panel); border: 1px solid var(--border-dim); '
f'padding: 6px 12px; border-radius: 4px; color: var(--arc-cyan); font-weight: 600; '
f'font-size: 13px;">{skill}</span>'
)
html_parts.append('</div></div>')
# Experience & Roles
if skills_data.get("experience_years") or skills_data.get("recent_roles"):
html_parts.append('<div style="margin-bottom: 20px;">')
html_parts.append('<h3 style="color: var(--arc-yellow); margin-bottom: 12px;">💼 EXPERIENCE</h3>')
if skills_data.get("experience_years"):
html_parts.append(f'<p style="color: var(--text-main); margin: 8px 0;"><strong>Years:</strong> {skills_data["experience_years"]}</p>')
if skills_data.get("recent_roles"):
html_parts.append('<p style="color: var(--text-main); margin: 8px 0;"><strong>Recent Roles:</strong></p>')
html_parts.append('<ul style="color: var(--text-dim); margin-top: 4px;">')
for role in skills_data["recent_roles"]:
html_parts.append(f'<li>{role}</li>')
html_parts.append('</ul>')
html_parts.append('</div>')
# Education
if skills_data.get("education") or skills_data.get("certifications"):
html_parts.append('<div style="margin-bottom: 20px;">')
html_parts.append('<h3 style="color: var(--arc-yellow); margin-bottom: 12px;">🎓 EDUCATION & CERTIFICATIONS</h3>')
if skills_data.get("education"):
html_parts.append('<p style="color: var(--text-main); margin: 8px 0;"><strong>Education:</strong></p>')
html_parts.append('<ul style="color: var(--text-dim); margin-top: 4px;">')
for edu in skills_data["education"]:
html_parts.append(f'<li>{edu}</li>')
html_parts.append('</ul>')
if skills_data.get("certifications"):
html_parts.append('<p style="color: var(--text-main); margin: 8px 0;"><strong>Certifications:</strong></p>')
html_parts.append('<ul style="color: var(--text-dim); margin-top: 4px;">')
for cert in skills_data["certifications"]:
html_parts.append(f'<li>{cert}</li>')
html_parts.append('</ul>')
html_parts.append('</div>')
# Domain Expertise
if skills_data.get("domain_expertise"):
html_parts.append('<div style="margin-bottom: 20px;">')
html_parts.append('<h3 style="color: var(--arc-yellow); margin-bottom: 12px;">🎯 DOMAIN EXPERTISE</h3>')
html_parts.append('<div style="display: flex; flex-wrap: wrap; gap: 8px;">')
for domain in skills_data["domain_expertise"]:
html_parts.append(
f'<span style="background: var(--bg-panel); border: 1px solid var(--border-dim); '
f'padding: 6px 12px; border-radius: 4px; color: var(--arc-orange); font-weight: 600; '
f'font-size: 13px;">{domain}</span>'
)
html_parts.append('</div></div>')
html_parts.append('</div>')
return ''.join(html_parts)
|