Spaces:
Sleeping
Sleeping
Commit ·
492af8a
1
Parent(s): a49ee8d
done
Browse files- enhancer_agent.py +86 -68
- evaluator_agent.py +4 -1
- linkedin_scraper.py +72 -41
- new_resume.py +124 -75
enhancer_agent.py
CHANGED
|
@@ -19,11 +19,10 @@ def enhancer_agent(profile_data, evaluation_feedback, target_role):
|
|
| 19 |
profile_sections = [
|
| 20 |
"Headline",
|
| 21 |
"Summary/About",
|
| 22 |
-
"Current Position/Experience"
|
| 23 |
"Education",
|
| 24 |
"Skills",
|
| 25 |
"Certifications",
|
| 26 |
-
|
| 27 |
"Recommendations",
|
| 28 |
]
|
| 29 |
|
|
@@ -83,6 +82,8 @@ Explanation:
|
|
| 83 |
temperature=0.3
|
| 84 |
)
|
| 85 |
agent_response = chat_completion.choices[0].message.content.strip()
|
|
|
|
|
|
|
| 86 |
except Exception as e:
|
| 87 |
st.error(f"❌ Groq API Error: {str(e)}")
|
| 88 |
if "organization_restricted" in str(e):
|
|
@@ -117,7 +118,8 @@ Explanation:
|
|
| 117 |
else:
|
| 118 |
positions_list = []
|
| 119 |
|
| 120 |
-
|
|
|
|
| 121 |
|
| 122 |
if isinstance(certifications_data, dict):
|
| 123 |
certifications_list = certifications_data.get('certificationHistory', [])
|
|
@@ -129,9 +131,10 @@ Explanation:
|
|
| 129 |
|
| 130 |
if isinstance(educations_data, dict):
|
| 131 |
educations_list = educations_data.get('educationHistory', [])
|
| 132 |
-
education
|
|
|
|
| 133 |
else:
|
| 134 |
-
|
| 135 |
|
| 136 |
if isinstance(recommendations_data, dict):
|
| 137 |
recommendations_count = recommendations_data.get('recommendationsCount', 0)
|
|
@@ -143,8 +146,8 @@ Explanation:
|
|
| 143 |
parsed_data = {
|
| 144 |
'Headline': person.get('headline', '-') or '-',
|
| 145 |
'Summary/About': summary,
|
| 146 |
-
'
|
| 147 |
-
'Education':
|
| 148 |
'Skills': skills_list,
|
| 149 |
'Certifications': certifications_list,
|
| 150 |
'Recommendations': recommendations_count,
|
|
@@ -181,58 +184,64 @@ Explanation:
|
|
| 181 |
if not value or value == '-' or (isinstance(value, list) and not any(value)):
|
| 182 |
st.markdown("_This section is currently missing or incomplete._")
|
| 183 |
else:
|
| 184 |
-
if key == "
|
| 185 |
-
|
| 186 |
-
|
| 187 |
-
|
| 188 |
-
|
| 189 |
-
|
| 190 |
-
|
| 191 |
-
|
| 192 |
-
|
| 193 |
-
|
| 194 |
-
|
| 195 |
-
|
| 196 |
-
|
| 197 |
-
|
| 198 |
-
|
| 199 |
-
|
| 200 |
-
|
| 201 |
-
|
| 202 |
-
|
| 203 |
-
|
| 204 |
-
|
| 205 |
-
|
| 206 |
-
duration = "
|
| 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 |
elif key == "Skills" and isinstance(value, list):
|
| 238 |
st.markdown("\n".join([f"- {skill}" for skill in value]))
|
|
@@ -273,6 +282,8 @@ Enhanced Evaluation Feedback:
|
|
| 273 |
temperature=0.3
|
| 274 |
)
|
| 275 |
enhanced_feedback = chat_completion.choices[0].message.content.strip()
|
|
|
|
|
|
|
| 276 |
return enhanced_feedback
|
| 277 |
except Exception as e:
|
| 278 |
st.error(f"❌ Groq API Error: {str(e)}")
|
|
@@ -298,9 +309,12 @@ Enhanced Evaluation Feedback:
|
|
| 298 |
|
| 299 |
for key in profile_sections:
|
| 300 |
current_content = parsed_profile.get(key, '-')
|
| 301 |
-
if key == "
|
| 302 |
-
|
| 303 |
-
|
|
|
|
|
|
|
|
|
|
| 304 |
recommendation, explanation = get_section_recommendation_and_explanation(key, current_content_str)
|
| 305 |
elif isinstance(current_content, list) and current_content != ['-']:
|
| 306 |
current_content_str = ", ".join(current_content)
|
|
@@ -354,12 +368,14 @@ Enhanced Evaluation Feedback:
|
|
| 354 |
profile_context = ""
|
| 355 |
for key, value in parsed_profile.items():
|
| 356 |
if value != '-' and value:
|
| 357 |
-
if key == "
|
| 358 |
-
|
| 359 |
-
|
| 360 |
-
|
| 361 |
-
|
| 362 |
-
|
|
|
|
|
|
|
| 363 |
elif isinstance(value, list):
|
| 364 |
profile_context += f"\n{key}:\n{', '.join(value)}"
|
| 365 |
elif isinstance(value, int):
|
|
@@ -388,6 +404,8 @@ User's current profile data:
|
|
| 388 |
)
|
| 389 |
|
| 390 |
agent_response = chat_completion.choices[0].message.content.strip()
|
|
|
|
|
|
|
| 391 |
st.session_state['conversation_history'].append({'role': 'assistant', 'content': agent_response})
|
| 392 |
|
| 393 |
display_conversation()
|
|
|
|
| 19 |
profile_sections = [
|
| 20 |
"Headline",
|
| 21 |
"Summary/About",
|
| 22 |
+
"Experience", # Changed from "Current Position/Experience" to "Experience"
|
| 23 |
"Education",
|
| 24 |
"Skills",
|
| 25 |
"Certifications",
|
|
|
|
| 26 |
"Recommendations",
|
| 27 |
]
|
| 28 |
|
|
|
|
| 82 |
temperature=0.3
|
| 83 |
)
|
| 84 |
agent_response = chat_completion.choices[0].message.content.strip()
|
| 85 |
+
# Clean markdown formatting from GPT-OSS response
|
| 86 |
+
agent_response = agent_response.replace('**', '').replace('*', '')
|
| 87 |
except Exception as e:
|
| 88 |
st.error(f"❌ Groq API Error: {str(e)}")
|
| 89 |
if "organization_restricted" in str(e):
|
|
|
|
| 118 |
else:
|
| 119 |
positions_list = []
|
| 120 |
|
| 121 |
+
# Get ALL positions, not just current one
|
| 122 |
+
all_positions = positions_list if positions_list else ['-']
|
| 123 |
|
| 124 |
if isinstance(certifications_data, dict):
|
| 125 |
certifications_list = certifications_data.get('certificationHistory', [])
|
|
|
|
| 131 |
|
| 132 |
if isinstance(educations_data, dict):
|
| 133 |
educations_list = educations_data.get('educationHistory', [])
|
| 134 |
+
# Get ALL education entries, not just the first one
|
| 135 |
+
all_education = educations_list if educations_list else ['-']
|
| 136 |
else:
|
| 137 |
+
all_education = ['-']
|
| 138 |
|
| 139 |
if isinstance(recommendations_data, dict):
|
| 140 |
recommendations_count = recommendations_data.get('recommendationsCount', 0)
|
|
|
|
| 146 |
parsed_data = {
|
| 147 |
'Headline': person.get('headline', '-') or '-',
|
| 148 |
'Summary/About': summary,
|
| 149 |
+
'Experience': all_positions, # Changed from current_position to all_positions
|
| 150 |
+
'Education': all_education, # Changed from single education to all_education
|
| 151 |
'Skills': skills_list,
|
| 152 |
'Certifications': certifications_list,
|
| 153 |
'Recommendations': recommendations_count,
|
|
|
|
| 184 |
if not value or value == '-' or (isinstance(value, list) and not any(value)):
|
| 185 |
st.markdown("_This section is currently missing or incomplete._")
|
| 186 |
else:
|
| 187 |
+
if key == "Experience" and isinstance(value, list) and value != ['-']:
|
| 188 |
+
# Display all experience entries
|
| 189 |
+
for i, position in enumerate(value):
|
| 190 |
+
st.markdown(f"### Experience {i+1}")
|
| 191 |
+
start_end_date = position.get('startEndDate', {})
|
| 192 |
+
start_date = start_end_date.get('start')
|
| 193 |
+
end_date = start_end_date.get('end')
|
| 194 |
+
|
| 195 |
+
if isinstance(start_date, dict):
|
| 196 |
+
start_month = start_date.get('month', '-')
|
| 197 |
+
start_year = start_date.get('year', '-')
|
| 198 |
+
else:
|
| 199 |
+
start_month = '-'
|
| 200 |
+
start_year = '-'
|
| 201 |
+
|
| 202 |
+
if isinstance(end_date, dict):
|
| 203 |
+
end_month = end_date.get('month', '-')
|
| 204 |
+
end_year = end_date.get('year', '-')
|
| 205 |
+
else:
|
| 206 |
+
end_month = 'Present'
|
| 207 |
+
end_year = ''
|
| 208 |
+
|
| 209 |
+
duration = f"{start_month}/{start_year} - {end_month}/{end_year}".strip()
|
| 210 |
+
if duration == "-/- - Present/":
|
| 211 |
+
duration = "Present"
|
| 212 |
+
|
| 213 |
+
position_details = f"**Title:** {position.get('title', '-')}\n\n" \
|
| 214 |
+
f"**Company Name:** {position.get('companyName', '-')}\n\n" \
|
| 215 |
+
f"**Duration:** {duration}\n\n" \
|
| 216 |
+
f"**Description:**\n{position.get('description', '-')}"
|
| 217 |
+
st.markdown(position_details)
|
| 218 |
+
st.markdown("---") # Separator between experiences
|
| 219 |
+
|
| 220 |
+
elif key == "Education" and isinstance(value, list) and value != ['-']:
|
| 221 |
+
# Display all education entries
|
| 222 |
+
for i, education in enumerate(value):
|
| 223 |
+
st.markdown(f"### Education {i+1}")
|
| 224 |
+
start_end_date = education.get('startEndDate', {})
|
| 225 |
+
start_date = start_end_date.get('start')
|
| 226 |
+
end_date = start_end_date.get('end')
|
| 227 |
+
|
| 228 |
+
if isinstance(start_date, dict):
|
| 229 |
+
start_year = start_date.get('year', '-')
|
| 230 |
+
else:
|
| 231 |
+
start_year = '-'
|
| 232 |
+
|
| 233 |
+
if isinstance(end_date, dict):
|
| 234 |
+
end_year = end_date.get('year', '-')
|
| 235 |
+
else:
|
| 236 |
+
end_year = '-'
|
| 237 |
+
|
| 238 |
+
duration = f"{start_year} - {end_year}".strip()
|
| 239 |
+
education_details = f"**School Name:** {education.get('schoolName', '-')}\n\n" \
|
| 240 |
+
f"**Degree Name:** {education.get('degreeName', '-')}\n\n" \
|
| 241 |
+
f"**Field of Study:** {education.get('fieldOfStudy', '-')}\n\n" \
|
| 242 |
+
f"**Duration:** {duration}"
|
| 243 |
+
st.markdown(education_details)
|
| 244 |
+
st.markdown("---") # Separator between education entries
|
| 245 |
|
| 246 |
elif key == "Skills" and isinstance(value, list):
|
| 247 |
st.markdown("\n".join([f"- {skill}" for skill in value]))
|
|
|
|
| 282 |
temperature=0.3
|
| 283 |
)
|
| 284 |
enhanced_feedback = chat_completion.choices[0].message.content.strip()
|
| 285 |
+
# Clean markdown formatting from GPT-OSS response
|
| 286 |
+
enhanced_feedback = enhanced_feedback.replace('**', '').replace('*', '')
|
| 287 |
return enhanced_feedback
|
| 288 |
except Exception as e:
|
| 289 |
st.error(f"❌ Groq API Error: {str(e)}")
|
|
|
|
| 309 |
|
| 310 |
for key in profile_sections:
|
| 311 |
current_content = parsed_profile.get(key, '-')
|
| 312 |
+
if key == "Experience" and isinstance(current_content, list) and current_content != ['-']:
|
| 313 |
+
# Handle multiple experience entries
|
| 314 |
+
experience_strs = []
|
| 315 |
+
for position in current_content:
|
| 316 |
+
experience_strs.append(f"Title: {position.get('title', '-')}\nCompany Name: {position.get('companyName', '-')}\nDescription: {position.get('description', '-')}")
|
| 317 |
+
current_content_str = "\n\n".join(experience_strs)
|
| 318 |
recommendation, explanation = get_section_recommendation_and_explanation(key, current_content_str)
|
| 319 |
elif isinstance(current_content, list) and current_content != ['-']:
|
| 320 |
current_content_str = ", ".join(current_content)
|
|
|
|
| 368 |
profile_context = ""
|
| 369 |
for key, value in parsed_profile.items():
|
| 370 |
if value != '-' and value:
|
| 371 |
+
if key == "Experience" and isinstance(value, list) and value != ['-']:
|
| 372 |
+
experience_details = []
|
| 373 |
+
for position in value:
|
| 374 |
+
position_detail = f"Title: {position.get('title', '-')}\n" \
|
| 375 |
+
f"Company Name: {position.get('companyName', '-')}\n" \
|
| 376 |
+
f"Description: {position.get('description', '-')}"
|
| 377 |
+
experience_details.append(position_detail)
|
| 378 |
+
profile_context += f"\n{key}:\n" + "\n\n".join(experience_details)
|
| 379 |
elif isinstance(value, list):
|
| 380 |
profile_context += f"\n{key}:\n{', '.join(value)}"
|
| 381 |
elif isinstance(value, int):
|
|
|
|
| 404 |
)
|
| 405 |
|
| 406 |
agent_response = chat_completion.choices[0].message.content.strip()
|
| 407 |
+
# Clean markdown formatting from GPT-OSS response
|
| 408 |
+
agent_response = agent_response.replace('**', '').replace('*', '')
|
| 409 |
st.session_state['conversation_history'].append({'role': 'assistant', 'content': agent_response})
|
| 410 |
|
| 411 |
display_conversation()
|
evaluator_agent.py
CHANGED
|
@@ -40,7 +40,10 @@ def evaluate_linkedin_profile(profile_data):
|
|
| 40 |
max_tokens=30000,
|
| 41 |
temperature=0.5
|
| 42 |
)
|
| 43 |
-
|
|
|
|
|
|
|
|
|
|
| 44 |
except Exception as e:
|
| 45 |
st.error(f"❌ Groq API Error: {str(e)}")
|
| 46 |
if "organization_restricted" in str(e):
|
|
|
|
| 40 |
max_tokens=30000,
|
| 41 |
temperature=0.5
|
| 42 |
)
|
| 43 |
+
# Clean markdown formatting from GPT-OSS response
|
| 44 |
+
content = completion.choices[0].message.content
|
| 45 |
+
content = content.replace('**', '').replace('*', '') # Remove markdown formatting
|
| 46 |
+
return content
|
| 47 |
except Exception as e:
|
| 48 |
st.error(f"❌ Groq API Error: {str(e)}")
|
| 49 |
if "organization_restricted" in str(e):
|
linkedin_scraper.py
CHANGED
|
@@ -15,8 +15,18 @@ class LinkedInScraper:
|
|
| 15 |
"""
|
| 16 |
|
| 17 |
def __init__(self):
|
| 18 |
-
|
| 19 |
-
self.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 20 |
self.headers = {
|
| 21 |
"Content-Type": "application/json"
|
| 22 |
}
|
|
@@ -34,39 +44,22 @@ class LinkedInScraper:
|
|
| 34 |
except:
|
| 35 |
return False
|
| 36 |
|
| 37 |
-
def
|
| 38 |
"""
|
| 39 |
-
|
| 40 |
-
|
| 41 |
-
Args:
|
| 42 |
-
linkedin_url (str): LinkedIn profile URL
|
| 43 |
-
|
| 44 |
-
Returns:
|
| 45 |
-
Dict containing profile data or error information
|
| 46 |
"""
|
| 47 |
-
|
| 48 |
-
# Validate URL
|
| 49 |
-
if not self.is_valid_linkedin_url(linkedin_url):
|
| 50 |
-
return {
|
| 51 |
-
"success": False,
|
| 52 |
-
"error": "Invalid LinkedIn URL format",
|
| 53 |
-
"url": linkedin_url
|
| 54 |
-
}
|
| 55 |
-
|
| 56 |
try:
|
| 57 |
-
# Prepare request
|
| 58 |
payload = {"url": linkedin_url}
|
| 59 |
-
full_url = f"{
|
| 60 |
|
| 61 |
-
print(f"[
|
| 62 |
start_time = time.time()
|
| 63 |
|
| 64 |
-
# Make API request
|
| 65 |
response = requests.post(
|
| 66 |
full_url,
|
| 67 |
headers=self.headers,
|
| 68 |
data=json.dumps(payload),
|
| 69 |
-
timeout=60
|
| 70 |
)
|
| 71 |
|
| 72 |
end_time = time.time()
|
|
@@ -75,11 +68,15 @@ class LinkedInScraper:
|
|
| 75 |
if response.status_code == 200:
|
| 76 |
data = response.json()
|
| 77 |
|
| 78 |
-
#
|
|
|
|
| 79 |
if 'linkedin_full_data' in data:
|
| 80 |
profile_data = data['linkedin_full_data']
|
| 81 |
-
|
| 82 |
-
|
|
|
|
|
|
|
|
|
|
| 83 |
print(f" Name: {profile_data.get('full_name', 'N/A')}")
|
| 84 |
print(f" Headline: {profile_data.get('headline', 'N/A')}")
|
| 85 |
print(f" Location: {profile_data.get('location', 'N/A')}")
|
|
@@ -88,19 +85,20 @@ class LinkedInScraper:
|
|
| 88 |
"success": True,
|
| 89 |
"data": profile_data,
|
| 90 |
"scrape_time": duration,
|
| 91 |
-
"url": linkedin_url
|
|
|
|
| 92 |
}
|
| 93 |
else:
|
| 94 |
return {
|
| 95 |
"success": False,
|
| 96 |
-
"error": "No profile data returned
|
| 97 |
"raw_response": data,
|
| 98 |
"url": linkedin_url
|
| 99 |
}
|
| 100 |
else:
|
| 101 |
return {
|
| 102 |
"success": False,
|
| 103 |
-
"error": f"API
|
| 104 |
"response_text": response.text,
|
| 105 |
"url": linkedin_url
|
| 106 |
}
|
|
@@ -108,27 +106,60 @@ class LinkedInScraper:
|
|
| 108 |
except requests.exceptions.Timeout:
|
| 109 |
return {
|
| 110 |
"success": False,
|
| 111 |
-
"error": "Request timed out after 60 seconds",
|
| 112 |
-
"url": linkedin_url
|
| 113 |
-
}
|
| 114 |
-
except requests.exceptions.RequestException as e:
|
| 115 |
-
return {
|
| 116 |
-
"success": False,
|
| 117 |
-
"error": f"Request failed: {str(e)}",
|
| 118 |
"url": linkedin_url
|
| 119 |
}
|
| 120 |
-
except
|
| 121 |
return {
|
| 122 |
"success": False,
|
| 123 |
-
"error": f"
|
| 124 |
"url": linkedin_url
|
| 125 |
}
|
| 126 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 127 |
return {
|
| 128 |
"success": False,
|
| 129 |
-
"error":
|
| 130 |
"url": linkedin_url
|
| 131 |
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 132 |
|
| 133 |
def extract_key_info(self, profile_data: Dict[str, Any]) -> Dict[str, Any]:
|
| 134 |
"""
|
|
|
|
| 15 |
"""
|
| 16 |
|
| 17 |
def __init__(self):
|
| 18 |
+
# Primary API (original)
|
| 19 |
+
self.primary_api = {
|
| 20 |
+
"url": "https://api-f1db6c.stack.tryrelevance.com/latest/studios/11116e42-9be9-4837-8753-c46a80458318/trigger_webhook",
|
| 21 |
+
"project_id": "f56ec267-8285-4bef-b8ab-4dce36204e5d"
|
| 22 |
+
}
|
| 23 |
+
|
| 24 |
+
# Fallback API (new account)
|
| 25 |
+
self.fallback_api = {
|
| 26 |
+
"url": "https://api-f1db6c.stack.tryrelevance.com/latest/studios/a1a00cf9-4102-4d76-99e5-8ce9b922b51c/trigger_webhook",
|
| 27 |
+
"project_id": "e5f9ef92-aa24-4626-a145-3fb746186504"
|
| 28 |
+
}
|
| 29 |
+
|
| 30 |
self.headers = {
|
| 31 |
"Content-Type": "application/json"
|
| 32 |
}
|
|
|
|
| 44 |
except:
|
| 45 |
return False
|
| 46 |
|
| 47 |
+
def _try_api(self, api_config: Dict[str, str], linkedin_url: str, api_name: str) -> Dict[str, Any]:
|
| 48 |
"""
|
| 49 |
+
Try scraping with a specific API configuration
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 50 |
"""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 51 |
try:
|
|
|
|
| 52 |
payload = {"url": linkedin_url}
|
| 53 |
+
full_url = f"{api_config['url']}?project={api_config['project_id']}"
|
| 54 |
|
| 55 |
+
print(f"[{api_name}] Trying to scrape: {linkedin_url}")
|
| 56 |
start_time = time.time()
|
| 57 |
|
|
|
|
| 58 |
response = requests.post(
|
| 59 |
full_url,
|
| 60 |
headers=self.headers,
|
| 61 |
data=json.dumps(payload),
|
| 62 |
+
timeout=60
|
| 63 |
)
|
| 64 |
|
| 65 |
end_time = time.time()
|
|
|
|
| 68 |
if response.status_code == 200:
|
| 69 |
data = response.json()
|
| 70 |
|
| 71 |
+
# Handle different response formats from different APIs
|
| 72 |
+
profile_data = None
|
| 73 |
if 'linkedin_full_data' in data:
|
| 74 |
profile_data = data['linkedin_full_data']
|
| 75 |
+
elif 'data' in data:
|
| 76 |
+
profile_data = data['data']
|
| 77 |
+
|
| 78 |
+
if profile_data:
|
| 79 |
+
print(f"[{api_name}] SUCCESS in {duration}s")
|
| 80 |
print(f" Name: {profile_data.get('full_name', 'N/A')}")
|
| 81 |
print(f" Headline: {profile_data.get('headline', 'N/A')}")
|
| 82 |
print(f" Location: {profile_data.get('location', 'N/A')}")
|
|
|
|
| 85 |
"success": True,
|
| 86 |
"data": profile_data,
|
| 87 |
"scrape_time": duration,
|
| 88 |
+
"url": linkedin_url,
|
| 89 |
+
"api_used": api_name
|
| 90 |
}
|
| 91 |
else:
|
| 92 |
return {
|
| 93 |
"success": False,
|
| 94 |
+
"error": f"{api_name}: No profile data returned",
|
| 95 |
"raw_response": data,
|
| 96 |
"url": linkedin_url
|
| 97 |
}
|
| 98 |
else:
|
| 99 |
return {
|
| 100 |
"success": False,
|
| 101 |
+
"error": f"{api_name}: API returned status {response.status_code}",
|
| 102 |
"response_text": response.text,
|
| 103 |
"url": linkedin_url
|
| 104 |
}
|
|
|
|
| 106 |
except requests.exceptions.Timeout:
|
| 107 |
return {
|
| 108 |
"success": False,
|
| 109 |
+
"error": f"{api_name}: Request timed out after 60 seconds",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 110 |
"url": linkedin_url
|
| 111 |
}
|
| 112 |
+
except Exception as e:
|
| 113 |
return {
|
| 114 |
"success": False,
|
| 115 |
+
"error": f"{api_name}: {str(e)}",
|
| 116 |
"url": linkedin_url
|
| 117 |
}
|
| 118 |
+
|
| 119 |
+
def scrape_profile(self, linkedin_url: str) -> Dict[str, Any]:
|
| 120 |
+
"""
|
| 121 |
+
Scrape a LinkedIn profile using primary API with fallback
|
| 122 |
+
|
| 123 |
+
Args:
|
| 124 |
+
linkedin_url (str): LinkedIn profile URL
|
| 125 |
+
|
| 126 |
+
Returns:
|
| 127 |
+
Dict containing profile data or error information
|
| 128 |
+
"""
|
| 129 |
+
|
| 130 |
+
# Validate URL
|
| 131 |
+
if not self.is_valid_linkedin_url(linkedin_url):
|
| 132 |
return {
|
| 133 |
"success": False,
|
| 134 |
+
"error": "Invalid LinkedIn URL format",
|
| 135 |
"url": linkedin_url
|
| 136 |
}
|
| 137 |
+
|
| 138 |
+
print(f"[SCRAPING] LinkedIn profile: {linkedin_url}")
|
| 139 |
+
|
| 140 |
+
# Try primary API first
|
| 141 |
+
result = self._try_api(self.primary_api, linkedin_url, "PRIMARY")
|
| 142 |
+
|
| 143 |
+
if result["success"]:
|
| 144 |
+
return result
|
| 145 |
+
|
| 146 |
+
print(f"[FALLBACK] Primary API failed: {result['error']}")
|
| 147 |
+
print(f"[FALLBACK] Trying secondary API...")
|
| 148 |
+
|
| 149 |
+
# Try fallback API
|
| 150 |
+
result = self._try_api(self.fallback_api, linkedin_url, "FALLBACK")
|
| 151 |
+
|
| 152 |
+
if result["success"]:
|
| 153 |
+
return result
|
| 154 |
+
|
| 155 |
+
# Both APIs failed
|
| 156 |
+
print(f"[FAILED] Both APIs failed!")
|
| 157 |
+
return {
|
| 158 |
+
"success": False,
|
| 159 |
+
"error": "Both primary and fallback APIs failed",
|
| 160 |
+
"primary_error": result.get('error', 'Unknown error'),
|
| 161 |
+
"url": linkedin_url
|
| 162 |
+
}
|
| 163 |
|
| 164 |
def extract_key_info(self, profile_data: Dict[str, Any]) -> Dict[str, Any]:
|
| 165 |
"""
|
new_resume.py
CHANGED
|
@@ -13,22 +13,52 @@ def generate_latex_resume(personal_info, education_list, experience_list, skills
|
|
| 13 |
"""Generate clean LaTeX code for resume"""
|
| 14 |
|
| 15 |
def clean_text(text):
|
| 16 |
-
"""Clean text for LaTeX -
|
| 17 |
if not text:
|
| 18 |
return ""
|
| 19 |
-
|
| 20 |
-
text = str(text)
|
| 21 |
-
|
| 22 |
-
|
|
|
|
|
|
|
|
|
|
| 23 |
text = text.replace('$', '\\$')
|
|
|
|
|
|
|
| 24 |
text = text.replace('#', '\\#')
|
|
|
|
| 25 |
text = text.replace('_', '\\_')
|
| 26 |
-
text = text.replace('
|
| 27 |
-
|
| 28 |
-
#
|
| 29 |
-
text = text.replace('
|
| 30 |
-
text = text.replace('
|
| 31 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 32 |
return text
|
| 33 |
|
| 34 |
# Get clean data
|
|
@@ -328,88 +358,98 @@ def parse_uploaded_resume(uploaded_file):
|
|
| 328 |
return None, f"Error parsing file: {str(e)}"
|
| 329 |
|
| 330 |
def structure_resume_content(raw_content):
|
| 331 |
-
"""Use AI to structure raw resume content into organized sections"""
|
| 332 |
try:
|
| 333 |
client = Groq(api_key=Config.GROQ_API_KEY)
|
| 334 |
|
| 335 |
-
prompt = f"""
|
| 336 |
-
You are an AI assistant that extracts and structures resume information from raw text.
|
| 337 |
-
Parse the following resume content and extract information into a structured format.
|
| 338 |
|
| 339 |
-
Raw Resume Content:
|
| 340 |
{raw_content}
|
| 341 |
|
| 342 |
-
|
| 343 |
{{
|
| 344 |
-
|
| 345 |
-
|
| 346 |
-
|
| 347 |
-
|
| 348 |
-
|
| 349 |
-
|
| 350 |
-
|
| 351 |
-
|
| 352 |
-
|
| 353 |
-
|
| 354 |
-
|
| 355 |
-
|
| 356 |
-
|
| 357 |
-
|
| 358 |
-
|
| 359 |
-
|
| 360 |
-
|
| 361 |
-
|
| 362 |
-
|
| 363 |
-
|
| 364 |
-
|
| 365 |
-
|
| 366 |
-
|
| 367 |
-
|
| 368 |
-
|
| 369 |
-
|
| 370 |
-
|
| 371 |
-
"certifications": ["cert1", "cert2"],
|
| 372 |
-
"projects": [
|
| 373 |
-
{{
|
| 374 |
-
"project_name": "project name",
|
| 375 |
-
"description": "project description"
|
| 376 |
-
}}
|
| 377 |
-
],
|
| 378 |
-
"languages": ["language1", "language2"],
|
| 379 |
-
"hobbies": ["hobby1", "hobby2"]
|
| 380 |
}}
|
| 381 |
|
| 382 |
-
|
| 383 |
-
-
|
| 384 |
-
-
|
| 385 |
-
-
|
| 386 |
-
-
|
| 387 |
-
"""
|
| 388 |
|
| 389 |
response = client.chat.completions.create(
|
| 390 |
-
messages=[
|
|
|
|
|
|
|
|
|
|
| 391 |
model="openai/gpt-oss-120b",
|
| 392 |
-
temperature=0.
|
|
|
|
| 393 |
)
|
| 394 |
|
| 395 |
-
#
|
| 396 |
-
|
| 397 |
-
|
| 398 |
|
| 399 |
-
#
|
| 400 |
-
|
| 401 |
-
|
| 402 |
|
| 403 |
-
|
| 404 |
-
|
| 405 |
-
|
| 406 |
-
|
| 407 |
-
else:
|
| 408 |
-
raise ValueError("No valid JSON found in response")
|
| 409 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 410 |
except Exception as e:
|
| 411 |
-
st.error(f"
|
| 412 |
-
return
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 413 |
|
| 414 |
def enhance_resume_with_ai(structured_data, target_role):
|
| 415 |
"""Use AI to enhance the structured resume data for the target role"""
|
|
@@ -451,6 +491,9 @@ Enhancement guidelines:
|
|
| 451 |
|
| 452 |
response_content = response.choices[0].message.content.strip()
|
| 453 |
|
|
|
|
|
|
|
|
|
|
| 454 |
# Extract enhanced data and summary
|
| 455 |
if "ENHANCED_DATA:" in response_content and "ENHANCEMENT_SUMMARY:" in response_content:
|
| 456 |
data_part = response_content.split("ENHANCED_DATA:")[1].split("ENHANCEMENT_SUMMARY:")[0].strip()
|
|
@@ -556,6 +599,12 @@ def display_editable_resume_data(structured_data):
|
|
| 556 |
st.markdown(f"**Project #{i+1}**")
|
| 557 |
proj_data = existing_projects[i] if i < len(existing_projects) else {}
|
| 558 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 559 |
project = {
|
| 560 |
'project_name': st.text_input(f"Project Name #{i+1}", value=proj_data.get('project_name', ''), key=f"edit_proj_name_{i}"),
|
| 561 |
'description': st.text_area(f"Project Description #{i+1}", value=proj_data.get('description', ''), key=f"edit_proj_desc_{i}")
|
|
|
|
| 13 |
"""Generate clean LaTeX code for resume"""
|
| 14 |
|
| 15 |
def clean_text(text):
|
| 16 |
+
"""Clean text for LaTeX - PRODUCTION READY VERSION"""
|
| 17 |
if not text:
|
| 18 |
return ""
|
| 19 |
+
|
| 20 |
+
text = str(text).strip()
|
| 21 |
+
|
| 22 |
+
# Handle special LaTeX characters in EXACT correct order
|
| 23 |
+
text = text.replace('\\', '\\textbackslash{}') # MUST be first!
|
| 24 |
+
text = text.replace('{', '\\{')
|
| 25 |
+
text = text.replace('}', '\\}')
|
| 26 |
text = text.replace('$', '\\$')
|
| 27 |
+
text = text.replace('&', '\\&')
|
| 28 |
+
text = text.replace('%', '\\%')
|
| 29 |
text = text.replace('#', '\\#')
|
| 30 |
+
text = text.replace('^', '\\textasciicircum{}')
|
| 31 |
text = text.replace('_', '\\_')
|
| 32 |
+
text = text.replace('~', '\\textasciitilde{}')
|
| 33 |
+
|
| 34 |
+
# Handle quotes properly for LaTeX
|
| 35 |
+
text = text.replace('"', "''")
|
| 36 |
+
text = text.replace('`', "'") # Prevent backtick issues
|
| 37 |
+
|
| 38 |
+
# Fix common Unicode characters that break LaTeX
|
| 39 |
+
unicode_fixes = {
|
| 40 |
+
'\u202f': ' ', # Narrow no-break space
|
| 41 |
+
'\u2013': '--', # En dash
|
| 42 |
+
'\u2014': '---', # Em dash
|
| 43 |
+
'\u2019': "'", # Right single quotation mark
|
| 44 |
+
'\u201c': '``', # Left double quotation mark
|
| 45 |
+
'\u201d': "''", # Right double quotation mark
|
| 46 |
+
'\u2026': '...', # Horizontal ellipsis
|
| 47 |
+
'\u00a0': ' ', # Non-breaking space
|
| 48 |
+
'\u2010': '-', # Hyphen
|
| 49 |
+
'\u2011': '-', # Non-breaking hyphen
|
| 50 |
+
'\u2012': '-', # Figure dash
|
| 51 |
+
}
|
| 52 |
+
|
| 53 |
+
for unicode_char, replacement in unicode_fixes.items():
|
| 54 |
+
text = text.replace(unicode_char, replacement)
|
| 55 |
+
|
| 56 |
+
# Convert to ASCII to remove any remaining problematic characters
|
| 57 |
+
text = text.encode('ascii', 'ignore').decode('ascii')
|
| 58 |
+
|
| 59 |
+
# Final cleanup - remove multiple spaces and trim
|
| 60 |
+
text = ' '.join(text.split())
|
| 61 |
+
|
| 62 |
return text
|
| 63 |
|
| 64 |
# Get clean data
|
|
|
|
| 358 |
return None, f"Error parsing file: {str(e)}"
|
| 359 |
|
| 360 |
def structure_resume_content(raw_content):
|
| 361 |
+
"""Use AI to structure raw resume content into organized sections - IMPROVED VERSION"""
|
| 362 |
try:
|
| 363 |
client = Groq(api_key=Config.GROQ_API_KEY)
|
| 364 |
|
| 365 |
+
prompt = f"""Extract information from this resume and return ONLY a JSON object:
|
|
|
|
|
|
|
| 366 |
|
|
|
|
| 367 |
{raw_content}
|
| 368 |
|
| 369 |
+
Return exactly this structure with real data from the resume:
|
| 370 |
{{
|
| 371 |
+
"personal_info": {{
|
| 372 |
+
"full_name": "",
|
| 373 |
+
"email": "",
|
| 374 |
+
"phone": "",
|
| 375 |
+
"address": "",
|
| 376 |
+
"summary": ""
|
| 377 |
+
}},
|
| 378 |
+
"experience": [
|
| 379 |
+
{{
|
| 380 |
+
"job_title": "",
|
| 381 |
+
"company": "",
|
| 382 |
+
"start_date": "",
|
| 383 |
+
"end_date": "",
|
| 384 |
+
"responsibilities": []
|
| 385 |
+
}}
|
| 386 |
+
],
|
| 387 |
+
"education": [
|
| 388 |
+
{{
|
| 389 |
+
"degree": "",
|
| 390 |
+
"institution": "",
|
| 391 |
+
"graduation_year": "",
|
| 392 |
+
"field": ""
|
| 393 |
+
}}
|
| 394 |
+
],
|
| 395 |
+
"skills": [],
|
| 396 |
+
"projects": [],
|
| 397 |
+
"certifications": []
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 398 |
}}
|
| 399 |
|
| 400 |
+
Rules:
|
| 401 |
+
- Return ONLY valid JSON, no markdown or explanations
|
| 402 |
+
- Use empty string "" for missing text fields
|
| 403 |
+
- Use empty array [] for missing list fields
|
| 404 |
+
- Extract real data only, don't invent anything"""
|
|
|
|
| 405 |
|
| 406 |
response = client.chat.completions.create(
|
| 407 |
+
messages=[
|
| 408 |
+
{"role": "system", "content": "Extract resume data and return only valid JSON. No explanations, no markdown formatting."},
|
| 409 |
+
{"role": "user", "content": prompt}
|
| 410 |
+
],
|
| 411 |
model="openai/gpt-oss-120b",
|
| 412 |
+
temperature=0.0,
|
| 413 |
+
max_tokens=1500
|
| 414 |
)
|
| 415 |
|
| 416 |
+
# Clean response
|
| 417 |
+
content = response.choices[0].message.content.strip()
|
| 418 |
+
content = content.replace('```json', '').replace('```', '').strip()
|
| 419 |
|
| 420 |
+
# Remove markdown formatting that GPT-OSS sometimes adds
|
| 421 |
+
content = content.replace('**', '') # Remove bold formatting
|
| 422 |
+
content = content.replace('*', '') # Remove italic formatting
|
| 423 |
|
| 424 |
+
# Parse JSON
|
| 425 |
+
import json
|
| 426 |
+
data = json.loads(content)
|
| 427 |
+
return data
|
|
|
|
|
|
|
| 428 |
|
| 429 |
+
except json.JSONDecodeError as e:
|
| 430 |
+
st.error(f"JSON parsing error: {str(e)}")
|
| 431 |
+
st.error("AI returned invalid JSON format")
|
| 432 |
+
return create_empty_resume_structure()
|
| 433 |
except Exception as e:
|
| 434 |
+
st.error(f"Extraction error: {str(e)}")
|
| 435 |
+
return create_empty_resume_structure()
|
| 436 |
+
|
| 437 |
+
def create_empty_resume_structure():
|
| 438 |
+
"""Create empty resume structure for fallback"""
|
| 439 |
+
return {
|
| 440 |
+
"personal_info": {
|
| 441 |
+
"full_name": "",
|
| 442 |
+
"email": "",
|
| 443 |
+
"phone": "",
|
| 444 |
+
"address": "",
|
| 445 |
+
"summary": ""
|
| 446 |
+
},
|
| 447 |
+
"experience": [],
|
| 448 |
+
"education": [],
|
| 449 |
+
"skills": [],
|
| 450 |
+
"projects": [],
|
| 451 |
+
"certifications": []
|
| 452 |
+
}
|
| 453 |
|
| 454 |
def enhance_resume_with_ai(structured_data, target_role):
|
| 455 |
"""Use AI to enhance the structured resume data for the target role"""
|
|
|
|
| 491 |
|
| 492 |
response_content = response.choices[0].message.content.strip()
|
| 493 |
|
| 494 |
+
# Clean markdown formatting from response
|
| 495 |
+
response_content = response_content.replace('**', '').replace('*', '')
|
| 496 |
+
|
| 497 |
# Extract enhanced data and summary
|
| 498 |
if "ENHANCED_DATA:" in response_content and "ENHANCEMENT_SUMMARY:" in response_content:
|
| 499 |
data_part = response_content.split("ENHANCED_DATA:")[1].split("ENHANCEMENT_SUMMARY:")[0].strip()
|
|
|
|
| 599 |
st.markdown(f"**Project #{i+1}**")
|
| 600 |
proj_data = existing_projects[i] if i < len(existing_projects) else {}
|
| 601 |
|
| 602 |
+
# Handle case where project data might be a string instead of dict
|
| 603 |
+
if isinstance(proj_data, str):
|
| 604 |
+
proj_data = {'project_name': proj_data, 'description': ''}
|
| 605 |
+
elif not isinstance(proj_data, dict):
|
| 606 |
+
proj_data = {'project_name': '', 'description': ''}
|
| 607 |
+
|
| 608 |
project = {
|
| 609 |
'project_name': st.text_input(f"Project Name #{i+1}", value=proj_data.get('project_name', ''), key=f"edit_proj_name_{i}"),
|
| 610 |
'description': st.text_area(f"Project Description #{i+1}", value=proj_data.get('description', ''), key=f"edit_proj_desc_{i}")
|