import os import time import random import re import json import requests from langchain_community.embeddings import OllamaEmbeddings from langchain_community.vectorstores import FAISS from langchain_core.prompts import ChatPromptTemplate from crawl4ai import AsyncWebCrawler import asyncio from langchain_groq import ChatGroq from langchain_core.prompts import PromptTemplate from langchain_core.output_parsers import StrOutputParser, PydanticOutputParser from langchain_core.runnables import RunnablePassthrough from langchain_community.document_loaders import PyPDFLoader, TextLoader, Docx2txtLoader, UnstructuredPowerPointLoader from PyPDF2 import PdfReader from langchain_experimental.text_splitter import SemanticChunker from pydantic import BaseModel, Field from typing import List, Optional, Dict from urllib.parse import urlparse from langchain_ollama import ChatOllama import shutil GROQ_API_KEY = "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" OLLAMA_MODEL = "nomic-embed-text" def map_skills_to_designations(skills_db, JobProfile, model_name="mixtral-8x7b-32768", temperature=0): # Initialize the chat model #chat = ChatGroq(temperature=temperature, groq_api_key=groq_api_key, model_name=model_name) chat = ChatOllama( base_url = 'http://127.0.0.1:11434', model = "llama3:8b" #model = "deepseek-r1:8b" ) #print(chat) # Define the system prompt system_prompt = PromptTemplate( input_variables=["skills_db", "JobProfile"], template=''' You are an expert in skill-to-Job Profile mapping. Your task is to analyze a given set of skills and map them to the most appropriate JobProfile. ### Input Skills: {skills_db} ### Designations: {JobProfile} ### Output Format: {{ "JobProfile 1": ["Skill 1", "Skill 2", "Skill 3"], "JobProfile 2": ["Skill 4", "Skill 5"], "JobProfile 3": ["Skill 6", "Skill 7", "Skill 8"] }} ### Notes: - Ensure that each skill is mapped only to the most relevant JobProfile(s). - If a skill fits multiple JobProfile, include it in both. - Use clear and concise mappings. - Respond only with the JSON object in the specified format, without any additional text. - If you found no skills for some JobProfile check does JobProfile at last one more time ### Instructions: - **Respond only with the JSON object in the specified format.** - **Do not include any additional text, explanations, or code fences (e.g., ```json).** - **Ensure that the JSON is properly formatted and valid.** ''' ) # Format the prompt with the input skills and designations formatted_prompt = system_prompt.format(skills_db=skills_db, JobProfile=JobProfile) #print(formatted_prompt) # Send the formatted prompt to ChatGroq try: output = chat.invoke(formatted_prompt) total_tokens = output.usage_metadata["output_tokens"] #token_consumed = output.response_metadata['token_usage']['total_tokens'] skill_to_jobProfile_mapping = output.content except Exception as e: print(f"An error occurred during chat invocation: {e}") return None, 0 return skill_to_jobProfile_mapping, total_tokens import json import re def extract_json_from_response(response_text): """Extracts the JSON object from a response string""" json_match = re.search(r"```json\s*(\{.*?\})\s*```", response_text, re.DOTALL) if json_match: return json_match.group(1) # Extract the JSON part only return None # Return None if no JSON found def map_skills_to_designations_non_empty(skills_db, JobProfile, mapped_skills): # Initialize the chat model #chat = ChatGroq(temperature=temperature, groq_api_key=groq_api_key, model_name=model_name) chat = ChatOllama( base_url = 'http://127.0.0.1:11434', model = "llama3:8b" #model = "deepseek-r1:8b" ) # Convert mapped_skills dictionary to a JSON string formatted_mapped_skills = json.dumps(mapped_skills, ensure_ascii=False) # Define the system prompt system_prompt = PromptTemplate( input_variables=["skills_db", "JobProfile", "mapped_skills"], template=''' You are an expert in skill-to-JobProfile mapping. Your task is to validate and refine a given mapping of skills to JobProfile to ensure accuracy, completeness, and clarity. ### Input: - **Existing Mapped Skills** (JSON format): {mapped_skills} ### Input Skills: {skills_db} ### JobProfile: {JobProfile} ### Output Format: {{ "JobProfile 1": ["Skill 1", "Skill 2", "Skill 3"], "JobProfile 2": ["Skill 4", "Skill 5"], "JobProfile 3": ["Skill 6", "Skill 7", "Skill 8"] }} ### Validation & Enhancement Guidelines: 1. **Ensure Relevance:** - Verify that each skill is appropriately mapped to the correct JobProfile(s). - Remove any incorrectly assigned skills. 2. **Check for Completeness:** - Identify and add missing skills relevant to each JobProfile. - Ensure no important skills are overlooked. 3. **Refine Overlapping Skills:** - If a skill belongs to multiple JobProfile, ensure it is correctly included in each. - Avoid redundant mappings unless justified. 4. **Improve Consistency & Clarity:** - Standardize skill naming to avoid duplicates with different names. - Use precise and industry-recognized terminology. 5. **Handle Unmapped JobProfile:** - If a JobProfile lacks sufficient mapped skills, reevaluate and attempt to match relevant ones. 6. **Respond in Valid JSON Format:** - **ONLY return the JSON object without any explanations, comments, or additional text.** - **DO NOT wrap the JSON in code blocks (```json ... ```).** ''' ) # Format the prompt safely formatted_prompt = system_prompt.format( skills_db=skills_db.replace("{", "{{").replace("}", "}}"), JobProfile=JobProfile.replace("{", "{{").replace("}", "}}"), mapped_skills=formatted_mapped_skills ) # Send the formatted prompt to ChatGroq try: output = chat.invoke(formatted_prompt) # Debugging: Print raw API response print(f"Raw API response: {output.content}") # Extract the JSON part from the response # json_data = extract_json_from_response(output.content) # if not json_data: # print("Error: No valid JSON found in the response") # return None, 0 # Ensure response is valid JSON skill_to_JobProfile_mapping = json.loads(json_data) # Convert response to dict #token_consumed = output.response_metadata['token_usage']['total_tokens'] token_consumed = output.usage_metadata["output_tokens"] except json.JSONDecodeError as e: print(f"Error decoding JSON response: {e}") return None, 0 except Exception as e: print(f"An error occurred during chat invocation: {e}") return None, 0 return skill_to_JobProfile_mapping, token_consumed # Example usage if __name__ == "__main__": GROQ_API_KEY = "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" # Replace with your actual Groq API key industry = "Infrastructure" #skill_tree,token_consumed = generate_skill_tree(prompt_text=industry, groq_api_key=GROQ_API_KEY) skillset =''' Budgeting, Cost Estimation, Civil Engineering, Structural Analysis, AutoCAD, Revit, Communication, Leadership, Teamwork, Urban Design, Water Supply and Distribution , SQL, C++ , Management, Housekeeping Management,Front Desk Management,Complaint Handling,Recruitment,Training & Development ''' JobProfile = ''' HR, Developer, Manager, Planner,CEO, Infrastructure Deverloper ''' mapped = { "Manager": [ "Time Management", "Decision Making", "Industry Trends", "Sector Expertise", "Business Acumen", "Stakeholder Communication", "Risk Management", "Agile Methodologies", "Leadership", "Teamwork", "Adaptability" ], "Engineer": [ "Data Encryption", "Incident Response", "Cloud Architecture", "Cloud Migration", "Cloud Monitoring", "Network Security", "Programming Languages", "Py" ], "CEO": [ "Negotiation", "Empathy", "Leadership", "Risk Management", "Business Acumen", "Sector Expertise", "Industry Trends", "Stakeholder Communication", "Creativity" ], "CO-Founder": [ "Creativity", "Sector Expertise", "Empathy", "Leadership", "Stakeholder Communication", "Risk Management", "Business Acumen", "Industry Trends", "Negotiation" ], "Developer": [ "Py", "Programming Languages", "Software Testing", "Version Control" ], "Senior Developer": [ "Network Security", "Py", "Programming Languages", "Data Encryption", "Incident Response" ], "HR Head": [ "Communication", "Stakeholder Communication" ], "Product Engineer": [ "Software Testing", "Version Control", "Programming Languages", "Cloud Migration", "Cloud Monitoring", "Py", "Cloud Architecture" ] } mapped_skills = { "Manager": [ "Time Management", "Decision Making", "Industry Trends", "Sector Expertise", "Business Acumen", "Stakeholder Communication", "Risk Management", "Agile Methodologies", "Leadership", "Teamwork", "Adaptability" ], "Engineer": [ "Data Encryption", "Incident Response", "Cloud Architecture", "Cloud Migration", "Cloud Monitoring", "Network Security", "Programming Languages", "Py" ], "CEO": [ "Negotiation", "Empathy", "Leadership", "Risk Management", "Business Acumen", "Sector Expertise", "Industry Trends", "Stakeholder Communication", "Creativity" ], "CO-Founder": [ "Creativity", "Sector Expertise", "Empathy", "Leadership", "Stakeholder Communication", "Risk Management", "Business Acumen", "Industry Trends", "Negotiation" ], "Developer": [ "Py", "Programming Languages", "Software Testing", "Version Control" ], "Senior Developer": [ "Network Security", "Py", "Programming Languages", "Data Encryption", "Incident Response" ], "HR Head": [ "Communication", "Stakeholder Communication" ], "Product Engineer": [ "Software Testing", "Version Control", "Programming Languages", "Cloud Migration", "Cloud Monitoring", "Py", "Cloud Architecture" ] } # print(type(mapped_skills)) json_data = json.dumps(mapped_skills) # 'indent' makes it readable print("OPO") print(json_data) print("LKL") # print(type(json_data)) # print(json_data) #print(mapped_skills) skill_tree,token_consumed = map_skills_to_designations_non_empty(skills_db= skillset,JobProfile =JobProfile, mapped_skills= json_data) #skill_tree,token_consumed = map_skills_to_designations(skills_db= skillset,JobProfile =JobProfile) print(token_consumed) print(skill_tree) print(type(skill_tree))