Spaces:
Build error
Build error
File size: 4,429 Bytes
7f0885b |
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 |
import os
from dotenv import load_dotenv
from huggingface_hub import InferenceClient
import re
from typing import List
import logging
# Set up logging
logging.basicConfig(filename="llm_errors.log", level=logging.ERROR)
# Load environment variables
load_dotenv()
class LLMHelper:
def __init__(self):
self.client = InferenceClient(
model=os.getenv("HF_MODEL", "HuggingFaceH4/zephyr-7b-beta"),
token=os.getenv("HF_TOKEN")
)
def generate_text(self, prompt: str, max_tokens: int = 500) -> str:
"""Generate text from the LLM with error handling and retry"""
for attempt in range(2): # Retry once on failure
try:
response = self.client.text_generation(
prompt,
max_new_tokens=max_tokens,
temperature=0.7,
do_sample=True
)
return response.strip()
except Exception as e:
logging.error(f"LLM Error (Attempt {attempt + 1}): {str(e)}")
if attempt == 1:
return ""
return ""
def parse_tech_stack(input_str: str) -> List[str]:
"""Parse tech stack input into a cleaned list of technologies"""
if not input_str.strip():
return []
replacements = {
"c#": "C#",
"c++": "C++",
"f#": "F#",
"golang": "Go",
"js": "JavaScript",
"ts": "TypeScript",
"nodejs": "Node.js",
"node": "Node.js",
"reactjs": "React",
"vuejs": "Vue.js",
"postgresql": "PostgreSQL",
"mysql": "MySQL"
}
techs = re.split(r'[,;/\n]', input_str)
cleaned = []
for tech in techs:
tech = tech.strip()
if tech:
tech = tech.lower()
tech = replacements.get(tech, tech)
if not re.match(r'^[a-z0-9+#]+$', tech):
tech = tech.title()
if tech not in cleaned and len(tech) > 1:
cleaned.append(tech)
return cleaned
def generate_tech_questions(tech_stack: List[str], years_experience: int) -> List[str]:
"""Generate technical questions based on tech stack and experience level"""
if not tech_stack:
return ["Please describe your technical experience."]
llm = LLMHelper()
difficulty = "beginner" if years_experience < 2 else "intermediate" if years_experience < 5 else "advanced"
questions = []
for tech in tech_stack:
prompt = f"""Generate exactly 3 technical questions about {tech} for a candidate with {years_experience} years of experience.
Difficulty level: {difficulty}
Format each question clearly numbered like:
1. [Question about {tech}]
2. [Question about {tech}]
3. [Question about {tech}]
The questions should:
- Be technical and specific to {tech}
- Cover different aspects (syntax, architecture, debugging)
- Require detailed answers
- Avoid simple yes/no or one-word answers
- Be unique and not repetitive
- Be relevant to real-world use cases
Example for Python (intermediate):
1. How would you optimize memory usage in a Python application processing large datasets?
2. Explain the differences between multiprocessing and threading in Python with examples.
3. Describe how you would implement and test a custom context manager in Python.
"""
response = llm.generate_text(prompt)
tech_questions = []
for line in response.split('\n'):
line = line.strip()
if line and re.match(r'^\d+\.\s*\[.*\]\s*$', line):
question = line.split('.', 1)[1].strip()[1:-1].strip()
if question:
tech_questions.append(f"{tech}: {question}")
if len(tech_questions) < 3:
default_questions = [
f"{tech}: Explain the most challenging {tech} project you've worked on and the key technical decisions you made.",
f"{tech}: How would you optimize performance in a {tech} application handling high traffic?",
f"{tech}: Describe your approach to debugging a complex issue in a {tech} application."
]
tech_questions.extend(default_questions[:3 - len(tech_questions)])
questions.extend(tech_questions[:3])
return questions[:15] # Limit to 15 questions max to avoid overwhelming candidates
|