job-hunter-bot（求职机器人）

Job Hunter - Automated Job Search System

Build a complete job hunting system: LinkedIn scraper, match scorer, and Telegram bot with inline buttons.

What This System Does

1. Scrapes real jobs from LinkedIn (public guest API, no login needed)
2. Scores each job 0-100% based on candidate profile (title, skills, experience, location)
3. Sends matching jobs to Telegram with action buttons (details, apply, remove)
4. Provides a clean foundation you can extend with CV generation later if needed

Setup Flow

1. Gather Candidate Profile

Ask the user for:

• Target roles (e.g. data analyst, BI developer, frontend developer)
• Core skills (e.g. SQL, Python, React, Power BI)
• Bonus skills (nice-to-have)
• Max years of experience they qualify for
• Preferred location and metro area cities
• Contact info (name, email, phone, LinkedIn URL)
• Work experience (companies, roles, dates, bullet points)
• Education (degree, institution, year)

2. Create Telegram Bot

Guide the user:

1. Open Telegram, search for @BotFather
2. Send /newbot, choose a name and username
3. Copy the bot token
4. Get their Telegram user ID (send a message to @userinfobot)
5. Optionally add more authorized users (e.g. the job seeker)

3. Deploy the System

Create a project directory and deploy these scripts (from scripts/):

job-hunter/
├── config.json          # Bot token, user IDs, candidate profile
├── jobs.db              # SQLite database (auto-created)
├── scorer.py            # Match scoring engine
├── linkedin_scraper.py  # LinkedIn job scraper
├── bot.py               # Telegram bot with inline buttons
└── notify_new_jobs.py   # Send new matches to Telegram

config.json Structure

{
  "telegram_bot_token": "TOKEN_FROM_BOTFATHER",
  "telegram_user_id": 123456789,
  "authorized_users": [123456789],
  "notify_users": [123456789],
  "candidate": {
    "name": "Full Name",
    "email": "email@example.com",
    "phone": "054-1234567",
    "linkedin": "linkedin.com/in/username",
    "location": "Tel Aviv, Israel",
    "target_titles": ["data analyst", "bi developer"],
    "good_titles": ["business analyst"],
    "core_skills": ["sql", "python", "power bi"],
    "bonus_skills": ["etl", "dax", "pandas"],
    "max_years": 2,
    "preferred_locations": ["tel aviv", "herzliya", "ramat gan"],
    "metro_locations": ["petah tikva", "rishon lezion"]
  }
}

4. Customize Scripts

After copying scripts from scripts/, customize:

• scorer.py - Update PROFILE dict with candidate's profile from config.json
• linkedin_scraper.py - Update DEFAULT_QUERIES with relevant search terms
• bot.py - Should work with just config.json changes
• notify_new_jobs.py - Verify notification flow and recipients

5. Install Dependencies

Install Python dependencies required by the included scripts. At minimum, verify the libraries imported by the scraper and bot are available in your environment.

6. Initialize Database

The database auto-creates on first run. Schema:

CREATE TABLE jobs (
    job_id TEXT PRIMARY KEY,
    title TEXT, company TEXT, location TEXT,
    url TEXT, career_url TEXT,
    description TEXT, requirements TEXT,
    required_years INTEGER,
    published_date TEXT, found_date TEXT,
    status TEXT DEFAULT 'new'
);

7. Start the Bot

nohup python3 -u bot.py > bot.log 2>&1 &

8. Set Up Daily Search (Cron)

# Run daily job search + notify at 9 AM
0 9 * * * cd /path/to/job-hunter && python3 linkedin_scraper.py && python3 notify_new_jobs.py

Or use OpenClaw cron:

openclaw cron add --name daily-job-search --schedule "0 9 * * *" --prompt "Run job search and notify"

Bot Commands

Scoring Weights

Thresholds: 🟢 >= 70% apply | 🟡 >= 50% review | 🔴 < 50% skip

Troubleshooting

• Bot not responding: Check only ONE instance is running (ps aux | grep bot.py)
• 409 Conflict: Multiple bot instances. Kill all, restart one.
• No jobs found: Check search queries match real LinkedIn job titles
• Scoring too high/low: Adjust weights in scorer.py