566 lines
18 KiB
Python
566 lines
18 KiB
Python
#!/usr/bin/env python3
|
|
"""
|
|
Generate a simple text-based HTML dashboard of all tracked jobs.
|
|
"""
|
|
|
|
import re
|
|
from datetime import datetime
|
|
from pathlib import Path
|
|
|
|
from db import Database
|
|
|
|
|
|
# Regions/locations we care about (case-insensitive matching)
|
|
DESIRED_REGIONS = [
|
|
"canada", "toronto", "vancouver",
|
|
"germany", "berlin", "munich",
|
|
"emea",
|
|
"americas", # includes North/South America
|
|
"north america",
|
|
"worldwide", "global", "anywhere",
|
|
]
|
|
|
|
# Locations to explicitly exclude (on-site or remote restricted to these)
|
|
EXCLUDED_LOCATIONS = [
|
|
# US cities/states (we don't want US-only jobs)
|
|
"san francisco", "new york", "nyc", "seattle", "austin", "boston",
|
|
"chicago", "denver", "los angeles", "atlanta", "dallas", "houston",
|
|
"california", "washington", "texas", "massachusetts", "colorado",
|
|
"united states", "usa", "u.s.", "us-", "usa-",
|
|
# UK
|
|
"london", "united kingdom", "uk", "dublin", "ireland",
|
|
# Australia/APAC (not EMEA)
|
|
"sydney", "melbourne", "australia", "singapore", "tokyo", "japan",
|
|
"india", "bangalore", "bengaluru", "hyderabad", "delhi",
|
|
"korea", "seoul", "taiwan", "taipei", "china", "beijing", "shenzhen",
|
|
# Other excluded
|
|
"israel", "tel aviv", "brazil", "sao paulo", "mexico",
|
|
"netherlands", "amsterdam", "france", "paris", "spain", "madrid",
|
|
"portugal", "lisbon", "poland", "warsaw", "italy",
|
|
"czech", "prague", "serbia", "belgrade", "cyprus", "limassol",
|
|
"austria", "vienna", "sweden", "stockholm", "denmark", "copenhagen",
|
|
"switzerland", "romania", "bucharest", "hungary", "greece",
|
|
"south africa", "indonesia", "jakarta", "malaysia",
|
|
]
|
|
|
|
|
|
def is_location_relevant(location: str, remote_type: str) -> bool:
|
|
"""
|
|
Strict location filter. Only keeps jobs available in Canada, Germany, EMEA, or Worldwide.
|
|
Filters out US-only jobs, UK jobs, APAC jobs, etc.
|
|
"""
|
|
if not location:
|
|
return False # No location info = probably US-based, filter out
|
|
|
|
loc_lower = location.lower()
|
|
|
|
# Check if any desired region is mentioned FIRST
|
|
has_desired = any(region in loc_lower for region in DESIRED_REGIONS)
|
|
|
|
# If it has a desired region, keep it (even if it also mentions excluded locations)
|
|
# e.g., "Remote (United States | Canada)" should be kept because of Canada
|
|
if has_desired:
|
|
return True
|
|
|
|
# If it just says "Remote" with nothing else, keep it (truly remote)
|
|
if loc_lower.strip() == "remote":
|
|
return True
|
|
|
|
# Check for excluded locations
|
|
has_excluded = any(excl in loc_lower for excl in EXCLUDED_LOCATIONS)
|
|
if has_excluded:
|
|
return False
|
|
|
|
# Check for patterns like "In-Office", "Hybrid", "On-site" without desired region
|
|
if any(x in loc_lower for x in ["in-office", "hybrid", "on-site", "onsite", "office based"]):
|
|
return False
|
|
|
|
# If we can't determine, filter it out (safer)
|
|
return False
|
|
|
|
|
|
def extract_location_tags(location: str, remote_type: str) -> tuple[list[str], str]:
|
|
"""
|
|
Extract relevant location tags and a short display location.
|
|
Returns (list of tag names, short location string)
|
|
"""
|
|
if not location:
|
|
return [], ""
|
|
|
|
loc_lower = location.lower()
|
|
tags = []
|
|
short_loc = ""
|
|
|
|
# Check for remote
|
|
is_remote = remote_type == "remote" or "remote" in loc_lower
|
|
if is_remote:
|
|
tags.append("remote")
|
|
|
|
# Check for Canada
|
|
if any(x in loc_lower for x in ["canada", "toronto", "vancouver"]):
|
|
tags.append("canada")
|
|
short_loc = "Canada"
|
|
|
|
# Check for Germany/Berlin
|
|
if any(x in loc_lower for x in ["germany", "berlin", "munich"]):
|
|
tags.append("germany")
|
|
short_loc = "Germany" if "germany" in loc_lower else "Berlin"
|
|
|
|
# Check for EMEA
|
|
if "emea" in loc_lower:
|
|
tags.append("emea")
|
|
short_loc = "EMEA"
|
|
|
|
# Check for Americas/North America
|
|
if "americas" in loc_lower or "north america" in loc_lower:
|
|
tags.append("americas")
|
|
short_loc = "Americas"
|
|
|
|
# Check for Worldwide
|
|
if any(x in loc_lower for x in ["worldwide", "global", "anywhere"]):
|
|
tags.append("worldwide")
|
|
short_loc = "Worldwide"
|
|
|
|
# If no specific region found but it's remote
|
|
if not short_loc and is_remote:
|
|
short_loc = "Remote"
|
|
|
|
return tags, short_loc
|
|
|
|
|
|
def generate_dashboard(output_path: str = "data/dashboard.html"):
|
|
"""Generate a static HTML dashboard."""
|
|
db = Database()
|
|
jobs = db.get_all_active_jobs()
|
|
|
|
# Get all monitored companies
|
|
all_company_names = db.get_all_companies()
|
|
|
|
# Track total jobs per company (before location filtering)
|
|
total_per_company = {}
|
|
for company_name, job in jobs:
|
|
total_per_company[company_name] = total_per_company.get(company_name, 0) + 1
|
|
|
|
# Group by company, filtering out irrelevant remote locations
|
|
companies = {}
|
|
filtered_count = 0
|
|
for company_name, job in jobs:
|
|
if not is_location_relevant(job.location, job.remote_type):
|
|
filtered_count += 1
|
|
continue
|
|
if company_name not in companies:
|
|
companies[company_name] = []
|
|
companies[company_name].append(job)
|
|
|
|
# Ensure all monitored companies are in the dict (even with 0 jobs)
|
|
for name in all_company_names:
|
|
if name not in companies:
|
|
companies[name] = []
|
|
if name not in total_per_company:
|
|
total_per_company[name] = 0
|
|
|
|
total_shown = sum(len(jobs) for jobs in companies.values())
|
|
total_scraped = sum(total_per_company.values())
|
|
|
|
# Sort companies by name
|
|
sorted_companies = sorted(companies.items())
|
|
|
|
html = f"""<!DOCTYPE html>
|
|
<html lang="en">
|
|
<head>
|
|
<meta charset="UTF-8">
|
|
<meta name="viewport" content="width=device-width, initial-scale=1.0">
|
|
<title>Job Board</title>
|
|
<style>
|
|
:root {{
|
|
--bg: #1a1a1a;
|
|
--fg: #e0e0e0;
|
|
--accent: #4a9eff;
|
|
--muted: #888;
|
|
--border: #333;
|
|
--highlight: #2a2a2a;
|
|
}}
|
|
* {{ box-sizing: border-box; margin: 0; padding: 0; }}
|
|
body {{
|
|
font-family: "SF Mono", "Monaco", "Inconsolata", "Fira Code", monospace;
|
|
font-size: 14px;
|
|
line-height: 1.6;
|
|
background: var(--bg);
|
|
color: var(--fg);
|
|
padding: 20px;
|
|
max-width: 1200px;
|
|
margin: 0 auto;
|
|
}}
|
|
header {{
|
|
border-bottom: 1px solid var(--border);
|
|
padding-bottom: 15px;
|
|
margin-bottom: 20px;
|
|
}}
|
|
h1 {{
|
|
font-size: 18px;
|
|
font-weight: normal;
|
|
color: var(--accent);
|
|
}}
|
|
.meta {{
|
|
color: var(--muted);
|
|
font-size: 12px;
|
|
margin-top: 5px;
|
|
}}
|
|
.filters {{
|
|
margin: 15px 0;
|
|
padding: 10px;
|
|
background: var(--highlight);
|
|
border-radius: 4px;
|
|
}}
|
|
.filters input {{
|
|
background: var(--bg);
|
|
border: 1px solid var(--border);
|
|
color: var(--fg);
|
|
padding: 8px 12px;
|
|
width: 100%;
|
|
max-width: 400px;
|
|
font-family: inherit;
|
|
font-size: 14px;
|
|
border-radius: 4px;
|
|
}}
|
|
.filters input:focus {{
|
|
outline: none;
|
|
border-color: var(--accent);
|
|
}}
|
|
.stats {{
|
|
display: flex;
|
|
gap: 20px;
|
|
margin: 10px 0;
|
|
font-size: 12px;
|
|
color: var(--muted);
|
|
}}
|
|
.company {{
|
|
margin-bottom: 25px;
|
|
}}
|
|
.company-header {{
|
|
display: flex;
|
|
align-items: baseline;
|
|
gap: 10px;
|
|
padding: 8px 0;
|
|
border-bottom: 1px solid var(--border);
|
|
cursor: pointer;
|
|
}}
|
|
.company-header:hover {{
|
|
color: var(--accent);
|
|
}}
|
|
.company-name {{
|
|
font-weight: bold;
|
|
color: var(--accent);
|
|
}}
|
|
.company-count {{
|
|
color: var(--muted);
|
|
font-size: 12px;
|
|
}}
|
|
.jobs {{
|
|
margin-left: 0;
|
|
}}
|
|
.job {{
|
|
padding: 6px 0;
|
|
border-bottom: 1px solid var(--border);
|
|
display: flex;
|
|
justify-content: space-between;
|
|
gap: 20px;
|
|
align-items: baseline;
|
|
}}
|
|
.job:last-child {{
|
|
border-bottom: none;
|
|
}}
|
|
.job:hover {{
|
|
background: var(--highlight);
|
|
}}
|
|
.job-title {{
|
|
overflow: hidden;
|
|
text-overflow: ellipsis;
|
|
}}
|
|
.job-title a {{
|
|
color: var(--fg);
|
|
text-decoration: none;
|
|
}}
|
|
.job-title a:hover {{
|
|
color: var(--accent);
|
|
text-decoration: underline;
|
|
}}
|
|
.job-location {{
|
|
color: var(--muted);
|
|
font-size: 12px;
|
|
text-align: right;
|
|
white-space: nowrap;
|
|
flex-shrink: 0;
|
|
}}
|
|
.tag {{
|
|
display: inline-block;
|
|
padding: 2px 6px;
|
|
border-radius: 3px;
|
|
font-size: 11px;
|
|
margin-left: 5px;
|
|
}}
|
|
.tag-remote {{
|
|
background: #1a4a1a;
|
|
color: #4ade80;
|
|
}}
|
|
.tag-canada {{
|
|
background: #4a1a1a;
|
|
color: #f87171;
|
|
}}
|
|
.tag-berlin {{
|
|
background: #4a4a1a;
|
|
color: #facc15;
|
|
}}
|
|
.tag-emea {{
|
|
background: #1a3a4a;
|
|
color: #60a5fa;
|
|
}}
|
|
.tag-americas {{
|
|
background: #3a1a4a;
|
|
color: #c084fc;
|
|
}}
|
|
.tag-worldwide {{
|
|
background: #1a4a3a;
|
|
color: #34d399;
|
|
}}
|
|
.hidden {{
|
|
display: none;
|
|
}}
|
|
.toc {{
|
|
margin: 20px 0;
|
|
padding: 15px;
|
|
background: var(--highlight);
|
|
border-radius: 4px;
|
|
}}
|
|
.toc-title {{
|
|
font-size: 12px;
|
|
color: var(--muted);
|
|
margin-bottom: 10px;
|
|
}}
|
|
.toc-links {{
|
|
display: flex;
|
|
flex-wrap: wrap;
|
|
gap: 10px;
|
|
}}
|
|
.toc-links a {{
|
|
color: var(--accent);
|
|
text-decoration: none;
|
|
font-size: 13px;
|
|
}}
|
|
.toc-links a:hover {{
|
|
text-decoration: underline;
|
|
}}
|
|
.toc-links .empty {{
|
|
color: var(--muted);
|
|
cursor: default;
|
|
}}
|
|
.toc-links .empty:hover {{
|
|
text-decoration: none;
|
|
}}
|
|
.filter-buttons {{
|
|
display: flex;
|
|
flex-wrap: wrap;
|
|
gap: 8px;
|
|
margin-top: 10px;
|
|
}}
|
|
.filter-btn {{
|
|
background: var(--bg);
|
|
border: 1px solid var(--border);
|
|
color: var(--muted);
|
|
padding: 4px 12px;
|
|
font-family: inherit;
|
|
font-size: 12px;
|
|
border-radius: 4px;
|
|
cursor: pointer;
|
|
transition: all 0.15s;
|
|
}}
|
|
.filter-btn:hover {{
|
|
border-color: var(--accent);
|
|
color: var(--fg);
|
|
}}
|
|
.filter-btn.active {{
|
|
background: var(--accent);
|
|
border-color: var(--accent);
|
|
color: var(--bg);
|
|
}}
|
|
</style>
|
|
</head>
|
|
<body>
|
|
<header>
|
|
<h1>$ job-board</h1>
|
|
<div class="meta">
|
|
Last updated: {datetime.now().strftime("%Y-%m-%d %H:%M:%S")} |
|
|
{total_shown}/{total_scraped} jobs (location filtered) | Monitoring {len(all_company_names)} companies
|
|
</div>
|
|
</header>
|
|
|
|
<div class="filters">
|
|
<input type="text" id="search" placeholder="Filter jobs... (e.g. 'senior engineer', 'remote', 'canada')" autofocus>
|
|
<div class="filter-buttons">
|
|
<button class="filter-btn" data-filter="">All</button>
|
|
<button class="filter-btn" data-filter="engineer">Engineering</button>
|
|
<button class="filter-btn" data-filter="senior engineer">Senior Eng</button>
|
|
<button class="filter-btn" data-filter="staff principal">Staff+</button>
|
|
<button class="filter-btn" data-filter="manager director">Management</button>
|
|
<button class="filter-btn" data-filter="product">Product</button>
|
|
<button class="filter-btn" data-filter="design">Design</button>
|
|
<button class="filter-btn" data-filter="security">Security</button>
|
|
<button class="filter-btn" data-filter="remote">Remote</button>
|
|
<button class="filter-btn" data-filter="canada">Canada</button>
|
|
<button class="filter-btn" data-filter="germany">Germany</button>
|
|
<button class="filter-btn" data-filter="emea">EMEA</button>
|
|
<button class="filter-btn" data-filter="americas">Americas</button>
|
|
<button class="filter-btn" data-filter="worldwide">Worldwide</button>
|
|
</div>
|
|
<div class="stats">
|
|
<span id="visible-count">{total_shown} jobs shown</span>
|
|
</div>
|
|
</div>
|
|
|
|
<div class="toc">
|
|
<div class="toc-title">Jump to company:</div>
|
|
<div class="toc-links">
|
|
"""
|
|
|
|
# Table of contents
|
|
for company_name, company_jobs in sorted_companies:
|
|
anchor = company_name.lower().replace(" ", "-")
|
|
filtered = len(company_jobs)
|
|
total = total_per_company.get(company_name, 0)
|
|
if filtered > 0:
|
|
html += f' <a href="#{anchor}">{company_name} ({filtered}/{total})</a>\n'
|
|
else:
|
|
html += f' <span class="empty">{company_name} (0/{total})</span>\n'
|
|
|
|
html += """ </div>
|
|
</div>
|
|
|
|
<main id="job-list">
|
|
"""
|
|
|
|
# Job listings (only for companies with jobs)
|
|
for company_name, company_jobs in sorted_companies:
|
|
if not company_jobs:
|
|
continue # Skip companies with no jobs after filtering
|
|
anchor = company_name.lower().replace(" ", "-")
|
|
html += f"""
|
|
<div class="company" id="{anchor}">
|
|
<div class="company-header">
|
|
<span class="company-name">{company_name}</span>
|
|
<span class="company-count">{len(company_jobs)} positions</span>
|
|
</div>
|
|
<div class="jobs">
|
|
"""
|
|
for job in sorted(company_jobs, key=lambda j: j.title):
|
|
location = job.location or ""
|
|
location_lower = location.lower()
|
|
|
|
# Extract tags and short location
|
|
tag_list, short_loc = extract_location_tags(location, job.remote_type)
|
|
|
|
# Build tag HTML
|
|
tags = ""
|
|
if "remote" in tag_list:
|
|
tags += '<span class="tag tag-remote">remote</span>'
|
|
if "canada" in tag_list:
|
|
tags += '<span class="tag tag-canada">canada</span>'
|
|
if "germany" in tag_list:
|
|
tags += '<span class="tag tag-berlin">germany</span>'
|
|
if "emea" in tag_list:
|
|
tags += '<span class="tag tag-emea">emea</span>'
|
|
if "americas" in tag_list:
|
|
tags += '<span class="tag tag-americas">americas</span>'
|
|
if "worldwide" in tag_list:
|
|
tags += '<span class="tag tag-worldwide">worldwide</span>'
|
|
|
|
html += f""" <div class="job" data-search="{job.title.lower()} {location_lower} {(job.department or '').lower()} {' '.join(tag_list)}">
|
|
<span class="job-title"><a href="{job.url}" target="_blank">{job.title}</a>{tags}</span>
|
|
<span class="job-location">{short_loc}</span>
|
|
</div>
|
|
"""
|
|
html += """ </div>
|
|
</div>
|
|
"""
|
|
|
|
html += """ </main>
|
|
|
|
<script>
|
|
const search = document.getElementById('search');
|
|
const jobs = document.querySelectorAll('.job');
|
|
const companies = document.querySelectorAll('.company');
|
|
const visibleCount = document.getElementById('visible-count');
|
|
const filterBtns = document.querySelectorAll('.filter-btn');
|
|
|
|
function filterJobs(query) {
|
|
let visible = 0;
|
|
const terms = query.toLowerCase().trim().split(/\\s+/).filter(t => t);
|
|
|
|
companies.forEach(company => {
|
|
const companyJobs = company.querySelectorAll('.job');
|
|
let companyVisible = 0;
|
|
|
|
companyJobs.forEach(job => {
|
|
const searchText = job.dataset.search;
|
|
// Match if ANY term matches (OR logic for filter buttons)
|
|
const matches = terms.length === 0 || terms.some(term => searchText.includes(term));
|
|
job.classList.toggle('hidden', !matches);
|
|
if (matches) {
|
|
companyVisible++;
|
|
visible++;
|
|
}
|
|
});
|
|
|
|
company.classList.toggle('hidden', companyVisible === 0);
|
|
});
|
|
|
|
visibleCount.textContent = `${visible} jobs shown`;
|
|
}
|
|
|
|
search.addEventListener('input', (e) => {
|
|
// Clear active button when typing
|
|
filterBtns.forEach(btn => btn.classList.remove('active'));
|
|
filterJobs(e.target.value);
|
|
});
|
|
|
|
// Filter buttons
|
|
filterBtns.forEach(btn => {
|
|
btn.addEventListener('click', () => {
|
|
const filter = btn.dataset.filter;
|
|
search.value = filter;
|
|
filterBtns.forEach(b => b.classList.remove('active'));
|
|
btn.classList.add('active');
|
|
filterJobs(filter);
|
|
});
|
|
});
|
|
|
|
// Keyboard shortcut: / to focus search
|
|
document.addEventListener('keydown', (e) => {
|
|
if (e.key === '/' && document.activeElement !== search) {
|
|
e.preventDefault();
|
|
search.focus();
|
|
}
|
|
if (e.key === 'Escape') {
|
|
search.value = '';
|
|
filterBtns.forEach(b => b.classList.remove('active'));
|
|
filterJobs('');
|
|
search.blur();
|
|
}
|
|
});
|
|
|
|
// Set "All" as active by default
|
|
filterBtns[0].classList.add('active');
|
|
</script>
|
|
</body>
|
|
</html>
|
|
"""
|
|
|
|
# Write the file
|
|
output = Path(output_path)
|
|
output.parent.mkdir(parents=True, exist_ok=True)
|
|
output.write_text(html)
|
|
print(f"Dashboard generated: {output_path}")
|
|
return output_path
|
|
|
|
|
|
if __name__ == "__main__":
|
|
generate_dashboard()
|