A crawl workstation: View, Control, and Crawl.

Inject seeds. Fetch politely. Parse with Tika. Store the frontier and pages in a Solr CrawlDB. Label pages relevant / not / highly so the next generate pass prefers the right hosts. A Vue 3 UI is the View; FastAPI is Control; the Python fetch loop is Crawl.

The mark is an ember in a crawl orbit — the spark is the page being fetched, the nodes on the ring are the frontier. Compass-rose geometry for navigating the web (search, Tika, Solr), not a campus seal. Named as Spark-Crawler; this tree still crawls, it just does it as a single workstation.

No Docker. No Spark. No Elasticsearch. No ArangoDB.

Needs Java 21+ (Solr 10), Java 11+ is enough for Tika if 21 is on PATH as JAVA_HOME for Solr, Python 3.10+, and Node 18+ to build the UI. Solr is downloaded on first use into data/ (gitignored) and bound to 127.0.0.1:8983. The app is 127.0.0.1:8180.

python3.12 -m venv .venv
.venv/bin/pip install -e .
cd web && npm install && npm run build && cd ..

bin/sparkler inject -id demo -sf demo/seeds.txt
bin/sparkler crawl  -id demo -tn 10 -i 1
bin/sparkler serve

Open http://127.0.0.1:8180/

bin/sparkler inject -id news -su 'https://example.com/'
bin/sparkler crawl  -id news -i 2 --same-host
bin/sparkler reset  -id news --yes
bin/sparkler solr   status

bin/sparkler uses .venv when it exists.

Inspired by USC IRDS Sparkler.