Gli esempi “runnable” sono nel repo sotto examples/pipelines/.
examples/pipelines/01-static-books.yamlexamples/pipelines/02-dynamic-quotes.yamlexamples/pipelines/03-iextract-prompt-only.yamlexamples/pipelines/04-attribute-resolvers.yamlexamples/pipelines/05-python-row-transforms.yamlexamples/pipelines/06-io-load-save-csv.yamlexamples/pipelines/07-searchengine-ean-enrich.yamlexamples/pipelines/08-fetch-traces-browser-actions.yamlexamples/pipelines/09-multi-source-seeds-union-dedup.yaml# unione insiemistica seed + dedupexamples/pipelines/10-multi-source-results-union-dedup.yaml# unione risultati multi-sorgente + dedupexamples/pipelines/11-vertical-source-a-offers.yaml# pipeline sorgente A (fetch+extract)examples/pipelines/12-vertical-source-b-offers.yaml# pipeline sorgente B (fetch+extract)examples/pipelines/13-vertical-stitch-union-dedup-offers.yaml# stitching (load_union + dedup)examples/pipelines/14-union-by-name-append-upstream.yaml# append upstream datasetexamples/pipelines/15-aggregation-group-by-key.yaml# aggregazione by-keyexamples/pipelines/16-aggregation-monthly.yaml# aggregazione mensileexamples/pipelines/17-single-pipeline-multi-source-union.yaml# multi-source con store/reset/union_withexamples/pipelines/18-single-pipeline-alternative-syntax.yaml# alternativa multi-sourceexamples/pipelines/19-price-comparison-5-sites.yaml# price comparison (5 siti) – output row-level + pivot downstreamexamples/pipelines/20-sports-betting-5-bookmakers.yaml# odds aggregation (5 bookmaker) – placeholder/complianceexamples/pipelines/21-surebet-intelligent-extraction.yaml# surebet – extraction intelligente, calcolo downstreamexamples/pipelines/22-real-estate-arbitrage-clustering.yaml# real estate arbitrage + clusteringexamples/pipelines/23-llm-finetuning-dataset.yaml# dataset fine-tuning LLM (No-CC policy)examples/pipelines/24-portfolio-management-90d-prediction.yaml# portfolio management 90g
Per la spiegazione completa di ogni esempio, usa la pagina in inglese o apri direttamente i file YAML.
- Gli esempi che citano siti/brand esterni usano spesso URL placeholder o pattern generici: usa solo fonti per cui hai autorizzazione.
- Per i dataset di training LLM, la guida di riferimento (
vertical-llm-finetuning.md) applica una policy No Creative Commons (No-CC): non usare contenuti CC* per training/redistribuzione se non espressamente autorizzato.