Polyglot Pipelines with Apache Nifi
summary
Shows a working pipeline that moves CSVs from the Chicago data portal into OpenSearch using Apache NiFi and Apache Tika, then runs a Python-based table detection model (Hugging Face Transformers + PyTorch + OpenCV) via py4j style IPC inside a per-plugin virtualenv. Demonstrates annotated images and COCO JSON output, packaged as a deployable Nar file (built with hatch), notes GPU support via Python libs and serialization/packaging tradeoffs.
AI-written summary · about
chapters (7)
- 0:00Introduction and speaker background
- 2:49Apache NiFi crash course demo
- 14:04Why mix Java and Python
- 21:08Py4J integration in NiFi
- 29:56AI use case with Python processors
- 41:51Building and deploying Python processors
- 48:20Q&A on performance and GPU support
related talks
41mAdding Generative AI to Real Time Streaming Pipelines
28mApache NiFi, Kafka, RisingWave, Iceberg + Stock Data
29mBuilding Apache NiFi 2.0 Python Processors
46mApache NiFi 101: Introduction and Best Practices
30mThe Fabulous World of Data – ETL Powered by Apache NiFi and Scrapy
34mData pipelines powered by Open source and fun
47mBuilding Modern Data Streaming Apps with Open Source
21mMeeting complex data load and data preparation challenges
