Mark Payne created NIFI-12959:
---------------------------------
Summary: Support loading Python processors from NARs
Key: NIFI-12959
URL: https://issues.apache.org/jira/browse/NIFI-12959
Project: Apache NiFi
Issue Type: Improvement
Components: Core Framework
Reporter: Mark Payne
Assignee: Mark Payne
Fix For: 2.0.0-M3
Currently, third-party dependencies for Python Processors can be handled in two
ways. Either they can be declared as dependencies in a Processor itself; or the
Processor can be in a module where a {{requirements.txt}} dictates the
requirements. These can be very helpful for developing Python based Processors.
However, in production environments, it is not uncommon to see environments
where {{pip}} is not installed. There is an inherent risk in allowing remote
code to be downloaded in an ad-hoc manner like this, without any sort of
vulnerability scanning, etc.
As such, we should allow users to also package python packages in NiFi's native
archiving format (NARs).
The package structure should be as follows:
{code:java}
my-nar.nar
+-- META-INF/
+-- MANIFEST.MF
+-- NAR-INF/
+-- bundled-dependencies/
+-- dependency1
+-- dependency2
+-- etc.
+-- MyProcessor.py{code}
Where {{MyProcessor.py}} could also be a python module / directory.
In this way, we allow a Python Processor to be packaged up with its third party
dependencies and dropped in the lib/ directory (or extensions) directory of a
NiFi installation in the same way that a Java processor would be.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
