Bryan, Thanks for your response.
1) Thank you for the feedback on the TestableGeoEnrichIP class. I see all that in the @Before setup() method of the TestGeoEnrichIP now. I was stuck looking at the setProperty() method and the value that was being set "./" wondering why can't I find the file. Anyway this cleared it up for me. 2) So I've been thinking about how to generalize the solution even further than I currently have it. Before I do that I'm going to finalize using my current approach making sure the tests are all in place. Then as a refactor and as you suggest I have been thinking about how to add to the PropertyDescriptor to make it a bit easier to interrogate to determine if the value represents a file. My hesitancy to put a boolean that has to be set is that the developer has to remember to set it so you could end up with a situation where it isn't set when it needs to be or it gets set when it shouldn't be. Thanks again for the suggestions these have definitely opened up some additional paths to go down. Ryan On Mon, Apr 22, 2019 at 9:19 AM Bryan Bende <[email protected]> wrote: > Hello, > > 1) In that test, there is a class TestableGeoEnrichIP which is a > subclass of the processor that overrides where the database reader > comes from, and replaces it with a mock reader. So the property for > the file location is not actually used for anything but does need to > be set for validation to pass. > > 2) I haven't fully thought this through, or looked at any of the work > being done, but in order to support this concept it seems like the > framework would have to know that a given PropertyDescriptor > represents a file or directory. I would think we need an option on the > PropertyDescriptor builder like ".identifiesFileSystemResource(true)" > then whenever a processor calls context.getProperty(foo).getValue() it > would know if it needs to do the expansion. If expression language is > involved, and a processor calls > context.getProperty(foo).evaluateAttributeExpressions().getValue() > then the framework would need to do the expansion after evaluating the > expressions. > > -Bryan > > On Fri, Apr 19, 2019 at 3:21 PM Ryan Withers <[email protected]> > wrote: > > > > Hello all, > > > > The basis for the story / jira I've been working on is around properly > > expanding the users home directory when the shortcut ~ is provided as a > > value for any PropertyDescriptor pointing to a file. So I've been > working > > through all the processors and implementing this. > > > > I have two questions as I'm nearing completion with this. > > > > 1. The class I'm looking at is TestGeoEnrichIP.java. It has a number of > > test cases that look at the base directory of the project for a maxmind > > database file. An example of this is: > > > > testRunner.setProperty(GeoEnrichIP.GEO_DATABASE_FILE, "./"); > > > > I've looked through the project for *.mmdb files as described here: > > https://github.com/maxmind/MaxMind-DB/tree/master/test-data, I didn't > find > > any references to files like that. I know I need to spend some > additional > > time with this but I wanted to ask the group to get some guidance and > > additional understanding. > > > > 2. My second question is related to file expansion and expression > > language. Specifically given the requirement to take ~/ and map it to > > /home/rwithers/ should this same behavior be enabled if the property > > descriptor has expression language enabled or only if expression language > > is disabled. I've recently run into an example where expression language > > is enabled on a property descriptor and it interfered with the path > > expansion. So naturally I started wondering if path expansion should > work > > when expression language is also enabled. > > > > Thanks in advance, > > > > -- > > Ryan Withers > > Senior Software Developer / Analyst > > > > http://www.linkedin.com/in/ryanwithers > -- Ryan Withers Senior Software Developer / Analyst http://www.linkedin.com/in/ryanwithers
