issues.org

(setq org-sync-github-auth ‘(“jstavel” . “aeLohjo9”))

Bugs of differ

jpylyzer - create unittest to check whether all significant properties are mapped right way

• see appCtx-differ-cmdline.xml. It is main source.

So there is bean metadataExtractors. See where it is used.

There is jpylyzerMetadataExtractor in the file. It will be called in a unittest to get output.

• please see, what structure the output is.

There is ArraySet extractorProperties, identificationProperties, validationProperties, characterizationProperties in a file metadataSignificantProperties. All of them will be checked in the unittest.

• Choose two images that jpylyzer can recognize. See property supportedFileExtensions in a file jpylyzerMetadataExtractor.
• call jpylyzer by hand and try to recognize what values of significant properties the extractor gives.
• save the values of significant properties for a given image into some structure. I think you can create spring application context for junit.
• write junit test for bean jpylyzerMetadataExtractor to check, whether outputs it gives for a given image are the same as values by hand.

If you have problem with collecting values of significant properties for given image I help you.

find one property that jpylyzer shows and that is not mapped yet

one property that does not appear in text report

fix refer to common.properties to proper place

proper link is differ-cmdline/src/main/resources/common.properties

implement some hash index to access properties by name

https://differ.readthedocs.org/en/latest/#id31

There are a lot of properties that program must collects and normalizes.
They are stored in ArrayList structure. It is often necessary to select given properties
by its own name.

Implementing some kind of index by property name must speed up an aplication.
At least ArrayHash by property name could help.

to create a thumbnail of an image and to add it into report

User could view image in a report. So it is easier to remember what images were compared.

result of comparion can be in other than text format

It could be great to have a comparison report in `html`.
Or in `pdf`. Report can be done using JasperReport.

See: https://differ.readthedocs.org/en/latest/#result-of-comparion-can-be-in-other-than-text-format

to improve an iteration process of recognizing new properties

https://differ.readthedocs.org/en/latest/#id35

It is difficult to observe all properties that metadata extractors offer.
So it is possible that some of properties remain unrecognized and they are not never used.

It would be great to offer some debug mode when an application shows all unused properties.
At the end of an iteration (ie. configuration) an application will show no unused property.

calling of external java extractors throught their methods using running JVM

https://differ.readthedocs.org/en/latest/#id32
Metadata extractors that are written in java can be called in running JVM
calling some class methods. It must speed up an application a lot.

save steps of jpylyzer installation into documentation

Please save your installations steps in documentation.
It is enought to copy and paste steps you did on your terminal.
The source file of a documentation regarding to jpylyzer is docs/jpylyzer.rst.

install sphinx documentation tool

documentation is written in restructured text.
[[https://docutils.sourceforge.net/rst.html]]

If you want to see rendered documentation, it is necessary to install sphinx tools.
[[https://sphinx-doc.org/]]

install jpylyzer

and try to run docs/bin/get-jpylyzer.sh
It is a script to collect outputs from jpylyzer.
It is the first step to collect significant properties - observing output of an external program.

install jpylyzer

please refer to documentation.
Please, write a short installation guide how to install.
Just copy and paste commands you did.

GSoC proposal

Hello. There is my code suggestion for start. It is not much.