Cephis has the following sorts of examples:
- extended Unit tests ("Integration Tests")
- projects decribed on ContentMine's "discuss" platform
Many cephis tests take a small problem and produce output and are referred to as Integration Tests (ITs). ITs have a special role in Maven and can be run or omitted as required. Typically many ITs:
- Take a "long" time to run. All tests > 5 sec on my machine will be ITs. This means that the other tests act as more frequent Regression tests.
- depend on complex files and documents. This means they can be fragile over time and different platforms. A faling IT may occasionally be a fragile test rather than a fundamental problem. For example file listing order is not determined and although I am retrofitting sorting it isn't universal.
- are used in an exploratory manner. This coincides with new approaches to TDD. Write the tasks addressing the generation of real results (query->results), get the hack working, then gradually coalesce into classes, and, when necessary, write tests for those.
- Occaasionally require "closed" resources such as paywalled scholarly articles. We cannot post these openly (Copyright) and these will fail for other developers. This doesn't matter.
We do most projects as Open Notebook Science and have several successful projects on our discussion forum. Some are run from norma or ami but are mainly based on cephis technology. Examples:
- OCR of 50-year old articles
- text mining with dictionaries
- OCR of very difficult ancient documents
- transformation with stylesheets (
normausescephis) - extraction and clipping of tables
- extracting data from vectors-based plots
There are many more - feel free to read the discuss and update this page.