Home

For documentation of PyElly software, see the PyEllyManual.pdf file on Github.

(150126) I am now building a new "marking" application to generate shallow XML from English text. This has uncovered many problems not picked up in previous PyElly unit or integration testing. If the problems can be quickly solved, then I will make the code available in a quick v.1.0.* release.

(150207) The "marking" application still needs extensive work, but has been helpful in debugging PyElly. We are now up to v1.0.6 and and may go further before getting to v1.1.

(150212) First stage of "marking" application is done. Moving on to v1.1, which will include two new Python modules for handling personal names in text.

(150321) v1.1 has been released with four new modules. v1.2 will focus on the handling of biochemical nomenclature, but this will be coming about slowly. The current plan is to expand the "indexing", "chinese", "marking", and "name" applications for integration testing.

(150403) v1.2 replaces Berkeley Database with SQLite to keep PyElly under a BSD license. It is no longer necessary to download anything extra to support PyElly external vocabulary tables.

(150415) v1.2.1 incorporates changes required to make PyElly MARKING application handle a greater variety of text data. This will the same approach taken for v1.2.2. A v1.2.3 release is likely. This is because of the importance of making MARKING credible.

(150515) Continued expansion of the MARKING application with more text samples from the Web, and making PyElly software changes as needed to handle them properly. This has been done for releases v1.2.2 through v1.2.4 and is expected for v1.2.5 and higher. My goal is a doubling of MARKING text samples.

(150525) Publishes PyElly bug fixes and other changes prompted by PyElly processing more test samples of text data from the Web. Release v1.2.6 is coming shortly.

(150629) MARKING can now handle the kinds of sentences found in "Alice in Wonderland" and Ted Cruz's official U.S. Senate bio. There may be another v1.2.* release before v1.3 comes out.

(150630) Still far from being able to process the 100+ sample sentences selected at the beginning of the year to test the MARKING application. Will continue chipping away at them.