This repository has been archived by the owner on Dec 1, 2017. It is now read-only.
0.4.1
There should not be any backwards incompatibilities from 0.3.0, but there are some places where the data parsed for the same field in the same CCDA may change.
Parser changes:
- Fixed a bug that led to some CCDA vitals, medication, and results sections being completely ignored (see #125)
- thanks to @obalyuk87 for reporting
- Include translation data for results (may be used for other code systems besides LOINC)
- Include the document title
- thanks to @anmolagar for contributing code for both of these
- Time precision is no longer lost if a source specifically provides a datetime with a zero time (midnight; see #133)
- thanks to @ufl-taeber for reporting
- Include the HL7 code for guardian relationships
- Correct the data being parsed as the "drug vehicle" for medications
- Correct the data being parsed as the "device" for procedures (see #17)
- Unescape special XML/HTML characters (e.g., "<")
- Parse non-numeric result values
- Add free-text medication names, which sometimes contains data not found anywhere else in the med data (
product.text
) - Falls back to text fields when structured data is missing much more often and does a better job parsing text fields which reference data from HTML tables
- thanks to @Smcner for reporting lots of fall downs
Generator changes:
- Many more data elements in the BB JSON are now supported by the generator, including drug vehicle, problem translation, problem age, and result reference range
- Better handling for some optional fields that were being treated as mandatory
- Now properly handles non-numeric result values
- Tests that will break on an ongoing basis if new data is added to the parser without being added to the generator