Refactored grammar to remove need for second transformer #385
Conversation
|
Hello @sognetic, nice to hear back from you and thank you very much already for this new contribution 👍! This PR pushed me to continue on task #357, see the latest #387 from today. One model in particular, rather unique in the way it span many lines, needed to be tested. I will get to your PR early next week at the very latest. In the meantime can you rebase to (1) get some more important tests run and (2) get a fix for the pre-commit hook CI (failing for you as well abvove)? Thanks. |
…el_alias_transformer_replacement
|
Hi @eduardo-rodrigues! I've also worked on a second feature which removes the need to maintain two lists of models in both |
|
Sounds great! Yeah, task #357 is a bit of a pain and repetitive work, but worth to ensure the parsing not only succeedes but also gives correctly parsed info. So far I've found 2-3 issues and made several improvements along the way. It does take time. |
Hi,
as promised (with some delay) I've had another think about the feature introduced in #100 which slowed down processing quite a bit by introducing an unnecessary transformer. I've removed this transformer again and refactored the grammar such that the
MODEL_NAMEterminal doesn't need to include the white space / semicolon anymore but just matches on a word boundary (\b). This simplifies the grammar quite a bit and speeds up parsing of the LHCb / Belle II decay files by 15%.I've removed the tests corresponding to the removed transformer, I'm not sure if / how I should replace them with something in this PR. All other tests pass.
Let me know what you think!
EDIT: I don't know why the
pre_commitcheck in the CI doesn't work, everything is okay locally.