Repository navigation
[OSD] OSDs with multiline format #83
Description
Following up #81 some multiline formatting variants do not work right with our regex patterns that are not multiline. The result is that some data elements will always be NA if they are not on the first line, because only the first line of the narrative is being returned.
It is possible we can preprocess this to get them on one line by concatenating all lines up to some line delimeter. In these examples we have horizon narrative ending with "." optionally followed by parenthetical thickness.
| OSD | JSON |
|---|---|
| ADIOS | ADIOS.json |
| BRINGMEE | BRINGMEE.json |
Metadata
Metadata
Assignees
Labels
No labels
Activity
dylanbeaudette commentedon Sep 8, 2025
Wow, that is some strange formatting. Over the years, I've found that some corner-cases like this are worth a call to the relevant SDQS or SRSS.
brownag commentedon Sep 8, 2025
I am planning on doing a more comprehensive look to find "all" of the ones that have the issue, using some method to make bulk corrections, then notifying NWR with a bundle of revised text files. We should be able to at least identify the scope of the problem.
brownag commentedon Sep 9, 2025
#84 added logging and error reporting to the "OSD sections" output as a first step towards closing this issue. There appear to be around 600 series that have this problem in some form.
Some are a lot less problematic than others as it is only the very end of a line that is not being parsed, and we don't yet parse the horizon thickness statements.
There are also probably cases where the first line of typical pedon is split, where the profile-level moisture statement is given as dry, that will need correction. This PR only addresses instances where the multiline statements affect horizon parsing, the first line is currently excluded from that logic.
My thinking is that we will probably be able to concatenate to fix many of these cases without having to fix the source.
brownag commentedon Sep 15, 2025
A decent chunk (~500 out of 594 identified in #84) of these are now corrected in horizon parsing with #87 . Now may need to update the logging to find just the ones with intractable issues. It may be that some of the other changes in recent PRs have shifted the expected number