update OSD parsing: effervescence, drop ordered factors #68
Conversation
brownag
changed the title
|
A further wrinkle with effervescence, similar to colors in moist/dry/concentrations etc. is that there can be far more than one or even two classes reported depending on what part of the matrix or other features are being described in that section of the narrative. |
- greediness: preferentially take first class (after comma or semicolon most likely to be matrix eff) - allow ranges of classes separated by "to" (like drainage class) - note there are a wide range of narrative comments still not yet handled e.g. "slightly effervescent but strongly effervescent in spots"
|
OK. A lot more effervescence information has now been been captured. Merging, will add some new stuff to typical pedon reporting tools to tabulate and identify particularly egregious issues. There are many cases with multiple classes without much qualification about what features were being tested, and even more cases where there are long-form descriptions of the pattern of effervescence. Possibly could parse the entire phrase between semicolons, but the formatting is not 100% consistent so this would likely lead to some problems/unintended inclusion of non-effervescence data |
Extension of #67 to cover non-standard, typos etc. effervescence classes.
This removes the creation of ordered factors from several parsed data elements.
The JSON output does not support factor encoding, and it is extremely common to have typos in OSDs. In general it can't be assumed that standard verbiage and choice lists are used. Returning the values as plain text will facilitate error checking and or writing rules to autoconvert classes to the standard set.