Lots of commits here... most are of little consequences. However! I did do a few really big things -
1) I added a new function in 'site_comp_test.R'. This is my fist pass at using network relationships to figure out if flags are occurring up/downstream of a given site, then using that as a means to confirm if a flag is warranted. This is used at the very end of the flagging workflow, and produces a new column called "cleaner_flag" (... feel free to change the name)
2) The slope flag is now based on the 1st and 99th percentile slope value, meaning the direction of slope (positive/negative) matters when we flag slope.
3) I added my own version of the anomaly flag, and renamed that flag to "suspect data". I tried to find all instances of the old flag's name across the entire project repo, but be careful! I may have missed some...
I pushed this PR into my (Juan) main fork. It has a merge conflict with the main branch that I will resolve on my end before pushing it back into the main branch.
Lots of commits here... most are of little consequences. However! I did do a few really big things -
1) I added a new function in 'site_comp_test.R'. This is my fist pass at using network relationships to figure out if flags are occurring up/downstream of a given site, then using that as a means to confirm if a flag is warranted. This is used at the very end of the flagging workflow, and produces a new column called "cleaner_flag" (... feel free to change the name)
2) The slope flag is now based on the 1st and 99th percentile slope value, meaning the direction of slope (positive/negative) matters when we flag slope.
3) I added my own version of the anomaly flag, and renamed that flag to "suspect data". I tried to find all instances of the old flag's name across the entire project repo, but be careful! I may have missed some...