bluegreen-labs / phenocamr Goto Github PK

https://rstudio.github.io/crosstalk/index.html

Example usage should be:

download_phenocam(site = "harvard",
vegetation = "DB",
frequency = 3,
phenophase = TRUE)

i.e. argument is 'phenophase' not 'phenophases'.

Worked example of how to deal with local post processing using the process_phenocam() function.

# install the latest development release!
library(phenocamr)

# I'll download some normal data. I'll use your
# home directory for this.

download_phenocam("spruceT0P19E",
                  frequency = 3,
                  veg_type = "EN",
                  smooth = FALSE,
                  outlier_detection = FALSE,
                  phenophase = FALSE,
                  out_dir = "~")
#> Downloading: spruceT0P19E_EN_1000_3day.csv


# List your snow flagged files
files <- list.files("~","*3day.csv", full.names = TRUE)

# loops over all files if multiple ones provided,
# snow_flags will be removed by the detect_outlier()
# routine if present.
lapply(files, function(file){
  
  # only do the smoothing (no phenophases)
  # write files to disk (will overwrite original if the same
  # directory is provided)
  process_phenocam(file,
                   out_dir = "~",
                   phenophase = FALSE,
                   internal = FALSE)
})
#> -- Flagging outliers!
#> -- Smoothing time series!

^{Created on 2019-04-09 by the reprex package (v0.2.1)}

Personal communications with Rebecca Killick [email protected]

"If you take the first differences of the data then you now have a 0-mean process and the problem reduces to a change in variance, i.e. cpt.var(gcc,method='PELT',penalty='SIC1'). I noticed in your code that you use smoothing splines on the data first. This is not advised as it changes the structure of the data and makes abrupt changes appear more gradual."

Include weather underground data for real time comparison of local weather with Gcc time series and / or model predictions.

The example download_phenocam() command grabs multiple sites which match the string 'harvard'. The help for this function should note that the site argument is a regular expression. And perhaps the example should use 'harvard$' as the site name.

Hi! I am intrigued by the "optimal_span" function you have written because I'm looking for a way to find the optimal span for a series of loess regressions I'm running.

In your documentation for the function, you write in the code "Internal function only, should not be used stand-alone." Could you explain if you mean I shouldn't use this function by itself or something else? Also, if possible could you please explain why this is the case? Thanks! Appreciate the work!
https://github.com/khufkens/phenocamr/blob/cd1ed0b708b551852830a22ab7cb772cddefd87a/R/optimal_span.r

Hi! I'm looking to add this project to the LF Energy Landscape ( l.lfenergy.org ) and was wondering if you had a pure SVG version of the logo. If you have an EPS or AI version that works too. Thanks!

Just a note that the example code on the README.md has phenophases = TRUE in the download_phenocam when it should just be phenophase = TRUE

I noticed when copying to make an instructional for someone and it threw up an error, figured I'd drop you a line just in case. Thanks!

A Hue based filter performs well in finding images with snow and cloud contamination, even on average daily values across an ROI (no image processing needed).

# HSV based snow flagging
library(phenocamr)
library(tidyverse)

if(!file.exists(file.path(tempdir(),"chibougamau_EN_1000_1day.csv"))){
  download_phenocam(site = "chibougamau",
                    out_dir = tempdir(),
                    frequency = 1,
                    roi_id = 1000)
}

# read data
df <- read.table(file.path(tempdir(),"chibougamau_EN_1000_1day.csv"),
                 sep = ",",
                 header = TRUE,
                 stringsAsFactors = FALSE)

df <- df %>% filter(!is.na(r_mean))

hsv <- rgb2hsv(df$r_mean, df$g_mean, df$b_mean)

df$snow_flag <- ifelse(hsv[1,] > 0.55 & hsv[1,] < 0.97,"snow","no snow")

p <- ggplot(df) +
  geom_point(aes(as.Date(date), gcc_90, col = as.factor(snow_flag))) +
  theme_minimal()

print(p)

Needed to avoid issues during download in unit tests

Not sure how this is supposed to work but when I use frequency=1 I get the following error message:

    > download_phenocam(site = "harvard$",
    +                   veg_type = "DB",
    +                   roi_id = "1000",
    +                   frequency = "1",
    +                   contract=FALSE)
    Downloading: harvard_DB_1000_1day.csv
    Error in contract_phenocam(data, internal = TRUE, no_padding = TRUE) : 
      This is a 1-day file, will not be contracted to 3-day frequency

And the resulting downloaded file is not padded with the +/- 90 days. My expectations were that the download would result in a file with padding as was the case with the 3-day frequency. This seems to be a problem with expand_phenocam() calling contract_phenocam() which fails for the 1-day file and stops the padding from occurring. I've got a local workaround but not sure if the logic matches your intent.

Some plotting routines can be rewritten more nicely.