All rasters (#4)

* incorporate all data

* add instead of copy habre.tif

* add citation

* updated test file with dates outside of example file range

.gitignore

@@ -2,3 +2,5 @@
 .Rhistory
 .RData
 .Ruserdata
+s3_downloads/
+habre.tif

Dockerfile

@@ -2,7 +2,7 @@ FROM rocker/r-ver:4.2.2
 
 # DeGAUSS container metadata
 ENV degauss_name="habre_pm"
-ENV degauss_version="0.1.0"
+ENV degauss_version="0.2.0"
 ENV degauss_description="weekly pm2.5 for California (Habre)"
 
 # add OCI labels based on environment variables too
@@ -28,7 +28,7 @@ COPY renv.lock .
 
 RUN R --quiet -e "renv::restore()"
 
-COPY habre.tif .
+ADD https://habre.s3-us-east-2.amazonaws.com/habre.tif habre.tif
 COPY pm25_iweek_startdate.csv .
 COPY entrypoint.R .
 

README.md

@@ -5,31 +5,31 @@
 
 ## Using
 
-If `my_address_file_geocoded.csv` is a file in the current working directory with coordinate columns named `lat` and `lon`, then the [DeGAUSS command](https://degauss.org/using_degauss.html#DeGAUSS_Commands):
+If `my_address_file_geocoded.csv` is a file in the current working directory with coordinate columns named `lat`, `lon`, (within the state of California) `start_date`, and `end_date` then the [DeGAUSS command](https://degauss.org/using_degauss.html#DeGAUSS_Commands):
 
 ```sh
-docker run --rm -v $PWD:/tmp ghcr.io/degauss-org/habre_pm:0.1.0 my_address_file_geocoded.csv
+docker run --rm -v $PWD:/tmp ghcr.io/degauss-org/habre_pm:0.2.0 my_address_file_geocoded.csv
 ```
 
-will produce `my_address_file_geocoded_habre_pm_0.1.0.csv` with added columns:
+will produce `my_address_file_geocoded_habre_pm_0.2.0.csv` with added columns:
 
-- **`pm`**: a definition of this geomarker
-- **`sd`**: a definition of that geomarker
-
-### Optional Argument
-
-- If this DeGAUSS container takes an optional argument, describe its usage and effects here.
-- Be sure to also update the example output file name with the argument value.
+- **`pm`**: time weighted average of weekly PM2.5
+- **`sd`**: standard deviation
 
 ## Geomarker Methods
 
-- If needed, put details here about the methods and assumptions used in the geomarker assessment process.
+- Geocoded points are overlaid with weekly PM2.5 rasters corresponding to the input date range.
+- For date ranges that span weeks, exposures are a time-weighted average. 
 
 ## Geomarker Data
 
-- List how geomarker was created, ideally including any scripts within the repo used to do so or linking to an external repository
-- If applicable, list where geomarker data is stored in S3 using a hyperlink like: [`s3://path/to/habre_pm.rds`](https://geomarker.s3.us-east-2.amazonaws.com/path/to/habre_pm.rds)
+- PM2.5 rasters were created using a model developed by Rima Habre and Lianfa Li. 
+
+> Li L, Girguis M, Lurmann F, Pavlovic N, McClure C, Franklin M, Wu J, Oman LD, Breton C, Gilliland F, Habre R. Ensemble-based deep learning for estimating PM2. 5 over California with multisource big data including wildfire smoke. Environment international. 2020 Dec 1;145:106143. https://doi.org/10.1016/j.envint.2020.106143
+
+- The raster stack used in this container is stored in S3 at [`s3://habre/habre.tif`](https://habre.s3-us-east-2.amazonaws.com/habre.tif)
+- Individual rasters that make up the raster stack are stored at [`s3://habre/li_2020/`](https://habre.s3-us-east-2.amazonaws.com/li_2020/)
 
 ## DeGAUSS Details
 
-For detailed documentation on DeGAUSS, including general usage and installation, please see the [DeGAUSS homepage](https://degauss.org).
+For detailed documentation on DeGAUSS, including general usage and installation, please see the [DeGAUSS homepage](https://degauss.org).

make_raster_stack.R

@@ -0,0 +1,12 @@
+library(terra)
+
+# weeks 410 to 931
+rast_files <- c(glue::glue("s3://habre/li_2020/week{410:931}_mean.tif"), 
+                glue::glue("s3://habre/li_2020/week{410:931}_std.tif"))
+
+rast_files <- s3::s3_get_files(rast_files)
+r <- terra::rast(rast_files$file_path)
+r <- round(r, digits = 2) # original 4 files: 7.7 MB to 4.2 MB
+terra::writeRaster(r, "habre.tif", overwrite = TRUE)
+
+fs::dir_delete("s3_downloads")

test/my_address_file_geocoded.csv

@@ -1,4 +1,4 @@
-id,lat,lon,start_date,end_date
-1,34.41471040449594,-115.04499837139416,2008-01-21,2008-01-27
-2,34.02192926572527,-116.056401323031,2008-01-28,2008-02-03
-3,35.22267356636723,-118.54506279596853,2008-01-25,2008-01-31
+id,lat,lon,start_date,end_date
+1,34.4147104,-115.0449984,2008-01-21,2008-01-27
+2,34.02192927,-116.0564013,2008-01-28,2008-02-03
+3,35.22267357,-118.5450628,2010-01-25,2010-01-31

test/my_address_file_geocoded_habre_pm_0.2.0.csv

@@ -0,0 +1,4 @@
+id,lat,lon,start_date,end_date,pm,sd
+1,34.4147104,-115.0449984,2008-01-21,2008-01-27,4.39,3.31
+2,34.02192927,-116.0564013,2008-01-28,2008-02-03,3.64,1.65
+3,35.22267357,-118.5450628,2010-01-25,2010-01-31,4.72,2.45