Elevatr

==========================

Biggest Changes To Note
- The prj argument relied on proj4string in the past. That support is going
away in order to keep up with changes in PROJ and the rest of R Spatial
ecosystem. Spatial reference will be pulled form "locations" if they have it
defined or will be pulled from an acceptable SRS_string. EPSG codes in the
form of "EPSG:4326" are a pretty safe bet. WKT strings work well on
PROJ > 5.2.0.

Bug Fixes
- Empty rasters were failing as elevatr was using nrow(locations) to get number
of features. It still does that as the default behavior, but if
nrow(locations) returns a null, it uses length(locations) instead. Thanks for
the catch, Gengping Zhu!
- Documentation fix on get_elev_raster, now correctly reports that the function
returns a raster, not points. Thanks AndyBunn!
- sfc objects getting missing in coercion in loc_check. Not anymore!
- epqs occasionally times out, but subsequent hits usually work fine. Added a
second hit when that happens and if that second one doesn't work then it
assigns elevation to NA and throws a warning, instead of erroring
- Was suppressing messages (and thus progress bar) on get_elev_pt src = "aws".
Turned that off so progress of building the DEM is tracked.
- proj_expand was using buffers to expand. Not great for geographic
projections.Now it adds the expansion to the max and subtracts from the min to
expand the bbox by the expand value. For raster retrievals with a single
point the resultant raster will be significant smaller than previous
(approximate 1km by 1km). Multiple points should see no difference. Thanks to
WithRegards on SO for helping me find this.
- In tests with spTransform, changed SRS_string to
CRS(SRS_string=paste0("EPSG:", ll_prj$epsg)). Details in
https://github.com/jhollist/elevatr/issues/56. Thanks to rsbivand and
Fonteh-Bonaventure for helping me with this.
- Raster locations were not returning correctly, that is now fixed.

Added Functionality
- Added access to OpenTopography Global Bathymetry SRTM15+ V2.1 with
src = "srtm15plus"
- serial loop for get_epqs was taking a long time (API returns are slow), so use
furrr::future_map_dbl to paralellize the gets. Defaults to 1 minus available
cores.
- Added argument to get_eqps to control serial vs parallel API calls. Defaults
to serial for 35 or fewer points, but can be set to TRUE for force serial.
- Added overwrite argument to get_elev_point() to check for existence of
elevation and elev_units columns. If either exist and overwrite not TRUE then
errors.
- Updated progress bars to use the progressr package.
- Converted all coordinate reference system handling to pull from locations or
an SRS string. Should take care of running on older versions of PROJ

Other Minor changes
- Removed message that reported out CRS, was too verbose and not necessarily
useful.
- Cleaned up message on units.

==========================
Errors
- updated tests to deal with build errors
- Parsing crs through st_crs()
- Fixed vignettes

==========================

Bug Fixes
- Rasters were not getting handled correctly by size estimation. Thanks to tteo for the catch (https://github.com/jhollist/elevatr/issues/37)
- Single point requests to Open Topography were failing. Expands now to capture small area around single point
- Switched to using wkt instead of proj4. I think this will help with the more recent versions of PROJ...
- Added rgdal_show_exportToProj4_warning=thin to options on load.
- Fixed https://github.com/jhollist/elevatr/issues/38. Thanks to cjcarlson, ACheysson, and jsta for helping track this one down.

==========================

Added Functionality
- Added user agents to httr requests
- Added new OpenTopography API access to three global datasets, SRTM GL1 and 3, and the ALOS World 3D 30m. These are accessed via the src argument
- Added new argument to convert any negative values to NA.

Bug Fixes
- Zoom levels 1 and 0 were throwing errors becuase tile selction was overzealous and was selecting tiles that existed. Conditionals to check fo this. Also zoom 0 returns as "image/tif", not "image/tiff" that all other levls return. More robust checking on return type.
- Tiles for points that fall exactly on the equator were returning NA on `get_elev_point()`. On tile selection in `get_tilexy()` a conditional was added to check for lat == 0 and projeciton being and acceptable proj4 alias of Lat/Long. If that is met a very small (~ 1meter) expansion to the bounding box is done. Thanks willgearty for the bug report <https://github.com/jhollist/elevatr/issues/25>.
- USGS epqs return -1000000 for areas without an elevation. `elevatr` now converts those values to NA. Thanks to George Moroz for the catch! <https://github.com/jhollist/elevatr/issues/24>
- Updated stale sp objects. Thank you Roger Bivand for making the update very easy!
- Updated links to mapzen documentation. Thanks to Alec Robitaille for the fix.
- Sped up merging and projecting via PR from Mike Johnson. Thanks for the contribution, Mike.

Added
- Ask for user confirmation before downloading large files.
- Add reprojection functionality to `get_elev_raster` function and `Raster` class.
- Added the ability to access the resolution of the raster through the `Raster` class via the `raster.resolution`
- Added the ability to access the imagery sources through the `Raster` class via the `raster.imagery_sources`
- Added two parameters to the `show` method of the `Raster` class: `show_extras` to control the display of additional elements, and `file_path` to specify the file path for saving the image.

Added
- Add clipping functionality to `get_elev_raster` function to clip the raster to the bounding box.

Changed
- Improved documentation

Fixed
- Update image links in README
- Removed unnecessary imports and path adjustments in examples.ipynb