-
Notifications
You must be signed in to change notification settings - Fork 2
/
quick_comparison_r_vs_python_grass_interfaces.qmd
279 lines (213 loc) · 8.56 KB
/
quick_comparison_r_vs_python_grass_interfaces.qmd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
---
title: "Quick comparison: R and Python GRASS packages"
author: "Veronica Andreo"
date: 2024-04-01
date-modified: today
format:
html:
toc: true
code-tools: true
code-copy: true
code-fold: false
categories: [GRASS GIS, Python, R, Intermediate]
engine: knitr
execute:
eval: false
---
In previous tutorials we have gone through some of the basics of using
GRASS GIS through/with R and Python. As you might recall then, there's
an R package called [rgrass](https://github.com/rsbivand/rgrass/) that provides
basic functionality to read and write data from and into GRASS database as well
as to execute GRASS tools in either existing or temporary GRASS projects.
The [GRASS Python API](https://grass.osgeo.org/grass-stable/manuals/libpython/index.html),
on the other hand, is composed of various packages that provide classes and
functions for low and high level tasks, including those that can be executed
with rgrass.
As you might have noticed there are some parallelisms between the
**rgrass** and **grass.script**/**grass.jupyter** packages, i.e.,
R and Python interfaces to GRASS GIS.
In this short tutorial we will go through these similarities in order
to highlight the equivalencies and streamline the use of GRASS extensive
functionality within R and Python communities.
Let's quickly review the equivalencies and go through some examples.
| Task | rgrass function | GRASS Python API function |
|------------------------------------------------------------|--------------------------------|---------------------------------------------------------|
| Load library | library(rgrass) | import grass.script as gs<br>import grass.jupyter as gj |
| Start GRASS and set all needed <br>environmental variables | initGRASS() | gs.setup.init(),<br>gj.init() |
| Execute GRASS commands | execGRASS() | gs.run_command(),<br>gs.read_command(),<br>gs.parse_command() |
| Read raster and vector data <br>from GRASS | read_RAST(),<br>read_VECT() | gs.array.array(),<br>n/a |
| Write raster and vector data<br>into GRASS | write_RAST(),<br>write_VECT() | gs.array.write(),<br>n/a |
| Get raster and vector info | n/a,<br>vInfo() | gs.raster_info(),<br>gs.vector_info() |
| Close GRASS session | unlink_.gislock() | gs.setup.finish(),<br>gj.finish() |
: R and Python GRASS interfaces compared {.striped .hover}
## Comparison examples
Let's see how usage examples would look like.
1. **Load the library**: Either if we are working in R or in Python, we need to
load the libraries that will, in this case, allow us to interface with GRASS GIS
functionality and (optionally) data. For the Python case, we first need to add
the GRASS python package path to our system's path.
::: {.panel-tabset}
## R
```{r}
library(rgrass)
```
## Python
```{python}
#| python.reticulate: FALSE
import sys
import subprocess
sys.path.append(
subprocess.check_output(["grass", "--config", "python_path"], text=True).strip()
)
import grass.script as gs
import grass.jupyter as gj
```
:::
2. **Start a GRASS GIS session**: Once we loaded or imported the packages, we
start a GRASS GIS session. In both cases we need to somehow pass the path to a
temporary or existing GRASS project.
In the case of R, we are also required to pass the path to GRASS GIS binaries,
that in the Python case was needed for library import above.
Here, it is worth noting that while grass.scrip and grass.jupyter init functions
take the same arguments, `gj.init` also sets other environmental variables to
streamline work within Jupyter Notebooks, e.g., overwrite is set to true so cells
can be executed multiple times.
::: {.panel-tabset}
## R
```{r}
session <- initGRASS(gisBase = "/usr/lib/grass83", # where grass binaries live, `grass --config path`
gisDbase = "/home/user/grassdata", # path to grass database or folder where your project lives
location = "nc_basic_spm_grass7", # existing project name
mapset = "PERMANENT" # mapset name
)
```
## Python
```{python}
#| python.reticulate: FALSE
# With grass.script
gs.setup.init(path="/home/user/grassdata",
location="nc_basic_spm_grass7",
mapset="PERMANENT")
# Optionally, the path to a mapset
gs.setup.init("/home/user/grassdata/nc_basic_spm_grass7/PERMANENT")
# With grass.jupyter
session = gj.init(path="/home/user/grassdata",
location="nc_basic_spm_grass7",
mapset="PERMANENT")
# Optionally, the path to a mapset
gj.init("/home/user/grassdata/nc_basic_spm_grass7/PERMANENT")
```
:::
3. **Execute GRASS commands**: Both interfaces work pretty similarly, the
first argument is always the GRASS module name and then we pass the parameters
and flags. While in R we basically use `execGRASS()` for all GRASS commands, in
the Python API, we have different wrappers to execute GRASS commands depending
on the nature of their output.
::: {.panel-tabset}
## R
```{r}
# Map output
execGRASS("r.slope.aspect",
elevation = "terra_elev",
slope = "slope",
aspect = "aspect")
# Text output
execGRASS("g.region",
raster = "elevation",
flags = "p")
```
## Python
```{python}
#| python.reticulate: FALSE
# Map output
gs.run_command("r.slope.aspect",
elevation = "terra_elev",
slope = "slope",
aspect = "aspect")
# Text output
gs.read_command("g.region",
raster = "elevation",
flags = "p")
# Text output - dictionary
gs.parse_command("g.region",
raster = "elevation",
flags = "p")
```
:::
4. **Read raster and vector data into other R or Python formats**:
*rgrass* functions `read_RAST()` and `read_VECT()` convert GRASS raster and
vector maps into terra's SpatRaster and SpatVector objects within R.
In the case of Python, we only have such a functionality for GRASS
raster maps that can be converted into numpy arrays through
`gs.array.array()`. Vector attribute data however can be converted into
pandas data frames using `StringIO` function within pandas `read_csv`.
::: {.panel-tabset}
## R
```{r}
# Raster
elevr <- read_RAST("elevation")
# Vector
schoolsr <- read_VECT("schools")
```
## Python
```{python}
#| python.reticulate: FALSE
# Raster
import numpy as np
elev = gs.array.array("elevation")
# Vector
import pandas as pd
from io import StringIO
schoolsp = gs.read_command("v.db.select", map="schools").strip()
df = pd.read_csv(StringIO(schoolsp), sep="|")
```
:::
5. **Write R or Python objects into GRASS raster and vector maps**: R terra's
SpatRaster and SpatVector objects can be written (back) into GRASS format with
`write_RAST()` and `write_VECT()` functions. Within the Python environment,
numpy arrays can also be written (back) into GRASS raster maps with the
`write()` method.
::: {.panel-tabset}
## R
```{r}
# Raster
write_RAST(elevr, "elevation_r")
# Vector
write_VECT(schoolsr, "schools_r")
```
## Python
```{python}
#| python.reticulate: FALSE
# Raster
elev.write(mapname="elev_np", overwrite=True)
```
:::
6. **Close GRASS GIS session**: In general, just closing R or Rstudio, as well
as shutting down Jupyter notebook, will clean up and close the GRASS session
properly. Sometimes, however, especially if the user changed mapset within the
workflow, it is better to clean up explicitly before closing.
::: {.panel-tabset}
## R
```{r}
unlink_.gislock()
```
## Python
```{python}
#| python.reticulate: FALSE
session.finish()
```
:::
## Final remarks
The examples and comparisons presented here are intended to facilitate the
combination of tools and languages as well as the exchange of data and format
conversions. We hope that's useful as starting point for the implementation
of different use cases and workflows that suit the needs of users.
## References
* [GRASS Python API docs](https://grass.osgeo.org/grass-stable/manuals/libpython/index.html)
* [rgrass docs](https://rsbivand.github.io/rgrass/)
***
:::{.smaller}
The development of this tutorial was funded by the US
[National Science Foundation (NSF)](https://www.nsf.gov/),
award [2303651](https://www.nsf.gov/awardsearch/showAward?AWD_ID=2303651).
:::