Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 8 Next »

Installing Python and Jupyter Notebook

You can download and install Python from https://www.python.org/downloads/

Check Add python.exe to PATH, then Install Now.

To confirm installation, from a command prompt enter:

py --version
python --version

Both commands should show the version that you downloaded.

Jupyter Notebook is a popular tool for running Python scripts, offering the added benefit of displaying plots directly within the interface. You can install it by running:

pip install notebook

You can then run:

jupyter lab /path/to/folder

Basic conversions

R

Python

Arithmetic operators

x + y
x - y
x * y
x / y
x ^ y
x + y
x - y
x * y
x / y
x ** y

The only difference is that R uses ^ for exponentiation, while Python uses **.

Comparison operators

# R and Python
x == y
x != y
x > y
x < y
x >= y
x <= y

No differences here.

Logical operators


x && y
x || y
!x
v1 & v2
v1 | v2
import numpy as np

x and y
x or y
not x
np.logical_and(v1, v2)
np.logical_or(v1, v2)

Python doesn’t have built-in logical operations for vectors.

numpy is a popular library for this purpose, but it’s not as clean as R’s syntax.

Variable assignment

name <- "John Doe"

name = "John Doe"

R does accept the = syntax for assignment. However, it’s non-idiomatic and can lead to unexpected errors.

For example, 

my_function <- function(y = 3, x = 5) {
  print(paste(y, x))
}

my_function(x = 10)   # Passes 10 to the x parameter, and outputs "3 10"

my_function(x <- 20)  # Assigns 20 to the global x variable, which it passes to the y parameter. Outputs "20 5"

To avoid confusion, always use <- for assignment, and use = for passing arguments within function calls.

Ranges

1:10
seq(1, 20, 2)
print(1:10)
range(1, 11)
range(1, 21, 2)
print(list(range(1, 11)))

Ouch. Python’s code is not only longer, but it excludes the final number.
range(1, 11) creates a range object in Python, which represents the numbers 1 through 10.

list(range(1, 11)) creates a list object in Python, which represents [1, 2, 3, … 10]
Both function similarly to the 1:10 vector in R. However, you need to change a range to a list in order to output it.

Conditional Statements

if (x > 5) {
  print("x is greater than 5")
} else if (x >= 2) {
  print("x is between 2 and 5")
} else {
  print("x is less than 2")
}
if x > 5:
    print("x is greater than 5")
elif x >= 2:
    print("x is between 2 and 5")
else:
    print("x is less than 2")

R requires the condition to be in parentheses. That would also work in Python, but it’s not required.

R begins a block statement with an opening curly brace ({), and it ends it with a closing curly brace (}).

Python is the oddball here, using a colon (:) to begin a block statement. It then uses indentation for the block’s statements. To end a block, you stop indenting.

This can take some getting used to, because very few computer languages have similar syntax.

Note that the recommended indentation for Python is 4 spaces, but consistency is all that matters.

For Loops

for (i in 1:5) {
  print(i)
}
for i in range(1, 6):
    print(i)

Note the similarity to the if statement comparisons above – particularly the absence of parentheses in the iteration expression, and the block syntax.

Functions

my_function <- function(x, y) {
  return(x + y)
}
def my_function(x, y):
    return x + y

Indexing

my_list[0]
my_list[1]

R uses 1-based indexing.

Python uses 0-based indexing.

R is actually the oddball here.

Example conversions

The following examples demonstrate how to convert the R code at https://github.com/jjennewein/MDA_2023-24 into equivalent Python code.

0.field buffer.R

0.field buffer.py

library(sf)
import geopandas as gpd

Loads a library for handling geospatial data, including working with shapefiles.

In the Python code, gpd is an alias for the geopandas library, which we use to reduce typing and improve code readability.

setwd(dirname(rstudioapi::getActiveDocumentContext()$path))

RStudio defaults to a global working directory, which means file paths must be specified explicitly.

Python defaults to the folder where the source code is located.

Jupyter Notebook defaults to where you launch the notebook server.

  • You could launch it from the command line within the source folder:

    cd \psa\MDA_2023-24\python
jupyter notebook

  • Or you could specify the folder directly when launching Jupyter:

    jupyter notebook \psa\MDA_2023-24\python

The original file path was "E:/Eastern_Shore/MD_enrollment_data_2023_24/fields/", but I don't have access to that location.

To avoid referencing "E:/Eastern_Shore..." explicitly, I saved all the necessary project files to the same folder as the source code, and I added a setwd() statement to the R source. I then removed all references to "E:/Eastern_Shore..."

This way, the R program can run without requiring the full path in filenames.

mda = st_read("MDAEnrollment2023-2024.shp") %>%
  st_transform(., crs = 32618) %>%
  st_buffer(., dist = -30) %>%
  dplyr::filter(!st_is_empty(.))
mda = gpd.read_file("MDAEnrollment2023-2024.shp") \
      .to_crs(epsg=32618)
mda["geometry"] = mda.geometry.buffer(-30)
mda = mda[~mda.geometry.is_empty]

This code:

  • Reads a shapefile into a spatial data frame (R) or a GeoDataFrame (Python).

  • Converts the coordinate reference system (CRS) to EPSG 32618 (WGS 84).

  • Applies a negative buffer of 30 meters to shrink the geometries.
    Note that sf objects in R have a designated geometry column, but in GeoPandas you need to explicitly reference the geometry column.

  • Filters out any empty geometries resulting from the buffering.
    In Python, ~ is a bitwise NOT operator.

R uses %>% for chaining operations, while Python uses . (dot notation).

In Python, a backslash (\) means line continuation. I used it here to align the code with the R version. But I could have also written it as:
mda = gpd.read_file("MDAEnrollment2023-2024.shp").to_crs(epsg=32618)

Also, note that the dots in st_transform(., crs = 32618) and st_buffer(., dist = -30) are optional but may improve readability.

st_write(mda, "MDAEnrollment2023-2024_Buff30m.shp", append=FALSE)
mda.to_file("MDAEnrollment2023-2024_Buff30m.shp")

Writes the buffered data frame to a new shapefile.

I added append=FALSE to the original source so the file would be overwritten if it exists. GeoPandas automatically overwrites existing files.

1.Extract_HLS_L30 .R

1.Extract_HLS_L30 .py

library(dplyr)
library(terra)
library(sf)
library(exactextractr)
import pandas as pd
import rasterio
import geopandas as gpd
from rasterstats import zonal_stats

Syntax is easy. Knowing which libraries to use – not so much!

setwd(dirname(rstudioapi::getActiveDocumentContext()$path))

Explained previously.

fields = st_read("MDAEnrollment2023-2024_Buff30m.shp")
fields = gpd.read_file("MDAEnrollment2023-2024_Buff30m.shp")

Reads the buffered shapefile.

plot(fields[1])
fields.plot(edgecolor='black')

Plots the shapefile.

edgecolor isn’t required in the Python code. It simply makes the output consistent with the R output.

  • No labels

0 Comments

You are not logged in. Any changes you make will be marked as anonymous. You may want to Log In if you already have an account.