195 update documentation and dependencies

CONTRIBUTING.md

@@ -43,9 +43,7 @@ Module names come from the names of the .py files containing function declaratio
 
 - Try to create modules in a way that each module contains only one functionality. Split this functionality into two function declarations: one for external use and one (the core functionality) for internal use. See e.g. implementation of [clipping functionality](./eis_toolkit/raster_processing/clipping.py) for reference.
 
-- For large or complex functionalities, it is okay to include multiple (helper) functions in one module/file. If you have a moderate amount of functions, you can put them in one file, but in case several helper functions are needed (and they are not general and don't belong in the utilities module), you can create a secondary file for your functionality, for example `clipping_functions.py` or `clipping_utilities.py` for `clipping.py`.
-
-3. Functions
+1. Functions
 
 Name each function according to what it is supposed to do. Try to express the purpose as simplistic as possible. In principle, each function should be creted for executing one task. We prefer modular structure and low hierarchy by trying to avoid nested function declarations. It is highly recommended to call other functions for executing sub tasks.
 

README.md

@@ -29,77 +29,17 @@ If you are contributing by implementing new functionalities, read the **For deve
 
 ## For developers
 
-### Prerequisites
-
 All contributing developers need git, and a copy of the repository.
 
 ```console
 git clone https://github.com/GispoCoding/eis_toolkit.git
 ```
 
 After this you have three options for setting up your local development environment.
-1. Docker
-2. Python venv
-3. Conda
-
-Docker is recommended as it containerizes the whole development environment, making sure it stays identical across different developers and operating systems. Using a container also keeps your own computer clean of all dependencies.
-
-### Setting up a local development environment with docker (recommended)
-Build and run the eis_toolkit container. Run this and every other command in the repository root unless otherwise directed.
-
-```console
-docker compose up -d
-```
-
-If you need to rebuild already existing container (e.g. dependencies have been updated), run
-
-```console
-docker compose up -d --build
-```
-
-### Working with the container
-
-Attach to the running container
-
-```console
-docker attach eis_toolkit
-```
-
-You are now in your local development container, and all your commands in the current terminal window interact with the container.
-
-**Note** that your local repository gets automatically mounted into the container. This means that:
-- The repository in your computer's filesystem and in the container are exactly the same
-- Changes from either one carry over to the other instantly, without any need for restarting the container
-
-For your workflow this means that:
-- You can edit all files like you normally would (on your own computer, with your favourite text editor etc.)
-- You must do all testing and running the code inside the container
+1. Docker - [instructions](./instructions/dev_setup_with_docker.md)
+2. Poetry - [instructions](./instructions/dev_setup_without_docker.md)
+3. Conda - [instructions](./instructions/dev_setup_without_docker_with_conda.md)
 
-### Python inside the container
-
-Whether or not using docker we manage the python dependencies with poetry. This means that a python venv is found in the container too. Inside the container, you can get into the venv like you normally would
-
-```console
-poetry shell
-```
-
-and run your code and tests from the command line. For example:
-
-```console
-python <path/to/your/file.py>
-```
-
-or
-
-```console
-pytest
-```
-
-You can also run commands from outside the venv, just prefix them with poetry run. For example:
-
-```console
-poetry run pytest
-```
 
 ### Additonal instructions
 
@@ -108,10 +48,8 @@ Here are some additional instructions related to the development of EIS toolkit:
 - [Generating documentation](./instructions/generating_documentation.md)
 - [Using jupyterlab](./instructions/using_jupyterlab.md)
 
-If you want to set up the development environment without docker, see:
-- [Setup without docker with poetry](./instructions/dev_setup_without_docker.md)
-- [Setup without docker with conda](./instructions/dev_setup_without_docker_with_conda.md)
-
+## For users
+TBD when first release is out.
 
 ## License
 

eis_toolkit/exploratory_analyses/k_means_cluster.py

@@ -18,17 +18,17 @@ def _k_means_clustering(
         # The elbow method
         k_max = 10
         inertia = np.array(
-            [KMeans(n_clusters=k, random_state=0).fit(coordinates).inertia_ for k in range(1, k_max + 1)]
+            [KMeans(n_clusters=k, random_state=0, n_init=10).fit(coordinates).inertia_ for k in range(1, k_max + 1)]
         )
 
         inertia = np.diff(inertia, 2)
         scaled_derivatives = [i * 100 for i in inertia]
         k_optimal = scaled_derivatives.index(min(scaled_derivatives))
 
-        kmeans = KMeans(n_clusters=k_optimal, random_state=random_state)
+        kmeans = KMeans(n_clusters=k_optimal, random_state=random_state, n_init=10)
 
     else:
-        kmeans = KMeans(n_clusters=number_of_clusters, random_state=random_state)
+        kmeans = KMeans(n_clusters=number_of_clusters, random_state=random_state, n_init=10)
 
     kmeans.fit(coordinates)
     data["cluster"] = kmeans.labels_

instructions/dev_setup_with_docker.md

@@ -0,0 +1,57 @@
+### Development with Docker
+
+Build and run the eis_toolkit container. Run this and every other command in the repository root unless otherwise directed.
+
+```console
+docker compose up -d
+```
+
+If you need to rebuild already existing container (e.g. dependencies have been updated), run
+
+```console
+docker compose up -d --build
+```
+
+### Working with the container
+
+Attach to the running container
+
+```console
+docker attach eis_toolkit
+```
+
+You are now in your local development container, and all your commands in the current terminal window interact with the container.
+
+**Note** that your local repository gets automatically mounted into the container. This means that:
+- The repository in your computer's filesystem and in the container are exactly the same
+- Changes from either one carry over to the other instantly, without any need for restarting the container
+
+For your workflow this means that:
+- You can edit all files like you normally would (on your own computer, with your favourite text editor etc.)
+- You must do all testing and running the code inside the container
+
+### Python inside the container
+
+Whether or not using docker we manage the python dependencies with poetry. This means that a python venv is found in the container too. Inside the container, you can get into the venv like you normally would
+
+```console
+poetry shell
+```
+
+and run your code and tests from the command line. For example:
+
+```console
+python <path/to/your/file.py>
+```
+
+or
+
+```console
+pytest
+```
+
+You can also run commands from outside the venv, just prefix them with poetry run. For example:
+
+```console
+poetry run pytest
+```

instructions/dev_setup_without_docker.md

@@ -1,5 +1,5 @@
-# Development without docker
-If you do not have docker, you can setup your local development environment as a python virtual environment.
+# Development with Poetry
+If you do not have docker, you can setup your local development environment as a python virtual environment using Poetry.
 
 ## Prerequisites
 

pyproject.toml

@@ -18,7 +18,7 @@ keywords = [
 ]
 
 [tool.poetry.dependencies]
-python = ">=3.8,<3.11"
+python = ">=3.9,<3.11"
 gdal = "3.4.3"
 rasterio = "^1.3.0"
 pandas = "^1.4.3"
@@ -29,7 +29,6 @@ statsmodels = "^0.13.2"
 keras = "^2.9.0"
 tensorflow = "^2.9.1"
 mkdocs-material = "^8.4.0"
-plotly = "^5.14.0"
 beartype = "^0.13.1"
 seaborn = "^0.12.2"
 pykrige = "^1.7.0"

tests/conversions/raster_to_dataframe_test.py

@@ -2,7 +2,6 @@
 
 import numpy as np
 import pandas as pd
-import pytest
 import rasterio
 
 from eis_toolkit.conversions.raster_to_dataframe import raster_to_dataframe
@@ -11,7 +10,6 @@
 test_dir = Path(__file__).parent.parent
 
 
-@pytest.mark.skip
 def test_raster_to_dataframe():
     """Test raster to pandas conversion by converting pandas dataframe and then back to raster data."""
     raster = rasterio.open(SMALL_RASTER_PATH)
@@ -32,7 +30,7 @@ def test_raster_to_dataframe():
     """Convert back to raster image."""
     df["id"] = df.index
     long_df = pd.wide_to_long(df, ["band_"], i="id", j="band").reset_index()
-    long_df.loc[:, ["col", "row"]] = long_df.loc[:, ["col", "row"]].astype(int)
+    long_df = long_df.astype({"col": int, "row": int})
     raster_img = np.empty((multiband_raster.count, multiband_raster.height, multiband_raster.width))
     raster_img[(long_df.band - 1).to_list(), long_df.row.to_list(), long_df.col.to_list()] = long_df.band_