migrate gh artifact actions to v4 (#20)

* migrate gh artifact actions to v4

from migration guide no breaking changes apply here.

* updates pipfile.lock dependency versions

* updates CI due to pytest issue see https://github.com/scipy/scipy/issues/22236

* bump to python 3.12

* revert to py3.10

.github/workflows/main.yml

@@ -36,13 +36,19 @@ jobs:
               run: |
                 python setup.py check
                 python setup.py bdist_wheel sdist
-
           - python: "3.10"
             task:
-              name: "Style"
+              name: "Lint"
               run: |
                 black --check .
-
+          - python: "3.10"
+            task:
+              name: "Test"
+              run: pytest --exitfirst --failed-first --assert=plain
+          - python: "3.8"
+            task:
+              name: "Test (3.8)"
+              run: pytest --exitfirst --failed-first --assert=plain
     steps:
       - uses: actions/checkout@v3
 
@@ -59,7 +65,7 @@ jobs:
 
       - name: Upload package distribution files
         if: matrix.task.name == 'Build'
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@v4
         with:
           name: package
           path: dist
@@ -93,7 +99,7 @@ jobs:
           echo "TAG=${GITHUB_REF#refs/tags/}" >> $GITHUB_ENV
 
       - name: Download package distribution files
-        uses: actions/download-artifact@v3
+        uses: actions/download-artifact@v4
         with:
           name: package
           path: dist

.pre-commit-config.yaml

@@ -0,0 +1,10 @@
+repos:
+  - repo: https://github.com/psf/black
+    rev: 22.3.0
+    hooks:
+      - id: black
+        # It is recommended to specify the latest version of Python
+        # supported by your project here, or alternatively use
+        # pre-commit's default_language_version, see
+        # https://pre-commit.com/#top_level-default_language_version
+        language_version: python3.9

Pipfile

@@ -13,7 +13,10 @@ scikit-learn = "*"
 
 [dev-packages]
 black = "*"
+pre-commit = "*"
+pytest = "*"
 wheel = "*"
+geoclustering = {editable = true, path = "."}
 
 [requires]
 python_version = "3.9"

README.md

@@ -10,40 +10,38 @@
 
 ### Clustering Method
 
-A cluster is created when a certain number of points (=> `--size`) each are within a given distance (=> `--distance`) of at least one other point in the cluster. 
+A cluster is created when a certain number of points (defined with `--size`) each are within a given distance (defined with `--distance`) of at least one other point in the cluster. 
 
 
 ## Install
 
-Clone the repository:
+Install with pip:
 
 ```sh
-git clone https://github.com/bellingcat/geoclustering
-cd geoclustering
+# with kepler.gl visualization support
+pip install geoclustering[full]
+
+# only text-based output
+pip install geoclustering
 ```
 
-Install keplergl build dependencies:
+If the `full` install fails, you might need to install kepler.gl build dependencies:
 
 ```sh
 # macos
 brew install proj gdal
 ```
 
-Install project with pip:
-```sh
-pip install .
-```
-
 ## Usage
 
 ```
 Usage: geoclustering [OPTIONS] FILENAME
 
   Tool to cluster geolocations. A cluster is created when a certain number of
-  points (--size) each are within a given distance (--distance) of at least
-  one other point in the cluster. Input is supplied as a csv file. At a
-  minimum, each row needs to have a 'lat' and a 'lon' column. Other rows are
-  reflected to the output.
+  points (defined with --size) each are within a given distance (defined with
+  --distance) of at least one other point in the cluster. Input is supplied as
+  a csv file. At a minimum, each row needs to have a 'lat' and a 'lon' column.
+  Other rows are reflected to the output.
 
 Options:
   -d, --distance FLOAT            (in km) Max. distance between two points in
@@ -64,7 +62,7 @@ Options:
 
 ## Input
 
-Inputs are supplied as a `.csv` file. The only required fields are `lat` and `lon`, all other fields are reflected to the output.
+Inputs are supplied as a `.csv` file. At a minimum, each row needs to have a `lat` and a `lon`` column. Other rows are reflected to the output.
 
 ```csv
 id,name,lat,lon
@@ -74,7 +72,7 @@ id,name,lat,lon
 
 ## Output
 
-If at least one cluster was found, the tool outputs a folder with `json`, `geojson`, `text` and a kepler.gl `html` files.
+If at least one cluster was found, the tool outputs a folder with output as `json`, `geojson`, `txt`, `csv` files. A kepler.gl `html` file is generated as well.
 
 ### JSON
 
@@ -123,7 +121,7 @@ Encodes a single `FeatureCollection`, containing all points as `Feature` objects
 }
 ```
 
-### txt
+### Text
 
 Encodes cluster as blocks separated by a newline, where each line in a cluster block contains one point.
 
@@ -134,6 +132,39 @@ id 9, name Rosanna Foggo, lat -6.2074293, lon 106.8915948
 // ...
 ```
 
+### CSV
+
+Encodes each event in one line with `cluster_id` information associated.
+
+```csv
+cluster_id,name,lat,lon
+9,Rosanna Foggo,-6.2074293,106.8915948
+...
+```
+
 ### kepler.gl
 
 ![kepler.gl instance](https://user-images.githubusercontent.com/1682504/176478177-c0446b51-4060-495c-803d-79e2bbd3e966.png)
+
+## Develop
+
+It is assumed that you are using **Python3.9+**. It is encouraged to [setup a virtualenv](https://wiki.archlinux.org/title/Python/Virtual_environment#venv>) for development.
+
+```sh
+    # install dependencies & dev-dependencies
+    # PIP
+    pip install -e .[dev,full]
+    # PIPENV
+    pipenv install --dev -e .
+
+    # install a git hook that runs the code formatter before each commit.
+    pre-commit install
+```
+
+We use [Black](https://github.com/psf/black) as our code formatter. If you don't want to use the `pre-commit` hook, you can run the formatter manually or via an editor plugin.
+
+## Release
+
+1. Update [version.py](geoclustering/version.py)
+2. Run `scripts/release.sh` 
+3. Confirm GH action completed successfully

geoclustering/__main__.py

@@ -1,6 +1,5 @@
 from pathlib import Path
 import click
-import os
 import webbrowser
 
 import geoclustering.clustering as clustering
@@ -8,12 +7,8 @@ import geoclustering.encoding as encoding
 import geoclustering.io as io
 
 
-def print_debug(s):
-    click.secho(s, fg="bright_black")
-
-
 @click.command(
-    help="Tool to cluster geolocations. A cluster is created when a certain number of points (--size) each are within a given distance (--distance) of at least one other point in the cluster. Input is supplied as a csv file. At a minimum, each row needs to have a 'lat' and a 'lon' column. Other rows are reflected to the output."
+    help="Tool to cluster geolocations. A cluster is created when a certain number of points (defined with --size) each are within a given distance (defined with --distance) of at least one other point in the cluster. Input is supplied as a csv file. At a minimum, each row needs to have a 'lat' and a 'lon' column. Other rows are reflected to the output."
 )
 @click.option(
     "--distance",
@@ -48,40 +43,53 @@ def print_debug(s):
 )
 @click.option(
     "--open",
+    "_open",
     is_flag=True,
     help="Open the generated visualization in the default browser automatically.",
 )
 @click.option("--debug", is_flag=True, help="Print debug output.")
 @click.argument("filename", type=click.Path(exists=True))
-def main(distance, size, output, filename, algorithm, open, debug):
-    if debug:
-        print_debug(f"Reading input from {Path(filename).absolute()}")
+def main(distance, size, output, filename, algorithm, _open, debug):
+    def print_debug(s):
+        if debug:
+            click.secho(s, fg="bright_black")
 
     df = io.read_csv_file(filename)
-    if debug:
-        print_debug(f"Read {len(df)} valid coordinates")
+    print_debug(f"Read {len(df)} valid coordinates from {Path(filename).absolute()}")
 
     clusters = clustering.cluster_locations(
         df=df, algorithm=algorithm, radius_km=distance, min_cluster_size=size
     )
 
     if not bool(clusters):
-        click.echo("Did not find clusters matching input parameters.")
+        click.secho("Did not find clusters matching input parameters.", fg="yellow")
         return
 
     print_debug(f"Found {len(clusters)} valid clusters using {algorithm}")
 
     encoded = encoding.encode_clusters(clusters)
-
     io.write_output_file(output, "result.txt", encoded["string"])
     io.write_output_file(output, "result.json", encoded["json"])
     io.write_output_file(output, "result.geojson", encoded["geojson"])
+    io.write_output_file(output, "result.csv", encoded["csv"])
+
     vis = io.write_visualization(output, "result.html", encoded["geojson"])
+    if vis is None:
+        print_debug("Skipped generating visualization: kepler is not installed.")
+
     click.echo(f"Output files saved to {Path(output).absolute()}")
 
-    if open:
-        print_debug(f"Opening visualization in default browser")
-        webbrowser.open_new_tab("file://" + str(vis.absolute()))
+    if _open:
+        if vis:
+            webbrowser.open_new_tab("file://" + str(vis.absolute()))
+            print_debug("Opened visualization in default browser.")
+        else:
+            click.secho(
+                "Can't open kepler.gl: package not installed. Please re-install geoclustering with `pip install geoclustering[full]`.",
+                fg="yellow",
+            )
+
+    click.secho("Clustering completed.", fg="green")
 
 
 if __name__ == "__main__":

geoclustering/encoding.py

@@ -1,6 +1,8 @@
 import json
 import numpy as np
 import geojson
+import csv
+import io  # not io.py
 
 
 class NpEncoder(json.JSONEncoder):
@@ -47,7 +49,7 @@ class JSONEncoder:
 
         for record in cluster:
             cluster_data["points"].append(record)
-            self.state.append(cluster_data)
+        self.state.append(cluster_data)
 
     def get(self):
         return json.dumps(self.state, cls=NpEncoder)
@@ -74,13 +76,37 @@ class GeoJSONEncoder:
         return json.dumps(geojson.FeatureCollection(self.state), cls=NpEncoder)
 
 
+class CSVEncoder:
+    """Encodes clustering result as a CSV"""
+
+    def __init__(self):
+        self.state = io.StringIO()
+        self.writer = False
+
+    def visitor(self, cluster_id, cluster):
+        if not self.writer:
+            self.writer = csv.DictWriter(
+                self.state,
+                fieldnames=["cluster_id"] + list(cluster[0].keys()),
+                quoting=csv.QUOTE_NONNUMERIC,
+                lineterminator="\n",
+            )
+            self.writer.writeheader()
+
+        for record in cluster:
+            self.writer.writerow({**record, "cluster_id": cluster_id})
+
+    def get(self):
+        return self.state.getvalue()
+
+
 def encode_clusters(clusters):
     json_encoder = JSONEncoder()
     geojson_encoder = GeoJSONEncoder()
     string_encoder = StringEncoder()
+    csv_encoder = CSVEncoder()
 
-    encoders = [json_encoder, geojson_encoder, string_encoder]
-
+    encoders = [json_encoder, geojson_encoder, string_encoder, csv_encoder]
     for cluster_id, cluster in clusters.items():
         for encoder in encoders:
             encoder.visitor(cluster_id, cluster)
@@ -89,4 +115,5 @@ def encode_clusters(clusters):
         "json": json_encoder.get(),
         "geojson": geojson_encoder.get(),
         "string": string_encoder.get(),
+        "csv": csv_encoder.get(),
     }

geoclustering/io.py

@@ -1,4 +1,3 @@
-from keplergl import KeplerGl
 from pathlib import Path
 from pkg_resources import resource_filename
 import json
@@ -7,6 +6,14 @@ import numpy as np
 import os
 import sys
 
+# kepler is optional, check if installed.
+try:
+    from keplergl import KeplerGl
+except:
+    has_kepler = False
+else:
+    has_kepler = True
+
 
 class HiddenPrints:
     """Disables stdout prints for a block of code."""
@@ -38,23 +45,35 @@ def is_valid_lon(val: str) -> bool:
         return False
 
 
+def is_not_none(val: any) -> bool:
+    return val is not None
+
+
 def read_csv_file(filename):
     """Read input csv file, dropping rows that don't have valid location data."""
-    df = pd.read_csv(filename)
-    initial_rows = len(df)
+    # replace NaN for all fields not to break kepler parsing.
+    df = pd.read_csv(filename).replace({np.nan: None})
 
-    df = df.dropna(subset=["lat", "lon"])
-    df = df.replace(
-        {np.nan: None}
-    )  # replace for other fields not to break kepler parsing
-    print(f"Ignored {initial_rows - len(df)} coordinates with NaN")
+    # construct an index of values with valid lat & lon.
+    valid_index = df.lat.apply(is_valid_lat) & df.lon.apply(is_valid_lon)
+    df_invalid = df[~valid_index]
+
+    count_invalid = len(df_invalid)
+    if count_invalid:
+        df_not_empty = df_invalid[
+            (df_invalid.lat.apply(is_not_none) | df_invalid.lon.apply(is_not_none))
+        ]
+
+        count_not_empty = len(df_not_empty)
+        count_empty = count_invalid - count_not_empty
+
+        if count_empty:
+            print(f"Removed {count_empty} empty coordinate pairs.")
+
+        if count_not_empty:
+            print(f"Removed {count_not_empty} invalid coordinate pairs:")
+            print(df_not_empty[["lat", "lon"]].to_string())
 
-    valid_index = df.lat.astype(str).apply(is_valid_lat) & df.lon.astype(str).apply(
-        is_valid_lon
-    )
-    if len(df_invalid := df[~valid_index]):
-        print(f"Found {len(df_invalid)} invalid coordinate pairs, ignoring:")
-        print(df_invalid[["lat", "lon"]].to_string())
     return df[valid_index]
 
 
@@ -77,6 +96,10 @@ def write_output_file(dirname, filename, data):
 
 def write_visualization(dirname, filename, data):
     """Write a visualization, ensuring parent directories."""
+
+    if not has_kepler:
+        return None
+
     # Hide kepler stdout output.
     with HiddenPrints():
         map = KeplerGl()

geoclustering/version.py

@@ -1,8 +1,8 @@
 _MAJOR = "0"
-_MINOR = "2"
+_MINOR = "4"
 # On main and in a nightly release the patch should be one ahead of the last
 # released build.
-_PATCH = "0"
+_PATCH = "1"
 # This is mainly for nightly builds which have the suffix ".dev$DATE". See
 # https://semver.org/#is-v123-a-semantic-version for the semantics.
 _SUFFIX = ""

pytest.ini

@@ -0,0 +1,3 @@
+[pytest]
+testpaths = tests/
+python_files = *.py

setup.py

@@ -28,12 +28,14 @@ setup(
     install_requires=[
         "click",
         "geojson",
-        "keplergl",
         "numpy",
         "pandas",
         "scikit-learn",
     ],
-    extras_require={"dev": ["black", "wheel"]},
+    extras_require={
+        "dev": ["black", "wheel", "pre-commit", "pytest"],
+        "full": ["keplergl"],
+    },
     include_package_data=True,
     zip_safe=False,
 )

tests/clustering.py

@@ -0,0 +1,41 @@
+from geoclustering.clustering import cluster_locations
+from tests.helpers import read_fixture_csv
+
+
+df = read_fixture_csv("clustering.csv")
+
+
+def has_member(list, name):
+    return any(x for x in list if x["name"] == name)
+
+
+def test_clustering_all():
+    # there should be one cluster with all members but Erin.
+    res = cluster_locations(
+        df=df, algorithm="dbscan", radius_km=1.97, min_cluster_size=4
+    )
+    assert len(res.values()) == 1
+    assert len(res[0]) == 4
+
+
+def test_clustering_split():
+    res = cluster_locations(
+        df=df, algorithm="dbscan", radius_km=0.5, min_cluster_size=2
+    )
+    # there should be two cluster: Alice & Bob and Carol & Dan
+    assert len(res.values()) == 2
+    cluster_one = res[0]
+    cluster_two = res[1]
+    assert len(cluster_one) == 2
+    assert has_member(cluster_one, "Alice")
+    assert has_member(cluster_one, "Bob")
+    assert has_member(cluster_two, "Carol")
+    assert has_member(cluster_two, "Dan")
+
+
+def test_clustering_none():
+    # there should be no clusters now.
+    res = cluster_locations(
+        df=df, algorithm="dbscan", radius_km=0.5, min_cluster_size=3
+    )
+    assert len(res.values()) == 0

tests/encoding.py

@@ -0,0 +1,30 @@
+from geoclustering.encoding import encode_clusters
+from tests.helpers import read_fixture_csv, read_fixture_content
+
+
+df = read_fixture_csv("clustering.csv")
+
+
+def test_encoders():
+    clusters = {
+        0: [
+            {"id": 1, "name": "Alice", "lat": 52.523955, "lon": 13.442362},
+            {"id": 2, "name": "Bob", "lat": 52.526659, "lon": 13.448097},
+        ],
+        1: [
+            {"id": 3, "name": "Carol", "lat": 52.525626, "lon": 13.419246},
+            {
+                "id": 4,
+                "name": "Dan",
+                "lat": 52.52443559865125,
+                "lon": 13.41261723049818,
+            },
+        ],
+    }
+
+    res = encode_clusters(clusters)
+
+    assert res["string"] == read_fixture_content("snapshots/result.txt")
+    assert res["json"] == read_fixture_content("snapshots/result.json")
+    assert res["geojson"] == read_fixture_content("snapshots/result.geojson")
+    assert res["csv"] == read_fixture_content("snapshots/result.csv")

tests/fixtures/clustering.csv

@@ -0,0 +1,6 @@
+id,name,lat,lon
+1,Alice,52.523955,13.442362
+2,Bob,52.526659,13.448097
+3,Carol,52.525626,13.419246
+4,Dan,52.52443559865125,13.41261723049818
+5,Erin,52.524838991760774,13.383188597040382

tests/fixtures/io.csv

@@ -0,0 +1,9 @@
+id,name,lat,lon
+1,Alice,,
+2,,52.523955,13.442362
+,,-90.12,132.23
+4,,78.234,-180.1212
+5,Bob,52.524838991760774,13.383188597040382
+6,Peter,91.234,
+7,Horst,,23.23
+7,Erin,foo,bar

tests/fixtures/snapshots/result.csv

@@ -0,0 +1,5 @@
+"cluster_id","id","name","lat","lon"
+0,1,"Alice",52.523955,13.442362
+0,2,"Bob",52.526659,13.448097
+1,3,"Carol",52.525626,13.419246
+1,4,"Dan",52.52443559865125,13.41261723049818

tests/fixtures/snapshots/result.geojson

@@ -0,0 +1 @@
+{"type": "FeatureCollection", "features": [{"type": "Feature", "geometry": {"type": "Point", "coordinates": [13.442362, 52.523955]}, "properties": {"id": 1, "name": "Alice", "cluster_id": 0}}, {"type": "Feature", "geometry": {"type": "Point", "coordinates": [13.448097, 52.526659]}, "properties": {"id": 2, "name": "Bob", "cluster_id": 0}}, {"type": "Feature", "geometry": {"type": "Point", "coordinates": [13.419246, 52.525626]}, "properties": {"id": 3, "name": "Carol", "cluster_id": 1}}, {"type": "Feature", "geometry": {"type": "Point", "coordinates": [13.412617, 52.524436]}, "properties": {"id": 4, "name": "Dan", "cluster_id": 1}}]}

tests/fixtures/snapshots/result.json

@@ -0,0 +1 @@
+[{"cluster_id": 0, "points": [{"id": 1, "name": "Alice", "lat": 52.523955, "lon": 13.442362}, {"id": 2, "name": "Bob", "lat": 52.526659, "lon": 13.448097}]}, {"cluster_id": 1, "points": [{"id": 3, "name": "Carol", "lat": 52.525626, "lon": 13.419246}, {"id": 4, "name": "Dan", "lat": 52.52443559865125, "lon": 13.41261723049818}]}]

tests/fixtures/snapshots/result.txt

@@ -0,0 +1,7 @@
+Cluster 0
+id 1, name Alice, lat 52.523955, lon 13.442362
+id 2, name Bob, lat 52.526659, lon 13.448097
+
+Cluster 1
+id 3, name Carol, lat 52.525626, lon 13.419246
+id 4, name Dan, lat 52.52443559865125, lon 13.41261723049818

tests/helpers.py

@@ -0,0 +1,16 @@
+import os
+from geoclustering.io import read_csv_file
+
+
+def get_fixture_path(filename):
+    dir_path = os.path.dirname(os.path.realpath(__file__))
+    return os.path.join(dir_path, "fixtures", filename)
+
+
+def read_fixture_csv(filename):
+    return read_csv_file(get_fixture_path(filename))
+
+
+def read_fixture_content(filename):
+    with open(get_fixture_path(filename)) as f:
+        return f.read()

tests/io.py

@@ -0,0 +1,25 @@
+from pathlib import Path
+import shutil
+from geoclustering.io import write_output_file
+from tests.helpers import read_fixture_csv
+
+
+def test_csv_filters():
+    df = read_fixture_csv("io.csv")
+    # entries 2 & 5 in fixture are valid.
+    assert len(df) == 2
+    assert df.iloc[0]["name"] == None
+    assert df.iloc[1]["name"] == "Bob"
+
+
+def test_write_output_file():
+    p = "./this/dir/does/not/exist"
+    f = "test.txt"
+    write_output_file(p, f, "test")
+
+    path = Path(p) / f
+
+    with open(path) as f:
+        assert f.read() == "test"
+
+    shutil.rmtree(Path("./this"))