clustereasy 0.1.0__tar.gz

clustereasy-0.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Shashikant
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

clustereasy-0.1.0/PKG-INFO

@@ -0,0 +1,305 @@
+Metadata-Version: 2.4
+Name: clustereasy
+Version: 0.1.0
+Summary: A beginner-friendly, consistent API for K-Means and Agglomerative Clustering on pandas DataFrames.
+Author: Shashikant
+License-Expression: MIT
+Project-URL: Homepage, https://github.com/shashikant-ds636/easycluster
+Project-URL: Issues, https://github.com/shashikant-ds636/easycluster/issues
+Keywords: clustering,kmeans,hierarchical,machine-learning,pandas
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Science/Research
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Scientific/Engineering
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: pandas>=1.5
+Requires-Dist: numpy>=1.23
+Requires-Dist: scikit-learn>=1.2
+Requires-Dist: scipy>=1.9
+Requires-Dist: matplotlib>=3.6
+Provides-Extra: dev
+Requires-Dist: pytest>=7; extra == "dev"
+Requires-Dist: build; extra == "dev"
+Requires-Dist: twine; extra == "dev"
+Dynamic: license-file
+
+# clustereasy
+
+A beginner-friendly, consistent Python API that combines **K-Means** and
+**Agglomerative (Hierarchical) Clustering** on pandas DataFrames into one
+class: `Cluster`.
+
+Instead of re-learning different scikit-learn/SciPy calling conventions for
+each algorithm, you get one object with one set of methods for fitting,
+visualizing, evaluating, and exporting results — for both algorithms.
+
+## 1. Why clustereasy
+
+- One consistent interface for two different clustering algorithms.
+- Keeps your original DataFrame untouched.
+- Clear, explicit errors instead of silently doing the wrong thing.
+- Advanced users can still pass through scikit-learn parameters.
+
+## 2. Installation
+
+Local development install (editable):
+
+```bash
+git clone https://github.com/shashikant-ds636/easycluster.git
+cd easycluster
+pip install -e ".[dev]"
+```
+
+Once published, it will be installable with:
+
+```bash
+pip install clustereasy
+```
+
+## 3. Requirements
+
+- Python >= 3.9
+- pandas >= 1.5
+- numpy >= 1.23
+- scikit-learn >= 1.2
+- scipy >= 1.9
+- matplotlib >= 3.6
+
+## 4. Quick start
+
+```python
+import pandas as pd
+from clustereasy import Cluster
+
+df = pd.read_csv("customers.csv")
+
+cluster = Cluster(
+    data=df,
+    features=["Annual Income", "Spending Score"],
+    scale=True,
+    missing="median",
+)
+
+cluster.elbow_method(max_clusters=10)
+cluster.fit_kmeans(n_clusters=4, random_state=42)
+cluster.plot_clusters(x="Annual Income", y="Spending Score")
+
+print(cluster.evaluate())
+result_df = cluster.get_clustered_data(column_name="Segment")
+```
+
+## 5. DataFrame input requirements
+
+`Cluster(data=df, features=[...], scale=..., missing=...)`:
+
+- `data` must be a non-empty `pandas.DataFrame`. It is never modified.
+- `features`: optional list of column names to cluster on. Defaults to all
+  columns. Columns must exist, be unique, and be numeric — clustereasy will
+  **not** silently convert text/categorical values into numbers. Encode
+  categorical columns yourself first (e.g. `pd.get_dummies`) if you want to
+  include them.
+- `missing`: one of `"raise"` (default), `"drop"`, `"mean"`, `"median"`.
+  Controls how missing values in the selected features are handled.
+- Infinite values always raise an explicit error — they are never silently
+  dropped or converted.
+- `scale`: `False` (default), `True` (same as `"standard"`), or one of
+  `"standard"`, `"minmax"`, `"robust"`.
+
+## 6. K-Means tutorial
+
+```python
+cluster = Cluster(data=df, features=["Income", "Score"], scale=True)
+
+# Explore how many clusters might make sense
+elbow_table = cluster.elbow_method(max_clusters=10, random_state=42)
+
+# Fit
+cluster.fit_kmeans(n_clusters=4, init="k-means++", n_init=10, random_state=42)
+
+cluster.cluster_centers()      # DataFrame, original feature scale by default
+cluster.inertia()              # WCSS for the fitted model
+cluster.predict(new_df)        # assign new rows to existing clusters
+```
+
+Supported `fit_kmeans` parameters: `n_clusters`, `init`, `n_init`, `max_iter`,
+`tol`, `random_state`, `algorithm` — all pass straight through to
+scikit-learn's `KMeans`.
+
+## 7. Elbow Method
+
+`elbow_method(max_clusters=10, random_state=42, show=True)` fits K-Means for
+k = 1..max_clusters, plots k vs. WCSS (inertia), and returns a DataFrame of
+the results so you can reuse them without refitting.
+
+**The elbow is a heuristic**, not a guaranteed optimum — use it together with
+`silhouette_score()` and your domain knowledge, not in isolation.
+
+## 8. Hierarchical Clustering tutorial
+
+```python
+Z = cluster.plot_dendrogram(method="ward", metric="euclidean")
+cluster.fit_agglomerative(n_clusters=3, linkage="ward")
+```
+
+Supported linkage methods: `ward`, `complete`, `average`, `single`.
+`ward` linkage requires `metric="euclidean"` — using any other metric with
+`ward` raises a clear error instead of silently producing wrong results.
+
+You can also cluster by distance instead of a fixed cluster count:
+
+```python
+cluster.fit_agglomerative(n_clusters=None, distance_threshold=5.0)
+```
+
+`distance_threshold` and `n_clusters` are mutually exclusive, matching
+scikit-learn's own requirement.
+
+**Important:** `AgglomerativeClustering` has no native `predict()` method in
+scikit-learn, and `Cluster.predict()` will raise an explicit error if called
+on an agglomerative model rather than faking support for it.
+
+## 9. Dendrogram interpretation
+
+- Each leaf is one observation.
+- Each join (inverted "U") is a merge of two clusters; its height is the
+  distance at which they were merged.
+- Cutting the dendrogram horizontally at a given height determines how many
+  clusters you get at that distance.
+- For large datasets, pass `truncate_mode="lastp"` with a small `p` to avoid
+  an unreadable plot — clustereasy will warn you above ~2000 samples if you
+  don't.
+
+## 10. Evaluation metrics and their limitations
+
+All metrics below work without ground-truth labels:
+
+| Method | Notes |
+|---|---|
+| `silhouette_score()` | Needs 2..n-1 distinct clusters. |
+| `silhouette_samples()` | Per-sample values, same constraints. |
+| `davies_bouldin_score()` | Lower is better. Needs >= 2 clusters. |
+| `calinski_harabasz_score()` | Higher is better. Needs >= 2 clusters. |
+| `inertia()` | K-Means only — not available for Agglomerative models. |
+| `evaluate()` | One DataFrame with every metric applicable to the fitted model. |
+
+These are **diagnostic tools**, not proof of a "correct" clustering — always
+sanity-check results against domain knowledge. If a metric can't be computed
+(e.g. only 1 distinct cluster), `evaluate()` shows `NaN` with a note instead
+of crashing; the individual metric methods raise a `ParameterError`.
+
+## 11. Feature scaling and missing-value handling
+
+- Distances (and therefore both K-Means and Agglomerative results) are scale
+  -sensitive. If your features are on different scales (e.g. income in
+  thousands vs. a 1-100 score), set `scale=True`.
+- K-Means cluster centers are available both in the transformed (scaled)
+  space (`cluster._model.cluster_centers_`) and inverse-transformed back to
+  the original feature scale (`cluster.cluster_centers()`, the default).
+- Missing values are never silently dropped by default — you must choose
+  `missing="drop"/"mean"/"median"` explicitly, or clean the data yourself.
+
+## 12. Visualization examples
+
+```python
+cluster.plot_clusters(x="Income", y="Score", show_centers=True)
+cluster.plot_cluster_sizes()
+cluster.plot_silhouette()
+cluster.plot_pca()   # for > 2 features; clearly labeled as a 2D projection
+```
+
+`plot_clusters` requires exactly two selected feature columns to display;
+for higher-dimensional data use `plot_pca()` instead, which is explicitly
+labeled as a projection, not the original feature space.
+
+## 13. Cluster labels and DataFrame output
+
+```python
+labels = cluster.get_labels()                       # pd.Series, original index
+clustered_df = cluster.get_clustered_data("Segment") # original df + label column
+summary = cluster.cluster_summary()                  # size, proportion, per-feature mean
+```
+
+Cluster IDs (0, 1, 2, ...) are arbitrary identifiers — they carry no
+inherent ranking or meaning.
+
+## 14. Common errors and solutions
+
+| Error | Likely cause | Fix |
+|---|---|---|
+| `InputValidationError: not numeric` | Selected a text/categorical column | Encode it first, or exclude it from `features` |
+| `InputValidationError: missing values` | NaNs present, `missing="raise"` (default) | Pass `missing="drop"/"mean"/"median"` |
+| `ParameterError: Ward linkage requires metric='euclidean'` | `linkage="ward"` with another metric | Use `metric="euclidean"` or a different linkage |
+| `NotFittedError` | Called an evaluation/plot/output method before `fit_kmeans`/`fit_agglomerative` | Fit a model first |
+| `ParameterError` on `predict()` | Called `predict()` on an Agglomerative model | Not supported — only K-Means supports `predict()` |
+
+## 15. Computational limitations
+
+- Dendrograms require O(n²) memory/time for the distance matrix and are
+  impractical much past a few thousand samples.
+- `AgglomerativeClustering` scales worse than `KMeans` for large n.
+- No GPU acceleration; this wraps scikit-learn/SciPy CPU implementations.
+
+## 16. API reference (summary)
+
+**Construction:** `Cluster(data, features=None, scale=False, missing="raise")`
+
+**K-Means:** `elbow_method`, `fit_kmeans`, `cluster_centers`, `inertia`, `predict`
+
+**Agglomerative:** `plot_dendrogram`, `get_linkage_matrix`, `fit_agglomerative`
+
+**Shared:** `plot_clusters`, `plot_cluster_sizes`, `plot_silhouette`, `plot_pca`
+
+**Evaluation:** `silhouette_score`, `silhouette_samples`, `davies_bouldin_score`,
+`calinski_harabasz_score`, `evaluate`
+
+**Output:** `get_labels`, `get_clustered_data`, `cluster_summary`, `elbow_results`
+
+## 17. Testing
+
+```bash
+pip install -e ".[dev]"
+pytest -q
+```
+
+Current suite: **59 tests, all passing**, covering validation, both
+algorithms, visualization, evaluation edge cases, and DataFrame-output
+guarantees (index preservation, immutability, reproducibility).
+
+## 18. Contribution guidelines
+
+1. Fork, branch from `main`.
+2. Add/update tests for any behavior change.
+3. Run `pytest -q` before opening a PR.
+4. Keep public methods documented with docstrings.
+
+## 19. Roadmap
+
+- `DBSCAN`, `OPTICS`, `GaussianMixture` support behind the same API.
+- Automatic comparison across multiple `k` / linkage configurations.
+- Model save/load (`joblib`).
+- Export results directly to CSV.
+- Optional connectivity-constrained agglomerative clustering helpers.
+
+## 20. Preparing a PyPI release
+
+```bash
+pip install build twine
+python -m build                 # creates dist/*.whl and dist/*.tar.gz
+twine check dist/*
+pip install dist/clustereasy-0.1.0-py3-none-any.whl  # test locally
+twine upload --repository testpypi dist/*             # TestPyPI first
+twine upload dist/*                                    # then real PyPI
+```
+
+Bump the `version` in `pyproject.toml` before each release and update
+`CHANGELOG.md`.
+
+## License
+
+MIT — see `LICENSE`.

clustereasy-0.1.0/README.md

@@ -0,0 +1,274 @@
+# clustereasy
+
+A beginner-friendly, consistent Python API that combines **K-Means** and
+**Agglomerative (Hierarchical) Clustering** on pandas DataFrames into one
+class: `Cluster`.
+
+Instead of re-learning different scikit-learn/SciPy calling conventions for
+each algorithm, you get one object with one set of methods for fitting,
+visualizing, evaluating, and exporting results — for both algorithms.
+
+## 1. Why clustereasy
+
+- One consistent interface for two different clustering algorithms.
+- Keeps your original DataFrame untouched.
+- Clear, explicit errors instead of silently doing the wrong thing.
+- Advanced users can still pass through scikit-learn parameters.
+
+## 2. Installation
+
+Local development install (editable):
+
+```bash
+git clone https://github.com/shashikant-ds636/easycluster.git
+cd easycluster
+pip install -e ".[dev]"
+```
+
+Once published, it will be installable with:
+
+```bash
+pip install clustereasy
+```
+
+## 3. Requirements
+
+- Python >= 3.9
+- pandas >= 1.5
+- numpy >= 1.23
+- scikit-learn >= 1.2
+- scipy >= 1.9
+- matplotlib >= 3.6
+
+## 4. Quick start
+
+```python
+import pandas as pd
+from clustereasy import Cluster
+
+df = pd.read_csv("customers.csv")
+
+cluster = Cluster(
+    data=df,
+    features=["Annual Income", "Spending Score"],
+    scale=True,
+    missing="median",
+)
+
+cluster.elbow_method(max_clusters=10)
+cluster.fit_kmeans(n_clusters=4, random_state=42)
+cluster.plot_clusters(x="Annual Income", y="Spending Score")
+
+print(cluster.evaluate())
+result_df = cluster.get_clustered_data(column_name="Segment")
+```
+
+## 5. DataFrame input requirements
+
+`Cluster(data=df, features=[...], scale=..., missing=...)`:
+
+- `data` must be a non-empty `pandas.DataFrame`. It is never modified.
+- `features`: optional list of column names to cluster on. Defaults to all
+  columns. Columns must exist, be unique, and be numeric — clustereasy will
+  **not** silently convert text/categorical values into numbers. Encode
+  categorical columns yourself first (e.g. `pd.get_dummies`) if you want to
+  include them.
+- `missing`: one of `"raise"` (default), `"drop"`, `"mean"`, `"median"`.
+  Controls how missing values in the selected features are handled.
+- Infinite values always raise an explicit error — they are never silently
+  dropped or converted.
+- `scale`: `False` (default), `True` (same as `"standard"`), or one of
+  `"standard"`, `"minmax"`, `"robust"`.
+
+## 6. K-Means tutorial
+
+```python
+cluster = Cluster(data=df, features=["Income", "Score"], scale=True)
+
+# Explore how many clusters might make sense
+elbow_table = cluster.elbow_method(max_clusters=10, random_state=42)
+
+# Fit
+cluster.fit_kmeans(n_clusters=4, init="k-means++", n_init=10, random_state=42)
+
+cluster.cluster_centers()      # DataFrame, original feature scale by default
+cluster.inertia()              # WCSS for the fitted model
+cluster.predict(new_df)        # assign new rows to existing clusters
+```
+
+Supported `fit_kmeans` parameters: `n_clusters`, `init`, `n_init`, `max_iter`,
+`tol`, `random_state`, `algorithm` — all pass straight through to
+scikit-learn's `KMeans`.
+
+## 7. Elbow Method
+
+`elbow_method(max_clusters=10, random_state=42, show=True)` fits K-Means for
+k = 1..max_clusters, plots k vs. WCSS (inertia), and returns a DataFrame of
+the results so you can reuse them without refitting.
+
+**The elbow is a heuristic**, not a guaranteed optimum — use it together with
+`silhouette_score()` and your domain knowledge, not in isolation.
+
+## 8. Hierarchical Clustering tutorial
+
+```python
+Z = cluster.plot_dendrogram(method="ward", metric="euclidean")
+cluster.fit_agglomerative(n_clusters=3, linkage="ward")
+```
+
+Supported linkage methods: `ward`, `complete`, `average`, `single`.
+`ward` linkage requires `metric="euclidean"` — using any other metric with
+`ward` raises a clear error instead of silently producing wrong results.
+
+You can also cluster by distance instead of a fixed cluster count:
+
+```python
+cluster.fit_agglomerative(n_clusters=None, distance_threshold=5.0)
+```
+
+`distance_threshold` and `n_clusters` are mutually exclusive, matching
+scikit-learn's own requirement.
+
+**Important:** `AgglomerativeClustering` has no native `predict()` method in
+scikit-learn, and `Cluster.predict()` will raise an explicit error if called
+on an agglomerative model rather than faking support for it.
+
+## 9. Dendrogram interpretation
+
+- Each leaf is one observation.
+- Each join (inverted "U") is a merge of two clusters; its height is the
+  distance at which they were merged.
+- Cutting the dendrogram horizontally at a given height determines how many
+  clusters you get at that distance.
+- For large datasets, pass `truncate_mode="lastp"` with a small `p` to avoid
+  an unreadable plot — clustereasy will warn you above ~2000 samples if you
+  don't.
+
+## 10. Evaluation metrics and their limitations
+
+All metrics below work without ground-truth labels:
+
+| Method | Notes |
+|---|---|
+| `silhouette_score()` | Needs 2..n-1 distinct clusters. |
+| `silhouette_samples()` | Per-sample values, same constraints. |
+| `davies_bouldin_score()` | Lower is better. Needs >= 2 clusters. |
+| `calinski_harabasz_score()` | Higher is better. Needs >= 2 clusters. |
+| `inertia()` | K-Means only — not available for Agglomerative models. |
+| `evaluate()` | One DataFrame with every metric applicable to the fitted model. |
+
+These are **diagnostic tools**, not proof of a "correct" clustering — always
+sanity-check results against domain knowledge. If a metric can't be computed
+(e.g. only 1 distinct cluster), `evaluate()` shows `NaN` with a note instead
+of crashing; the individual metric methods raise a `ParameterError`.
+
+## 11. Feature scaling and missing-value handling
+
+- Distances (and therefore both K-Means and Agglomerative results) are scale
+  -sensitive. If your features are on different scales (e.g. income in
+  thousands vs. a 1-100 score), set `scale=True`.
+- K-Means cluster centers are available both in the transformed (scaled)
+  space (`cluster._model.cluster_centers_`) and inverse-transformed back to
+  the original feature scale (`cluster.cluster_centers()`, the default).
+- Missing values are never silently dropped by default — you must choose
+  `missing="drop"/"mean"/"median"` explicitly, or clean the data yourself.
+
+## 12. Visualization examples
+
+```python
+cluster.plot_clusters(x="Income", y="Score", show_centers=True)
+cluster.plot_cluster_sizes()
+cluster.plot_silhouette()
+cluster.plot_pca()   # for > 2 features; clearly labeled as a 2D projection
+```
+
+`plot_clusters` requires exactly two selected feature columns to display;
+for higher-dimensional data use `plot_pca()` instead, which is explicitly
+labeled as a projection, not the original feature space.
+
+## 13. Cluster labels and DataFrame output
+
+```python
+labels = cluster.get_labels()                       # pd.Series, original index
+clustered_df = cluster.get_clustered_data("Segment") # original df + label column
+summary = cluster.cluster_summary()                  # size, proportion, per-feature mean
+```
+
+Cluster IDs (0, 1, 2, ...) are arbitrary identifiers — they carry no
+inherent ranking or meaning.
+
+## 14. Common errors and solutions
+
+| Error | Likely cause | Fix |
+|---|---|---|
+| `InputValidationError: not numeric` | Selected a text/categorical column | Encode it first, or exclude it from `features` |
+| `InputValidationError: missing values` | NaNs present, `missing="raise"` (default) | Pass `missing="drop"/"mean"/"median"` |
+| `ParameterError: Ward linkage requires metric='euclidean'` | `linkage="ward"` with another metric | Use `metric="euclidean"` or a different linkage |
+| `NotFittedError` | Called an evaluation/plot/output method before `fit_kmeans`/`fit_agglomerative` | Fit a model first |
+| `ParameterError` on `predict()` | Called `predict()` on an Agglomerative model | Not supported — only K-Means supports `predict()` |
+
+## 15. Computational limitations
+
+- Dendrograms require O(n²) memory/time for the distance matrix and are
+  impractical much past a few thousand samples.
+- `AgglomerativeClustering` scales worse than `KMeans` for large n.
+- No GPU acceleration; this wraps scikit-learn/SciPy CPU implementations.
+
+## 16. API reference (summary)
+
+**Construction:** `Cluster(data, features=None, scale=False, missing="raise")`
+
+**K-Means:** `elbow_method`, `fit_kmeans`, `cluster_centers`, `inertia`, `predict`
+
+**Agglomerative:** `plot_dendrogram`, `get_linkage_matrix`, `fit_agglomerative`
+
+**Shared:** `plot_clusters`, `plot_cluster_sizes`, `plot_silhouette`, `plot_pca`
+
+**Evaluation:** `silhouette_score`, `silhouette_samples`, `davies_bouldin_score`,
+`calinski_harabasz_score`, `evaluate`
+
+**Output:** `get_labels`, `get_clustered_data`, `cluster_summary`, `elbow_results`
+
+## 17. Testing
+
+```bash
+pip install -e ".[dev]"
+pytest -q
+```
+
+Current suite: **59 tests, all passing**, covering validation, both
+algorithms, visualization, evaluation edge cases, and DataFrame-output
+guarantees (index preservation, immutability, reproducibility).
+
+## 18. Contribution guidelines
+
+1. Fork, branch from `main`.
+2. Add/update tests for any behavior change.
+3. Run `pytest -q` before opening a PR.
+4. Keep public methods documented with docstrings.
+
+## 19. Roadmap
+
+- `DBSCAN`, `OPTICS`, `GaussianMixture` support behind the same API.
+- Automatic comparison across multiple `k` / linkage configurations.
+- Model save/load (`joblib`).
+- Export results directly to CSV.
+- Optional connectivity-constrained agglomerative clustering helpers.
+
+## 20. Preparing a PyPI release
+
+```bash
+pip install build twine
+python -m build                 # creates dist/*.whl and dist/*.tar.gz
+twine check dist/*
+pip install dist/clustereasy-0.1.0-py3-none-any.whl  # test locally
+twine upload --repository testpypi dist/*             # TestPyPI first
+twine upload dist/*                                    # then real PyPI
+```
+
+Bump the `version` in `pyproject.toml` before each release and update
+`CHANGELOG.md`.
+
+## License
+
+MIT — see `LICENSE`.

clustereasy-0.1.0/pyproject.toml

@@ -0,0 +1,43 @@
+[build-system]
+requires = ["setuptools>=68", "wheel"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "clustereasy"
+version = "0.1.0"
+description = "A beginner-friendly, consistent API for K-Means and Agglomerative Clustering on pandas DataFrames."
+readme = "README.md"
+license = "MIT"
+requires-python = ">=3.9"
+authors = [{ name = "Shashikant" }]
+keywords = ["clustering", "kmeans", "hierarchical", "machine-learning", "pandas"]
+classifiers = [
+    "Development Status :: 3 - Alpha",
+    "Intended Audience :: Science/Research",
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3.9",
+    "Programming Language :: Python :: 3.10",
+    "Programming Language :: Python :: 3.11",
+    "Programming Language :: Python :: 3.12",
+    "Topic :: Scientific/Engineering",
+]
+dependencies = [
+    "pandas>=1.5",
+    "numpy>=1.23",
+    "scikit-learn>=1.2",
+    "scipy>=1.9",
+    "matplotlib>=3.6",
+]
+
+[project.optional-dependencies]
+dev = ["pytest>=7", "build", "twine"]
+
+[project.urls]
+Homepage = "https://github.com/shashikant-ds636/easycluster"
+Issues = "https://github.com/shashikant-ds636/easycluster/issues"
+
+[tool.setuptools.packages.find]
+where = ["src"]
+
+[tool.pytest.ini_options]
+testpaths = ["tests"]

clustereasy-0.1.0/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

clustereasy-0.1.0/src/clustereasy/__init__.py

@@ -0,0 +1,15 @@
+"""easycluster: a beginner-friendly, consistent API for K-Means and
+Agglomerative Clustering on pandas DataFrames."""
+
+from .cluster import Cluster
+from .exceptions import EasyClusterError, InputValidationError, NotFittedError, ParameterError
+
+__version__ = "0.1.0"
+
+__all__ = [
+    "Cluster",
+    "EasyClusterError",
+    "InputValidationError",
+    "NotFittedError",
+    "ParameterError",
+]