distfitlab 0.2.0__tar.gz

distfitlab-0.2.0/.gitignore

@@ -0,0 +1,40 @@
+# Virtual environments
+venv/
+.venv/
+env/
+
+# Python bytecode
+__pycache__/
+*.py[cod]
+
+# Jupyter
+.ipynb_checkpoints/
+*/.ipynb_checkpoints/
+
+# IDE
+.idea/
+.vscode/
+
+# macOS
+.DS_Store
+
+# R
+.Rhistory
+
+# Generated run outputs (timestamped fit results are reproducible; keep them local)
+outputs/
+
+# Packaging
+build/
+dist/
+*.egg-info/
+
+# Test cache
+.pytest_cache/
+
+# Kept locally, not part of the published package
+archive/
+artifacts/
+INVENTORY.md
+requirements.txt
+.Rhistory

distfitlab-0.2.0/CHANGELOG.md

@@ -0,0 +1,21 @@
+# Changelog
+
+## 0.2.0 — 2026-09-13
+
+- First release on PyPI: `pip install distfitlab`.
+- **Data view**: new **Sorted run chart** — values in ascending order against rank, with
+  mean and median lines and the same pan/zoom toolbar as the run chart.
+- **grid lines** option for both run charts, applied live to the chart on screen.
+- Run charts are narrower (7 in, was 9 in), since they can be zoomed.
+- README: "After installing" walkthrough and troubleshooting; links now work on the PyPI page.
+
+## 0.1.0 — first public release
+
+- Two fitting apps behind one switch: 18 continuous distributions and 5 count distributions.
+- **Fit All** ranks every distribution by max |empirical CDF − theoretical CDF|, best first.
+- **Fit** scores a single distribution against parameters you type, with CDF/PDF overlays.
+- **Data view**: histogram / counts bar chart, ECDF, box plot, Q-Q plot, run chart and
+  summary statistics, optionally split by group. The run chart has a pan/zoom toolbar.
+- File data: multi-format reader (CSV, Excel, JSON, Parquet, Feather, SPSS/Stata/SAS, HDF5),
+  a stack of filter conditions, cleaning options, and group-by that fits every group.
+- **Save results** writes the table, the samples and the figures to a timestamped folder.

distfitlab-0.2.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Reza Berangi
+
+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.

distfitlab-0.2.0/PKG-INFO

@@ -0,0 +1,161 @@
+Metadata-Version: 2.5
+Name: distfitlab
+Version: 0.2.0
+Summary: Interactive distribution fitting in Jupyter: fit, rank and inspect continuous and count distributions
+Project-URL: Homepage, https://github.com/rberangi/distfitlab
+Project-URL: Source, https://github.com/rberangi/distfitlab
+Project-URL: Issues, https://github.com/rberangi/distfitlab/issues
+Author-email: Reza Berangi <rberangi@gmail.com>
+License-Expression: MIT
+License-File: LICENSE
+Keywords: distribution fitting,goodness of fit,ipywidgets,jupyter,poisson,scipy,statistics,weibull
+Classifier: Development Status :: 4 - Beta
+Classifier: Framework :: Jupyter
+Classifier: Intended Audience :: Science/Research
+Classifier: Programming Language :: Python :: 3
+Classifier: Topic :: Scientific/Engineering :: Mathematics
+Requires-Python: >=3.9
+Requires-Dist: ipywidgets>=8.0
+Requires-Dist: matplotlib>=3.9
+Requires-Dist: numpy>=1.23
+Requires-Dist: pandas>=1.5
+Requires-Dist: scipy>=1.9
+Provides-Extra: all
+Requires-Dist: ipympl>=0.9; extra == 'all'
+Requires-Dist: openpyxl; extra == 'all'
+Requires-Dist: pyarrow; extra == 'all'
+Requires-Dist: pyreadstat; extra == 'all'
+Requires-Dist: tables; extra == 'all'
+Provides-Extra: dev
+Requires-Dist: jupyter; extra == 'dev'
+Requires-Dist: nbconvert; extra == 'dev'
+Requires-Dist: nbformat; extra == 'dev'
+Requires-Dist: pytest>=7; extra == 'dev'
+Provides-Extra: excel
+Requires-Dist: openpyxl; extra == 'excel'
+Provides-Extra: hdf5
+Requires-Dist: tables; extra == 'hdf5'
+Provides-Extra: parquet
+Requires-Dist: pyarrow; extra == 'parquet'
+Provides-Extra: spss
+Requires-Dist: pyreadstat; extra == 'spss'
+Provides-Extra: zoom
+Requires-Dist: ipympl>=0.9; extra == 'zoom'
+Description-Content-Type: text/markdown
+
+# distfitlab
+
+Interactive distribution fitting in Jupyter. Simulate or load data, fit many
+probability distributions at once, see which one fits best, and inspect the data —
+from a single widget UI.
+
+Two apps behind one switch:
+
+- **Continuous** — 18 distributions: Normal, Exponential, Gamma, Rayleigh, Weibull,
+  Lognormal, Loglogistic, Inverse Gaussian, Beta, GEV, Logistic, Laplace, Chi-squared,
+  Chi, Nakagami, Rician, Cauchy, Student-T. Fitted with `scipy.stats`.
+- **Discrete** — 5 count distributions: Poisson, Binomial, Geometric, Negative Binomial,
+  Zero-Inflated Poisson. SciPy has no `.fit` for these, so the fitters are written here.
+
+## Install
+
+```bash
+pip install distfitlab
+```
+
+Optional readers and the interactive run charts are extras:
+
+```bash
+pip install "distfitlab[all]"
+```
+
+For the latest development version, install from GitHub instead:
+`pip install "distfitlab[all] @ git+https://github.com/rberangi/distfitlab.git"`.
+
+| Extra | Adds |
+|---|---|
+| `excel` | `.xlsx`, `.xls` |
+| `parquet` | `.parquet`, `.feather`, `.arrow` |
+| `spss` | `.sav`, `.dta`, `.sas7bdat`, `.xpt` |
+| `hdf5` | `.h5`, `.hdf5` |
+| `zoom` | pan/zoom toolbar on the run charts |
+| `all` | all of the above |
+
+## After installing
+
+distfitlab is a widget app, so it runs inside a notebook — not in a plain Python script
+or terminal.
+
+**1. Open a notebook** in the same Python environment you installed into:
+
+```bash
+pip install jupyterlab      # skip if you already have Jupyter
+jupyter lab
+```
+
+VS Code works too: create a `.ipynb` file and select that Python as the kernel.
+Notebook 7 is also fine — anywhere `ipywidgets` renders.
+
+**2. Start the app** in a notebook cell:
+
+```python
+from distfitlab import main
+
+main()                # opens on the continuous app
+# main("Discrete")    # opens on the discrete app
+```
+
+**3. Use the UI** that appears below the cell:
+
+1. **Data source** — simulate data, or choose **From file** to upload a file (or enter a
+   path) and pick the column.
+2. **Fit All** — fit every distribution and get a table ranked best-first.
+3. **Fit** — try one distribution with parameters you type.
+4. **Visualize data** — histogram, ECDF, box plot, Q-Q plot, run charts and summary stats.
+5. **Save results** — write the table and figures to an `outputs/` folder next to the
+   notebook.
+
+A ready notebook is in [`examples/quickstart.ipynb`](https://github.com/rberangi/distfitlab/blob/main/examples/quickstart.ipynb)
+(in the repository; a pip install does not include it), and every control is explained in
+[`docs/guide.md`](https://github.com/rberangi/distfitlab/blob/main/docs/guide.md).
+
+### If something doesn't work
+
+| Symptom | Fix |
+|---|---|
+| `ModuleNotFoundError: No module named 'distfitlab'` | The notebook kernel is a different Python from the one pip installed into. Run `%pip install "distfitlab[all]"` in a cell, then restart the kernel. |
+| Text such as `VBox(children=...)` instead of the UI | The frontend can't render widgets. Use JupyterLab 3+, Notebook 7, or VS Code with the Jupyter extension. |
+| No pan/zoom toolbar on the run charts | `ipympl` is missing. Install the `zoom` or `all` extra, then restart the kernel. |
+
+## What you get
+
+- **Fit All** fits every distribution and ranks them best-first by the largest gap
+  between the empirical and theoretical CDF, with the parameters for each.
+- **Fit** scores one distribution against parameters you type, with CDF/PDF overlays.
+- **Data view** — histogram (or counts bar chart), ECDF, box plot, Q-Q plot against the
+  distribution you chose, run chart and sorted run chart with a pan/zoom toolbar, and
+  summary statistics.
+- **File data** — CSV, TSV, Excel, JSON, Parquet, Feather, SPSS/Stata/SAS and HDF5; a
+  stack of filter conditions across columns; cleaning (drop NaN, drop ≤ 0, percentile
+  trim, drop duplicates); and **Group by**, which fits every group and ranks them.
+- **loc = 0** option, for when a distribution should be anchored at the origin.
+- **Save results** writes the table, the samples and the figures to a timestamped folder.
+
+## Requirements
+
+Python 3.9+, with `numpy`, `pandas`, `scipy`, `matplotlib` and `ipywidgets`. Developed
+and tested on Python 3.12.
+
+## Documentation
+
+The full walkthrough — every control, the parameter conventions, supported file types,
+troubleshooting and how the fitting works — is in [`docs/guide.md`](https://github.com/rberangi/distfitlab/blob/main/docs/guide.md).
+
+## License
+
+MIT — see [LICENSE](https://github.com/rberangi/distfitlab/blob/main/LICENSE).
+
+## Citation
+
+If this helps with published work, please cite the libraries it stands on: SciPy, NumPy,
+pandas, Matplotlib and ipywidgets.

distfitlab-0.2.0/README.md

@@ -0,0 +1,116 @@
+# distfitlab
+
+Interactive distribution fitting in Jupyter. Simulate or load data, fit many
+probability distributions at once, see which one fits best, and inspect the data —
+from a single widget UI.
+
+Two apps behind one switch:
+
+- **Continuous** — 18 distributions: Normal, Exponential, Gamma, Rayleigh, Weibull,
+  Lognormal, Loglogistic, Inverse Gaussian, Beta, GEV, Logistic, Laplace, Chi-squared,
+  Chi, Nakagami, Rician, Cauchy, Student-T. Fitted with `scipy.stats`.
+- **Discrete** — 5 count distributions: Poisson, Binomial, Geometric, Negative Binomial,
+  Zero-Inflated Poisson. SciPy has no `.fit` for these, so the fitters are written here.
+
+## Install
+
+```bash
+pip install distfitlab
+```
+
+Optional readers and the interactive run charts are extras:
+
+```bash
+pip install "distfitlab[all]"
+```
+
+For the latest development version, install from GitHub instead:
+`pip install "distfitlab[all] @ git+https://github.com/rberangi/distfitlab.git"`.
+
+| Extra | Adds |
+|---|---|
+| `excel` | `.xlsx`, `.xls` |
+| `parquet` | `.parquet`, `.feather`, `.arrow` |
+| `spss` | `.sav`, `.dta`, `.sas7bdat`, `.xpt` |
+| `hdf5` | `.h5`, `.hdf5` |
+| `zoom` | pan/zoom toolbar on the run charts |
+| `all` | all of the above |
+
+## After installing
+
+distfitlab is a widget app, so it runs inside a notebook — not in a plain Python script
+or terminal.
+
+**1. Open a notebook** in the same Python environment you installed into:
+
+```bash
+pip install jupyterlab      # skip if you already have Jupyter
+jupyter lab
+```
+
+VS Code works too: create a `.ipynb` file and select that Python as the kernel.
+Notebook 7 is also fine — anywhere `ipywidgets` renders.
+
+**2. Start the app** in a notebook cell:
+
+```python
+from distfitlab import main
+
+main()                # opens on the continuous app
+# main("Discrete")    # opens on the discrete app
+```
+
+**3. Use the UI** that appears below the cell:
+
+1. **Data source** — simulate data, or choose **From file** to upload a file (or enter a
+   path) and pick the column.
+2. **Fit All** — fit every distribution and get a table ranked best-first.
+3. **Fit** — try one distribution with parameters you type.
+4. **Visualize data** — histogram, ECDF, box plot, Q-Q plot, run charts and summary stats.
+5. **Save results** — write the table and figures to an `outputs/` folder next to the
+   notebook.
+
+A ready notebook is in [`examples/quickstart.ipynb`](https://github.com/rberangi/distfitlab/blob/main/examples/quickstart.ipynb)
+(in the repository; a pip install does not include it), and every control is explained in
+[`docs/guide.md`](https://github.com/rberangi/distfitlab/blob/main/docs/guide.md).
+
+### If something doesn't work
+
+| Symptom | Fix |
+|---|---|
+| `ModuleNotFoundError: No module named 'distfitlab'` | The notebook kernel is a different Python from the one pip installed into. Run `%pip install "distfitlab[all]"` in a cell, then restart the kernel. |
+| Text such as `VBox(children=...)` instead of the UI | The frontend can't render widgets. Use JupyterLab 3+, Notebook 7, or VS Code with the Jupyter extension. |
+| No pan/zoom toolbar on the run charts | `ipympl` is missing. Install the `zoom` or `all` extra, then restart the kernel. |
+
+## What you get
+
+- **Fit All** fits every distribution and ranks them best-first by the largest gap
+  between the empirical and theoretical CDF, with the parameters for each.
+- **Fit** scores one distribution against parameters you type, with CDF/PDF overlays.
+- **Data view** — histogram (or counts bar chart), ECDF, box plot, Q-Q plot against the
+  distribution you chose, run chart and sorted run chart with a pan/zoom toolbar, and
+  summary statistics.
+- **File data** — CSV, TSV, Excel, JSON, Parquet, Feather, SPSS/Stata/SAS and HDF5; a
+  stack of filter conditions across columns; cleaning (drop NaN, drop ≤ 0, percentile
+  trim, drop duplicates); and **Group by**, which fits every group and ranks them.
+- **loc = 0** option, for when a distribution should be anchored at the origin.
+- **Save results** writes the table, the samples and the figures to a timestamped folder.
+
+## Requirements
+
+Python 3.9+, with `numpy`, `pandas`, `scipy`, `matplotlib` and `ipywidgets`. Developed
+and tested on Python 3.12.
+
+## Documentation
+
+The full walkthrough — every control, the parameter conventions, supported file types,
+troubleshooting and how the fitting works — is in [`docs/guide.md`](https://github.com/rberangi/distfitlab/blob/main/docs/guide.md).
+
+## License
+
+MIT — see [LICENSE](https://github.com/rberangi/distfitlab/blob/main/LICENSE).
+
+## Citation
+
+If this helps with published work, please cite the libraries it stands on: SciPy, NumPy,
+pandas, Matplotlib and ipywidgets.

distfitlab-0.2.0/docs/guide.md

@@ -0,0 +1,147 @@
+# distfitlab — user guide
+
+This is the full walkthrough. For installation and a one-minute start, see the
+[README](../README.md).
+
+## Using the UI
+0) Pick Continuous or Discrete with the switch at the top.
+
+1) Data source  (on-screen block: DATA SOURCE)
+
+Pick **Data Source**, which has two options:
+
+Simulated Random Data: choose a **Distribution** and N; a single compact row shows just its parameters.
+UI convention: loc is always first.
+
+From file: upload a file or enter a local path, then choose a Column. Use the panel below it to stack filters, clean the column and group rows (all optional). The Distribution list, N and Seed are hidden in this mode, since they only apply to simulated data. CDF bins lives in the **Visualize data** block and stays visible in both modes, because it sets the empirical CDF that every fit is scored against, file data included.
+
+Reproducibility: for synthetic data, set Seed; for external data, “Fit” and “Fit All” share a cached empirical baseline so results match.
+
+2) Fit all distributions  (on-screen block: FIT ALL DISTRIBUTIONS)
+
+Continuous only: choose **loc free** (scipy fits loc) or **loc = 0** (loc held at zero) with the radio buttons next to Fit All. With loc = 0, distributions that need positive data cannot be fitted when the data has values ≤ 0; they are skipped and listed in the status line, and the table shows loc = 0 (fixed).
+
+Click Fit All to:
+
+Fit every supported distribution to the current data.
+
+See a parameter table (with Parameter 1 = loc where applicable) and Fitting Error (max CDF deviation).
+
+View CDF/PDF overlays for the best fit.
+
+Inspect a bar chart of errors (labels printed vertically).
+
+3) Fit one distribution  (on-screen block: FIT ONE DISTRIBUTION)
+
+Pick a distribution, enter parameters, and click Fit to:
+
+Compute the max CDF error against the same empirical baseline.
+
+Display CDF and PDF overlays for those parameters.
+
+4) Visualize data  (on-screen block: VISUALIZE DATA, results in the Data view tab)
+
+The **Visualize** buttons draw the data you are about to fit - after filters, cleaning and group selection - into the **Data view** tab. **Q-Q plot** uses the distribution and parameters currently in the Fit row, so it answers "do my parameters match this data?" before you fit. **Run chart** plots values in row order, which shows drift or steps that a histogram hides. **Sorted run chart** plots the same values in ascending order against their rank, with mean and median lines, so the range, gaps, ties and outliers stand out; like the run chart it has a pan/zoom toolbar. Tick **grid lines** to add a light grid to both run charts; it applies to a chart already on screen, keeping the current zoom. With a **Group by** column set, tick **split by group** to draw one series per group (up to 10).
+
+5) Save results
+
+Click Save results to write, into a timestamped folder under outputs/:
+
+fitted_parameters.csv, samples.csv
+
+fit_errors.png, cdf_overlay.png, pdf_overlay.png
+
+Links appear inline in the notebook. Group mode produces no CDF/PDF overlay, so only the files that exist are written.
+
+Parameter conventions
+
+UI and synthetic generation always show loc first.
+
+Fitted parameter table makes Parameter 1 the loc value (when defined).
+
+Special notes:
+
+Exponential
+
+Simulated input uses scale (Distribution → Exponential has “scale”).
+
+Fit expects λ (rate); it’s clearly labeled as “λ (rate)”.
+
+Internally we convert rate ↔ scale as needed.
+
+Beta expects (loc, α, β, scale); support for general location/scale form.
+
+Weibull UI uses (loc, k(shape), λ(scale)).
+
+Lognormal UI uses (loc, s(shape), scale).
+
+GEV UI uses (loc, ξ(shape), scale).
+
+All other distributions follow the labels shown in the UI and table.
+
+Supported file types (and add-ons)
+Type	Ext	Notes / Add-on
+CSV/TSV/PSV/TXT	.csv, .tsv, .psv, .txt	TXT auto-sniffs delimiter; falls back to whitespace
+Excel	.xlsx, .xls	openpyxl for .xlsx
+JSON	.json	Standard JSON
+JSON Lines	.jsonl, .ndjson	lines=True
+Parquet	.parquet	pyarrow recommended
+Feather / Arrow IPC	.feather, .arrow	pyarrow required
+SPSS / Stata / SAS	.sav, .dta, .sas7bdat, .xpt	pyreadstat
+HDF5	.h5, .hdf5	pandas.read_hdf
+MATLAB	.mat	v5 via scipy.io; v7.3 via h5py
+
+Column selection: the column dropdown prioritizes numeric columns; non-numeric are kept for filtering but are not usable for fitting until coerced to numeric.
+
+Troubleshooting
+
+“Load error: …requires pyarrow/pyreadstat/openpyxl/h5py”
+Install the indicated package (see Quick start).
+
+No numeric data in selected column
+Choose a different column or clean/filter data first. The app coerces to numeric and drops non-numeric rows for modeling.
+
+Fit vs Fit All mismatch
+
+Ensure you didn’t change bins, filter, or seed between actions.
+
+The app caches the empirical baseline so both use the same CDF/PDF reference.
+
+Widget UI not rendering
+Ensure ipywidgets is installed and enabled; reload the notebook kernel.
+
+How it works (internals snapshot)
+
+pdf_cdf(x, bins, PDF=True) builds a histogram-based empirical CDF (stable in the tails) and derives an empirical PDF by differencing.
+
+datasource(...) generates synthetic arrays using SciPy/NumPy or returns a numeric column from the uploaded DataFrame.
+
+theory_cdf/pdf(process, params, x) call SciPy CDF/PDF for each distribution (parameter order mapped from the UI).
+
+fit_subset_and_summarize(r, bins, subset):
+
+Fits each distribution with scipy.stats.<dist>.fit.
+
+Computes max absolute deviation between empirical CDF and theoretical CDF.
+
+Builds a table (with loc in Parameter 1 where applicable).
+
+A small cache ensures the empirical baseline used in Fit All is reused by Fit.
+
+Notes & tips
+
+CDF bins (in the **Visualize data** block): affects empirical CDF/PDF smoothness and the histogram detail; 50–200 is a sensible range. It also sets the empirical curve that Fit and Fit All are scored against, for file data as well as simulated data.
+
+Scale/shape positivity: the UI enforces positive values for these parameters.
+
+Large catalogs: the error bar chart turns labels vertical to stay readable.
+
+## Where files are written
+
+`Save results` and the formula sheets write under your **current working directory**:
+
+- `outputs/fit_outputs_<timestamp>/` — continuous fits
+- `outputs/discrete_fit_outputs_<timestamp>/` — discrete fits
+- `artifacts/` — the rendered formula sheets
+
+Run the notebook from the folder you want those in.

distfitlab-0.2.0/examples/quickstart.ipynb

@@ -0,0 +1,67 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "0db36152",
+   "metadata": {},
+   "source": [
+    "# distfitlab\n",
+    "\n",
+    "Fit probability distributions to your data and compare them, from one widget UI.\n",
+    "\n",
+    "- **Continuous** — 18 distributions (Normal … Student-T), fitted with `scipy.stats`\n",
+    "- **Discrete** — 5 count distributions (Poisson, Binomial, Geometric, Negative Binomial, Zero-Inflated Poisson)\n",
+    "\n",
+    "Use the switch at the top to choose the family. **Fit All** fits every distribution and\n",
+    "ranks them; **Fit** checks parameters you enter; the **Data view** tab plots the data itself.\n",
+    "\n",
+    "Files you save land under your current working directory (`outputs/`, `artifacts/`)."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "1e496f87",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "try:\n",
+    "    from distfitlab import main\n",
+    "except ModuleNotFoundError:        # opened from a clone without installing: use the sources\n",
+    "    import sys; sys.path.insert(0, \"../src\")\n",
+    "    from distfitlab import main\n",
+    "\n",
+    "main()                     # Run chart is interactive (pan/zoom); other figures are images"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "7bcf3489-dd26-41ef-9c5f-f80e4d091468",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.12.7"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

distfitlab-0.2.0/pyproject.toml

@@ -0,0 +1,52 @@
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[project]
+name = "distfitlab"
+dynamic = ["version"]
+description = "Interactive distribution fitting in Jupyter: fit, rank and inspect continuous and count distributions"
+readme = "README.md"
+requires-python = ">=3.9"
+license = "MIT"
+license-files = ["LICENSE"]
+authors = [{ name = "Reza Berangi", email = "rberangi@gmail.com" }]
+keywords = ["statistics", "distribution fitting", "jupyter", "ipywidgets",
+            "scipy", "goodness of fit", "poisson", "weibull"]
+classifiers = [
+    "Development Status :: 4 - Beta",
+    "Framework :: Jupyter",
+    "Intended Audience :: Science/Research",
+    "Programming Language :: Python :: 3",
+    "Topic :: Scientific/Engineering :: Mathematics",
+]
+dependencies = [
+    "numpy>=1.23",
+    "pandas>=1.5",
+    "scipy>=1.9",
+    "matplotlib>=3.9",      # boxplot(tick_labels=...)
+    "ipywidgets>=8.0",      # RadioButtons(orientation=...), Tab titles
+]
+
+[project.optional-dependencies]
+excel   = ["openpyxl"]              # .xlsx / .xls
+parquet = ["pyarrow"]               # .parquet, .feather, .arrow
+spss    = ["pyreadstat"]            # .sav, .dta, .sas7bdat, .xpt
+hdf5    = ["tables"]                # .h5, .hdf5
+zoom    = ["ipympl>=0.9"]           # pan/zoom toolbar on the run chart
+all     = ["openpyxl", "pyarrow", "pyreadstat", "tables", "ipympl>=0.9"]
+dev     = ["pytest>=7", "nbformat", "nbconvert", "jupyter"]
+
+[project.urls]
+Homepage = "https://github.com/rberangi/distfitlab"
+Source   = "https://github.com/rberangi/distfitlab"
+Issues   = "https://github.com/rberangi/distfitlab/issues"
+
+[tool.hatch.version]
+path = "src/distfitlab/__init__.py"
+
+[tool.hatch.build.targets.wheel]
+packages = ["src/distfitlab"]
+
+[tool.hatch.build.targets.sdist]
+include = ["src/distfitlab", "examples", "docs", "tests", "README.md", "LICENSE", "CHANGELOG.md"]

distfitlab-0.2.0/src/distfitlab/__init__.py

@@ -0,0 +1,29 @@
+"""distfitlab - interactive distribution fitting in Jupyter.
+
+Fit probability distributions to your data, compare them by the largest gap between
+the empirical and theoretical CDF, and inspect the data itself - all from one widget
+UI with two halves:
+
+    continuous  18 distributions (Normal, Gamma, Weibull, GEV, Student-T, ...)
+    discrete     5 count distributions (Poisson, Binomial, Geometric,
+                 Negative Binomial, Zero-Inflated Poisson)
+
+Usage:
+
+    from distfitlab import main
+    main()                 # opens on the continuous app
+    main("Discrete")       # opens on the discrete app
+
+Building the widgets takes a couple of seconds, so it is deferred until `main` is
+actually looked up rather than done on `import distfitlab`.
+"""
+
+__version__ = "0.2.0"
+__all__ = ["main"]
+
+
+def __getattr__(name):          # PEP 562: keeps `import distfitlab` cheap
+    if name == "main":
+        from .app import main
+        return main
+    raise AttributeError(f"module {__name__!r} has no attribute {name!r}")