elessar 1.0.0__tar.gz

elessar-1.0.0/.gitignore

@@ -0,0 +1,14 @@
+__pycache__/
+*.py[cod]
+*.egg-info/
+.venv/
+venv/
+build/
+dist/
+.pytest_cache/
+.coverage
+htmlcov/
+elessar_report/
+elessar_report.*
+.DS_Store
+fuzz/

elessar-1.0.0/CHANGELOG.md

@@ -0,0 +1,36 @@
+# Changelog
+
+## 1.0.0 (2026-10-30)
+
+First public release on PyPI.
+
+- **Any input, any size**: pandas, polars, pyarrow, dask, NumPy objects and files (CSV / TSV, JSON, Parquet,
+  Feather, ORC, Excel, ODS, SQLite, SAS, Stata, SPSS, HDF5, pickle, XML, compressed or not, globs, folders, URLs).
+  Tables bigger than memory are typed progressively on as few rows as the evidence allows, then read once in
+  parallel chunks (dask for text files) with exact counts over every row; time series are never cut into
+  scattered rows.
+- **Semantic typing by an explicit decision tree**, with the path and the certainty (certain, probable,
+  uncertain) of every decision: identifiers, booleans in disguise, numbers and dates stored as text in any
+  format, emails, URLs, phones, postal codes, JSON, lists of labels, free text, ordinal scales, counts, and
+  integers that are really category codes (region 11, store 7, MCC 5411). Types can be set by hand
+  (`StoneConfig(column_types=...)`, `--type COL=TYPE`).
+- **Dates of birth** told apart from event dates and turned into ages at the right reference date, with their
+  own checks.
+- **Statistics, distributions and associations**: descriptive statistics with confidence intervals, MLE fits
+  ranked by AIC, normality tests, every pairwise association with tests, Benjamini-Hochberg correction and
+  confidence intervals, VIF, exact linear dependencies, missingness mechanisms.
+- **Target analysis and feature selection**: ranking, leakage detection, informative missingness, five
+  cross-validated selection methods with a vote.
+- **Multivariate statistics**: PCA verdict (Bartlett, KMO, parallel analysis), a factor analysis run step by step
+  (KMO and Bartlett, communalities, number of factors, varimax, cross-loadings, Cronbach's alpha, scale scores),
+  redundant groups, robust multivariate outliers.
+- **Outliers** by three methods, point by point, with keep / remove / reassign.
+- **Time series** recognised on their own (single series, panel, event log): trend with the Hamed-Rao
+  correction, seasonality by effect size, ADF and KPSS read together, decomposition, validation folds on the
+  real dates.
+- **Data quality alerts and a preparation plan**, every step with its reason and its code.
+- **Report**: a terminal report, an HTML page in a dark "data observatory" design (sidebar, key figures,
+  numbered sections, charts that open full size) and its Markdown twin, report-quality charts, in a dated folder
+  per table and target.
+- Tested on Linux, macOS and Windows, Python 3.10 to 3.13, on 55 public datasets and 17 reference datasets for
+  the factor analysis.

elessar-1.0.0/LICENSE

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