lessPython 0.1.0__py3-none-any.whl

lessPy/__init__.py

@@ -0,0 +1,60 @@
+# Python analog of NAMESPACE: names imported here are the public API.
+
+from .bc_plotly import bc_plotly
+from .bubble_plotly import bubble_plotly
+from .Chart import Chart
+from .dn_plotly import dn_plotly
+from .dot_plotly import dot_plotly
+from .freq_poly_plotly import freq_poly_plotly
+from .hier_plotly import hier_plotly
+from .hs_plotly import hs_plotly
+from .Logit import Logit
+from .pie_plotly import pie_plotly
+from .plt_plotly import plt_plotly
+from .radar_plotly import radar_plotly
+from .Regression import Regression
+from .utils import get_option, set_option
+from .vbs_plotly import vbs_plotly
+from .X import X
+from .ANOVA import ANOVA
+from .ttest import ttest
+from .datasets import datasets, read_data
+from .reshape import reshape_long, reshape_wide
+from .pivot import pivot
+from .corEFA import corEFA
+from .corCFA import corCFA
+from .corScree import corScree
+from .corReorder import corReorder
+from .corProp import corProp
+from .Correlation import Correlation
+from .corReflect import corReflect
+from .corRead import corRead
+from .Prop_test import Prop_test
+from .corPrint import corPrint
+from .Flows import Flows
+from .date_infer import date_infer, format_date_labels
+from .rename import rename
+from .showColors import showColors
+from .getColors import getColors
+from .simCLT import simCLT
+from .simMeans import simMeans
+from .simFlips import simFlips
+from .simCImean import simCImean
+from .order_by import order_by
+from .prob_norm import prob_norm
+from .prob_znorm import prob_znorm
+from .prob_tcut import prob_tcut
+from .details import details
+from .VariableLabels import VariableLabels
+from .XY import XY
+
+__version__ = "0.1.0"
+
+__all__ = ["Chart", "X", "XY", "ANOVA", "ttest", "Regression", "Logit",
+           "read_data", "datasets", "reshape_long", "reshape_wide", "pivot", "corEFA", "corCFA", "corScree", "corReorder", "corProp", "Correlation", "corReflect", "corRead", "Prop_test", "corPrint", "Flows", "date_infer", "format_date_labels", "rename", "showColors", "getColors", "simCLT", "simMeans", "simFlips", "simCImean", "order_by", "prob_norm", "prob_znorm", "prob_tcut", "details",
+           "VariableLabels",
+           "bc_plotly", "bubble_plotly", "dn_plotly", "dot_plotly",
+           "freq_poly_plotly", "hier_plotly", "hs_plotly",
+           "pie_plotly", "plt_plotly", "radar_plotly",
+           "vbs_plotly",
+           "get_option", "set_option"]

lessPy/anova_rmd.py

@@ -0,0 +1,227 @@
+# anova_rmd.py — Quarto (.qmd) report for ANOVA(), the analog of
+# R's Rmd= for ANOVA (av.Rmd.R). Unlike the Regression report,
+# av.Rmd.R uses no external prose templates and no toggles
+# (explain/results are hardcoded on), so this generator is
+# correspondingly simple. Same design choices as the other lessPy
+# reports: Quarto target, direct per-section emitters, live code
+# chunks for tables and plots (recomputed from the result object
+# / data), narrative from R's prose. Deviation from av.Rmd.R: the
+# cell-means plot is included (R's ANOVA report has no graphics).
+
+from pathlib import Path
+
+from .reg_rmd import _chunk, _maybe_render, _xAnd
+from .utils import fmt
+
+
+def anova_rmd(result, y_name, facs, formula_str, design,
+              data_cols, rmd, rmd_data, rmd_format, rmd_browser):
+    """Write a Quarto report reproducing the ANOVA. Returns the
+    path of the .qmd file written."""
+    Y = y_name
+    X = _xAnd(facs)
+    d = result.digits_d
+    data_name = "d"
+    read_expr = (f'pd.read_csv("{rmd_data}")' if rmd_data
+                 else 'pd.read_csv("your_data.csv")')
+    call = f'ANOVA("{formula_str}", data={data_name})'
+
+    tx = _front_matter(Y)
+    tx += [
+        "The purpose of this analysis is the analysis of "
+        f"variance of the values of {Y} for the different "
+        f"levels of {X}.", ""]
+
+    # data
+    tx += ["## The Data", "",
+           "Read the data into a pandas data frame.", ""]
+    tx += _chunk(["import pandas as pd",
+                  f"{data_name} = {read_expr}", data_name],
+                 echo=True)
+    tx += ["Data from the following variables are available for "
+           f"analysis: {_xAnd(data_cols)}.", ""]
+
+    # analysis
+    tx += ["## Analysis of Variance", "",
+           "Obtain the analysis with the _lessPy_ function "
+           "_ANOVA()_.", "", _design_sentence(design, Y, facs),
+           ""]
+    tx += _chunk(["import pandas as pd",
+                  "from lessPy import ANOVA",
+                  f"r = {call}"], echo=True, output=False)
+
+    # background
+    tx += ["### Background", "",
+           "The output begins with a specification of the "
+           "variables in the model and a brief description of "
+           "the data.", "",
+           f"The response variable is {Y}. The "
+           f"{'factor' if len(facs) == 1 else 'factors'} "
+           f"{X} define the groups. Of {result.n_obs} cases, "
+           f"{result.n_keep} are retained for analysis.", ""]
+
+    tx += _sec_descriptive(result, Y, facs, design, d)
+    tx += _sec_summary(Y, X)
+    tx += _sec_effects(result, facs, design, d)
+    tx += _sec_pairwise(result, X, design, facs)
+    tx += _sec_residuals()
+
+    path = Path(rmd)
+    if path.suffix.lower() != ".qmd":
+        path = path.with_suffix(".qmd")
+    path.write_text("\n".join(tx))
+    print(f"\nQuarto report written:  {path}")
+    _maybe_render(path, rmd_format, rmd_browser)
+    return path
+
+
+def _front_matter(Y):
+    return [
+        "---",
+        f'title: "ANOVA of {Y}"',
+        "format:",
+        "  html:",
+        "    toc: true",
+        "    toc-depth: 4",
+        "    embed-resources: true",
+        "engine: jupyter",
+        "---",
+        "",
+        "------",
+        ""]
+
+
+def _design_sentence(design, Y, facs):
+    if design == "oneway":
+        return (f"This is a one-way ANOVA of {Y} with treatment "
+                f"factor {facs[0]}.")
+    if design == "two-between":
+        return (f"This is a two-way between groups ANOVA of {Y} "
+                f"with treatment factors {_xAnd(facs)}.")
+    return (f"This is a randomized blocks ANOVA of {Y} with one "
+            f"treatment factor, {facs[0]}, and one blocking "
+            f"factor, {facs[1]}.")
+
+
+def _sec_descriptive(result, Y, facs, design, d):
+    out = ["### Descriptive Statistics", "",
+           "The descriptive statistics of "
+           f"{Y} for the different levels of {_xAnd(facs)} "
+           "begin the analysis.", ""]
+    if design == "oneway":
+        out += _chunk(["r.descriptive"], echo=False)
+        out += [f"Grand Mean: {fmt(result.grand_mean, d + 1)}",
+                ""]
+        out += _plot_chunk(result, "means")
+        return out
+
+    f1, f2 = facs
+    if design == "two-between":
+        out += ["First, the number of data values in each "
+                "cell.", ""]
+        out += _chunk(
+            [f'd.pivot_table(index="{f2}", columns="{f1}", '
+             f'values="{Y}", aggfunc="size")'], echo=False)
+        out += ["The cell means follow.", ""]
+        out += _chunk(
+            [f'd.pivot_table(index="{f2}", columns="{f1}", '
+             f'values="{Y}", aggfunc="mean")'], echo=False)
+    out += ["The marginal means of each factor assist in "
+            "interpreting any main effects.", ""]
+    out += _chunk(
+        [f'd.groupby("{f1}")["{Y}"].mean()'], echo=False)
+    out += _chunk(
+        [f'd.groupby("{f2}")["{Y}"].mean()'], echo=False)
+    out += [f"The grand mean of all the data is "
+            f"{fmt(result.grand_mean, d + 1)}.", ""]
+    if design == "two-between":
+        out += ["The variation in each cell, its standard "
+                "deviation, follows.", ""]
+        out += _chunk(
+            [f'd.pivot_table(index="{f2}", columns="{f1}", '
+             f'values="{Y}", aggfunc="std")'], echo=False)
+        out += _plot_chunk(result, "interaction")
+    else:
+        out += _plot_chunk(result, "data")
+        out += _plot_chunk(result, "fitted")
+    return out
+
+
+def _sec_summary(Y, X):
+    return ["### Summary Table", "",
+            "The analysis of variance (ANOVA) partitions the "
+            f"total sum of squares for {Y} into the residual "
+            "variability, $\\sum e^2_i$, and the sum of squares "
+            f"for {X}. The ANOVA table displays these sources "
+            "of variation.", "",
+            *_chunk(["r.anova"], echo=False)]
+
+
+def _sec_effects(result, facs, design, d):
+    out = ["### Effects", "",
+           "The ANOVA table gives the significance test for the "
+           "factor; also of interest is the size of the effect.",
+           ""]
+    e = result.effects
+    if design == "oneway":
+        out += [
+            f"R Squared: {fmt(e['R_squared'], 3)}  \n"
+            f"R Sq Adjusted: {fmt(e['R_sq_adjusted'], 3)}  \n"
+            f"Omega Squared: {fmt(e['omega_squared'], 3)}  \n"
+            f"Cohen's f: {fmt(e['cohen_f'], 3)}", ""]
+    elif design == "two-between":
+        f1, f2 = facs
+        out += [
+            f"Partial Omega Squared for {f1}: "
+            f"{fmt(e['omega_sq_' + f1], 3)}  \n"
+            f"Partial Omega Squared for {f2}: "
+            f"{fmt(e['omega_sq_' + f2], 3)}  \n"
+            f"Partial Omega Squared for {f1} & {f2}: "
+            f"{fmt(e['omega_sq_interaction'], 3)}", ""]
+    else:
+        f1, f2 = facs
+        out += [
+            f"Partial Omega Squared for {f1}: "
+            f"{fmt(e['omega_sq_' + f1], 3)}  \n"
+            f"Partial Intraclass Correlation for {f2}: "
+            f"{fmt(e['intraclass_' + f2], 3)}", ""]
+    return out
+
+
+def _sec_pairwise(result, X, design, facs):
+    out = ["### Pairwise Differences", "",
+           f"A significant effect shows that {X} impacts the "
+           "response, but do all levels differ or just some? "
+           "The Tukey pairwise comparisons, adjusted for the "
+           "overall significance level, are the search for "
+           "Honestly Significant Differences (HSD).", ""]
+    if result.tukey is None:
+        return out
+    if design == "oneway":
+        out += _chunk(["r.tukey"], echo=False)
+    else:
+        f1 = facs[0]
+        out += [f"Factor {f1}:", ""]
+        out += _chunk([f'r.tukey["{f1}"]'], echo=False)
+        if design == "two-between":
+            f2 = facs[1]
+            out += [f"Factor {f2}:", ""]
+            out += _chunk([f'r.tukey["{f2}"]'], echo=False)
+            out += ["Cell means:", ""]
+            out += _chunk(['r.tukey["cells"]'], echo=False)
+    return out
+
+
+def _sec_residuals():
+    return ["### Residuals", "",
+            "ANOVA is a form of regression: each data value has "
+            "a fitted value, its cell mean, and the difference "
+            "is the residual. The cases with the largest "
+            "standardized residuals are listed first.", "",
+            *_chunk(["r.residuals"], echo=False)]
+
+
+def _plot_chunk(result, key):
+    if key not in result.plots:
+        return []
+    return _chunk([f'r.plots["{key}"]'], echo=False)