Origin/dev/refactor code

.github/workflows/draft-pdf.yml

@@ -1,5 +1,18 @@
 name: Draft PDF
-on: [push]
+
+on:
+  push:
+    paths:
+      - paper.md
+      - paper.bib
+      - figures/**
+      - .github/workflows/draft-pdf.yml
+  pull_request:
+    paths:
+      - paper.md
+      - paper.bib
+      - figures/**
+      - .github/workflows/draft-pdf.yml
 
 permissions:
   contents: read

.gitignore

@@ -22,3 +22,4 @@ htmlcov/
 /metbit/__pycache__
 *.pyc
 *.pyc
+.claude

manuscript/figures/generate_graphical_abstract.py

@@ -201,13 +201,7 @@ def mini_heatmap(ax, cx, cy, w=1.0, h=0.6):
         color=WHITE, fontweight="bold", zorder=8)
 
 # ── footer ────────────────────────────────────────────────────────────────────
-ax.text(8, 0.62,
-        "pip install metbit   ·   github.com/aeiwz/metbit   ·   MIT License",
-        ha="center", va="center", fontsize=9, color=DARK, alpha=0.7, zorder=5)
 
-ax.text(0.35, 0.62,
-        "Bubpamala T. · kawa-technology · 2024  (AI-assisted writing: Claude Sonnet 4.6, GPT-5.2)",
-        ha="left", va="center", fontsize=8, color=DARK, alpha=0.55, zorder=5)
 
 # ── save ──────────────────────────────────────────────────────────────────────
 out = "graphical_abstract.png"

manuscript/figures/generate_workflow.py

@@ -346,11 +346,6 @@ def node(x, y, w, color, title, items,
 # ─────────────────────────────────────────────────────────────────────────────
 # FOOTER
 # ─────────────────────────────────────────────────────────────────────────────
-ax.text(CX, LEG_Y - 0.32,
-        "Bubpamala T.  ·  kawa-technology  ·  github.com/aeiwz/metbit  "
-        "(AI writing assistance: Claude Sonnet 4.6, GPT-5.2)",
-        ha="center", va="center", fontsize=7.5,
-        color=DARK, alpha=0.55, zorder=5)
 
 # ─────────────────────────────────────────────────────────────────────────────
 # SAVE  – clip canvas to content

manuscript/metbit_manuscript.md

@@ -4,12 +4,12 @@
 
 **Authors**
 
-Theerayut Bubpamala¹\*
+Theerayut Bubpamala<sup>1,2,*</sup>\*
+<sup>1</sup> Medical Biochemistry and Molecular Biology Graduate Study Program, Faculty of Medicine, Khon Kaen University, Khon Kaen, Thailand
+<sup>2</sup> kawa-technology, Independent Research & Development
 
-¹ kawa-technology, Independent Research & Development
-
-\* Corresponding author: theerayut_aeiw_123@hotmail.com  
-  GitHub: https://github.com/aeiwz/metbit  
+\* <sup>*</sup> Corresponding author: theerayut_aeiw_123@hotmail.com<br>
+  GitHub: https://github.com/aeiwz/metbit<br>
   PyPI: https://pypi.org/project/metbit/
 
 ---
@@ -34,7 +34,7 @@ Theerayut Bubpamala¹\*
 
 ## 1. Introduction
 
-Proton nuclear magnetic resonance (¹H NMR) spectroscopy is an essential platform for untargeted metabolomics due to its quantitative accuracy and non-destructive sample handling (Emwas et al., 2019). However, the journey from raw free-induction decay (FID) files to interpretable multivariate models involves complex sequential steps, including preprocessing, normalization, and alignment. While open-source tools such as NMRglue (Helmus and Jaroniec, 2013) and MetaboAnalyst (Pang et al., 2022) address parts of this pipeline, the Python ecosystem lacks a single, scriptable library that integrates the entire NMR metabolomics workflow natively.
+Proton nuclear magnetic resonance (<sup>1</sup>H NMR) spectroscopy is an essential platform for untargeted metabolomics due to its quantitative accuracy and non-destructive sample handling (Emwas et al., 2019). However, the journey from raw free-induction decay (FID) files to interpretable multivariate models involves complex sequential steps, including preprocessing, normalization, and alignment. While open-source tools such as NMRglue (Helmus and Jaroniec, 2013) and MetaboAnalyst (Pang et al., 2022) address parts of this pipeline, the Python ecosystem lacks a single, scriptable library that integrates the entire NMR metabolomics workflow natively.
 
 We present **metbit**, an end-to-end Python package that consolidates NMR data processing and multivariate modeling into a coherent API. metbit bridges the gap between raw spectral acquisitions and biological conclusions, delivering publication-quality interactive visualizations and reproducible workflows.
 

metbit/__init__.py

@@ -1,50 +1,71 @@
 # -*- coding: utf-8 -*-
 
-# Public classes/functions re-exported for convenience
-from .metbit import opls_da, pca
-from .utility import UnivarStats, Normalise
-from .lazy_opls_da import lazy_opls_da
-
-# Backwards-compat star imports (intended public surface)
-from .spec_norm import *  # noqa: F401,F403
-from .peak_processe import peak_chops
-from .STOCSY import STOCSY
-from .ui_stocsy import STOCSY_app
-from .ui_picky_peak import pickie_peak
-from .take_intensity import *  # noqa: F401,F403
-from .nmr_preprocess import nmr_preprocessing
-from .calibrate import calibrate
-from .annotate_peak import annotate_peak
-from .baseline import baseline_correct, bline
-from .alignment import detect_multiplets, icoshift_align, PeakAligner
-
 # Package metadata
 __author__ = "aeiwz"
 __email__ = "theerayut_aeiw_123@hotmail.com"
 __maintainer__ = "aeiwz"
 __status__ = "Development"
 __copyright__ = "Copyright 2024, Theerayut"
+__version__ = "9.0.0-dev"
+
+# Core analysis
+from metbit.analysis.opls_da import opls_da
+from metbit.analysis.pca import pca
+from metbit.analysis.stocsy import STOCSY
+
+# NMR processing
+from metbit.nmr.preprocess import nmr_preprocessing
+from metbit.nmr.calibrate import calibrate
+from metbit.nmr.alignment import detect_multiplets, icoshift_align, PeakAligner
+from metbit.nmr.peaks import peak_chops
+
+# Statistics
+from metbit.stats.normalise import Normalise
+from metbit.stats.univariate import UnivarStats
+
+# Preprocessing
+from metbit.preprocessing.baseline import baseline_correct, bline
+from metbit.preprocessing.normalize import Normalization
+
+# Dash apps (optional heavy deps)
+try:
+    from metbit.apps.annotate import annotate_peak
+except Exception:
+    pass
+
+try:
+    from metbit.apps.stocsy_app import STOCSY_app
+except Exception:
+    pass
 
-# Optional version string to keep in sync with setup.py
-__version__ = "8.7.7"
+try:
+    from metbit.apps.peak_picker import pickie_peak
+except Exception:
+    pass
+
+# Legacy flat-layout class not yet moved to a sub-package
+try:
+    from metbit.lazy_opls_da import lazy_opls_da
+except Exception:
+    pass
 
-# Explicit export list for primary API (star imports remain for compatibility)
 __all__ = [
     "opls_da",
     "pca",
-    "UnivarStats",
-    "Normalise",
-    "lazy_opls_da",
-    "peak_chops",
     "STOCSY",
-    "STOCSY_app",
-    "pickie_peak",
     "nmr_preprocessing",
     "calibrate",
-    "annotate_peak",
-    "baseline_correct",
-    "bline",
     "detect_multiplets",
     "icoshift_align",
     "PeakAligner",
+    "peak_chops",
+    "Normalise",
+    "UnivarStats",
+    "baseline_correct",
+    "bline",
+    "Normalization",
+    "annotate_peak",
+    "STOCSY_app",
+    "pickie_peak",
+    "lazy_opls_da",
 ]

metbit/_compat.py

@@ -0,0 +1,70 @@
+# Backwards-compatibility re-exports from sub-package paths.
+# Importing from old flat paths (e.g. `from metbit.nmr_preprocess import ...`)
+# still works via the originals; this module provides the NEW canonical paths
+# so external consumers can start migrating to the sub-package imports.
+
+# --- nmr/ ---
+from .nmr.preprocess import (  # noqa: F401
+    nmr_preprocessing,
+    read_fid,
+    remove_digital_filter,
+    generate_ppm_scale,
+    phasing,
+)
+from .nmr.calibrate import calibrate  # noqa: F401
+from .nmr.alignment import (  # noqa: F401
+    detect_multiplets,
+    icoshift_align,
+    PeakAligner,
+    Multiplet,
+)
+from .nmr.denoise import Denoise  # noqa: F401
+from .nmr.peaks import peak_chops  # noqa: F401
+
+# --- models/ ---
+from .models.base import nipals  # noqa: F401
+from .models.pls import PLS  # noqa: F401
+from .models.opls import OPLS  # noqa: F401
+from .models.vip import vip_scores  # noqa: F401
+from .models.cross_validation import CrossValidation  # noqa: F401
+
+# --- preprocessing/ ---
+from .preprocessing.scaler import Scaler as SimpleScaler  # noqa: F401
+from .preprocessing.scaler_ext import Scaler  # noqa: F401
+from .preprocessing.normalize import Normalization  # noqa: F401
+from .preprocessing.baseline import baseline_correct, bline  # noqa: F401
+
+# --- analysis/ ---
+from .analysis.stocsy import STOCSY  # noqa: F401
+
+# --- viz/ ---
+from .viz.ellipse import confidence_ellipse  # noqa: F401
+
+# --- higher-level classes (canonical sub-package paths) ---
+from .analysis.opls_da import opls_da  # noqa: F401
+from .analysis.pca import pca  # noqa: F401
+
+try:
+    from .lazy_opls_da import lazy_opls_da  # noqa: F401
+except Exception:
+    pass
+
+# --- utility classes (canonical sub-package paths) ---
+from .stats.univariate import UnivarStats  # noqa: F401
+from .stats.normalise import Normalise  # noqa: F401
+
+# --- Dash apps (optional heavy deps) ---
+try:
+    from .apps.annotate import annotate_peak  # noqa: F401
+except Exception:
+    pass
+
+try:
+    from .apps.stocsy_app import STOCSY_app  # noqa: F401
+except Exception:
+    pass
+
+try:
+    from .apps.peak_picker import pickie_peak  # noqa: F401
+except Exception:
+    pass

metbit/_internal/pairs.py

@@ -0,0 +1,57 @@
+# -*- coding: utf-8 -*-
+
+import pandas as pd
+
+
+class lazypair:
+    """Utility for generating all pairwise groupings and related dataset splits."""
+
+    def __init__(self, dataset: pd.DataFrame, column_name: str):
+        if not isinstance(dataset, pd.DataFrame):
+            raise ValueError("meta should be a pandas dataframe")
+        if not isinstance(column_name, str):
+            raise ValueError("column_name should be a string")
+        if column_name not in dataset.columns:
+            raise KeyError(f"{column_name} not found in dataset")
+        if dataset[column_name].nunique() < 2:
+            raise ValueError("Group should contain at least 2 groups")
+
+        self.meta = dataset
+        self.column_name = column_name
+
+        groups = dataset[column_name].unique()
+        self.pairs = [(g1, g2) for i, g1 in enumerate(groups) for g2 in groups[i + 1:]]
+        self.index_ = [
+            dataset.index[dataset[column_name] == g1].tolist()
+            + dataset.index[dataset[column_name] == g2].tolist()
+            for g1, g2 in self.pairs
+        ]
+        self.names = [f"{g1}_vs_{g2}".replace("/", "_") for g1, g2 in self.pairs]
+
+    def get_index(self):
+        index_ = self.index_
+        return index_
+
+    def get_name(self):
+        names = self.names
+        return names
+
+    def get_meta(self):
+        meta = self.meta
+        column_name = self.column_name
+        return meta[column_name]
+
+    def get_column_name(self):
+        column_name = self.column_name
+        return column_name
+
+    def get_dataset(self):
+        df = self.meta
+        index_ = self.index_
+        list_of_df = []
+        for i in range(len(index_)):
+            list_of_df.append(df.loc[index_[i]])
+
+        #Create object attribute
+        self.list_of_df = list_of_df
+        return list_of_df