Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

Unreleased

Added

Pipeline

• pipeline.run_mipar() gains a phase_aliases parameter to normalise phase name capitalisation variants across files.

• pipeline.run_mipar() gains a missing_phase parameter ("raise" / "warn" / "skip") to handle states where a phase is absent entirely.

Notebooks

• 04_mipar_feature_analysis — new section (§7) reports nearest-neighbour spacing and particle number density per phase from MIPAR’s Average Neighbor Distance column and per-FOV feature counts: direct geometric measurements that give physically intuitive gap sizes for dilute phases (MX, ZPhase) where the intercept-based mean free path in notebook 05 is dominated by empty stretches of matrix.

• 04_mipar_feature_analysis — new section (§8) reports total phase fraction per FOV per phase, computed by summing individual feature area fractions; cross-checked against the batch measurements in notebook 05.

Changed

Notebooks

• 04_mipar_feature_analysis and 05_mipar_image_analysis now cover all 7 material states (0 kh, 12 kh, 51 kh, 81 kh, 139 kh gauge, 139 kh thread, 139 kh fracture) instead of the previous three; all outputs re-executed.

• 05_mipar_image_analysis masks zero or negative mean-intercept values before computing stereological quantities to prevent division errors.

Fixed

• Untracked CSV files under notebooks/data/ removed from the repository; .gitignore tightened to exclude all CSV, Excel, PNG, PDF, SVG, TIFF, and TeX outputs in that directory.

0.1.0a6 - 2026-05-18

Changed

Docs

• Landing page (index.rst) redesigned with a description, tab-set install snippet, sphinx-design card tiles for Examples, API Reference, Installation, and Contributing, plus a Citing STAMP section.

• Installation instructions corrected to use the PyPI package name nanoshot-stamp (was incorrectly stamp); notebook extras documented.

• Contributing guide and changelog now rendered directly in the docs rather than linking out to GitHub.

• Added sphinx-design>=0.5 to docs dependencies for grid/card layout.

Notebooks

• 04_mipar_feature_analysis and 05_mipar_image_analysis — updated to reflect a 4-phase MIPAR segmentation where MX ZPhase is split into two independent phases (MX and ZPhase); FILES and STATE_LABELS trimmed to the three available material states (GOO220_51 gauge, _52 thread, _53 fracture)

• run_mipar pipeline gains phase_aliases and missing_phase parameters so notebooks can normalise phase name capitalisation variants and handle states where a phase is absent without raising an error

• Notebook titles simplified (prefix STAMP — removed from all five notebooks)

Fixed

Notebooks

• Corrected mojibake in 05_mipar_image_analysis (axis labels, print statements, and LaTeX captions displayed garbled characters such as â€" and µ instead of — and µ)

0.1.0a5 - 2026-05-17

Added

Stereology

• Derive volume fraction ($V_V$), surface area density ($S_V$), mean caliper diameter ($\bar{D}$), and 3-D mean free path ($\lambda_{3D}$) from MIPAR image-level measurements — applies Delesse (1848), Underwood (1970), and Fullman (1953) relations directly to per-FOV DataFrame columns (stamp.stereo.volume_fraction, surface_area_density, mean_caliper_diameter, mean_free_path_3d)

I/O

• Load MIPAR image-measurement batch CSVs (one row per FOV, all phases as column suffixes) into a tidy long-format DataFrame with one row per (FOV × phase) combination (stamp.io.load_mipar_image())

Export

• Publication-ready figure styling and table export — apply journal formatting (default, Nature preset, or custom overrides) to any STAMP plot and export results tables as CSV or LaTeX booktabs (stamp.export)

Notebooks

• 05_mipar_image_analysis — new notebook demonstrating load_mipar_image across three material states; derives and compares 2-D quantities (phase fraction, mean particle size, interparticle spacing) and 3-D stereological quantities ($V_V$, $S_V$, $\bar{D}$, $\lambda_{3D}$) per phase with Nature journal-style box plots and LaTeX table exports

• 04_mipar_feature_analysis — plots converted to Nature journal style (B&W, hatch-differentiated boxes, correct column widths); summary table now also exported as a LaTeX booktabs file via stamp.export.to_latex

0.1.0a4 - 2026-05-16

Added

• Install JupyterLab and ipykernel alongside STAMP with pip install "nanoshot-stamp[notebooks]"

Fixed

• Minor development and packaging fixes

0.1.0a2 - 2026-05-16

Added

• Optional notebooks extra — installs JupyterLab and ipykernel for running the example notebooks (pip install "nanoshot-stamp[notebooks]")

0.1.0a1 - 2026-05-16

Added

Pipeline

• Multi-state scripted analysis pipeline — run the full load → statistics workflow across any number of material states (heat treatments, compositions, processing routes, etc.), each with multiple fields-of-view (stamp.pipeline.run())

• Side-by-side box plot comparing per-field-of-view mean statistics across all material states, with individual data points overlaid (stamp.pipeline.boxplot())

• Export the per-field-of-view summary table (arithmetic mean, geometric mean, median, CIs, percentiles) to CSV (stamp.pipeline.export_csv())

• Run the full pipeline directly on MIPAR feature-measurement CSVs — groups rows by image/FOV, filters to a chosen precipitate phase, and returns the same PipelineResult as run() (stamp.pipeline.run_mipar())

I/O

• Load grain measurements from CSV, Excel (.xlsx/.xls), or plain-text files — stamp.io.load() now returns a single-column pd.DataFrame with physical unit and label stored in df.attrs, making the loading API consistent with load_mipar_features()

• Load a MIPAR feature-measurement CSV into a pandas DataFrame for custom filtering and inspection (stamp.io.load_mipar_features())

Stereology

• Convert 2-D projected areas to equivalent circle diameters (stamp.stereo.ecd_from_area())

• Fullman (1953) linear intercept correction for 2-D → 3-D mean grain diameter (stamp.stereo.linear_intercept_correction())

• Saltykov/Wicksell matrix unfolding — recovers the 3-D sphere-diameter frequency distribution from 2-D circle measurements, including a volume-weighted CDF (stamp.stereo.saltykov())

• Two-step lognormal fit (Lopez-Sanchez & Llana-Funez 2016) — iterates Saltykov over a range of bin counts and returns the best-fit geometric mean and log-shape σ with a ±3σ uncertainty band (stamp.stereo.two_step())

Statistics

• Arithmetic mean with ASTM, GCI, and mCox confidence intervals (stamp.stats.amean())

• Geometric mean with CLT and Bayesian confidence intervals (stamp.stats.gmean())

• Median with IQR and Hollander–Wolfe confidence interval (stamp.stats.median())

• KDE mode estimation with Silverman, Scott, or user-supplied bandwidth (stamp.stats.freq_peak())

• MLE distribution fitting for normal and lognormal with KS goodness-of-fit test (stamp.stats.fit())

• Single-call summary of all descriptive statistics (stamp.stats.describe())

Plots

• Histogram + KDE with annotated averages and optional fitted distribution overlay (stamp.plot.distribution())

• Dual-panel 3-D frequency and volume-weighted CDF figure for Saltykov results (stamp.plot.saltykov_plot())

• Lognormal fit curve with ±3σ uncertainty band for two-step results (stamp.plot.twostep_plot())

• PDF or empirical CDF profile (stamp.plot.distribution_profile())

• Quantile-quantile plot against normal or lognormal reference (stamp.plot.qq_plot())

• Dual-panel comparison of 2-D apparent vs corrected 3-D distributions with recovery error annotation (stamp.plot.comparison_plot())

Simulation

• Monte Carlo Wicksell corpuscle simulation — generates a synthetic lognormal or normal 3-D grain population and random 2-D cross-sections for validating stereological corrections (stamp.simulate.simulate_section())

Notebooks

• notebooks/01_quickstart.ipynb — end-to-end workflow: load measurements, compute statistics, apply Saltykov and two-step corrections, generate all plots

• notebooks/02_simulation_validation.ipynb — Monte Carlo validation of stereological corrections including Wicksell bias demo, recovery accuracy table, and sample-size sweep

Documentation

• Added STAMP logo (stylised grain-boundary pattern) shown in the docs navbar and README

• Switched documentation theme to PyData Sphinx Theme with a top navigation bar (Installation, Examples, Contributing, Changelog, API Reference) and GitHub link