Fitter

Uniform LUT fitting and C code generation for analytic functions and table data, with GUI and CLI workflows.

Current Project State

The project now includes:

• a Tkinter GUI with fit controls, plotting, history, batch runs, and multilingual labels/tooltips
• a CLI for scripted use
• a split pipeline (fit compute + render/codegen) so code output can be regenerated without re-running the fit
• startup dependency preflight from requirements.txt
• portable runtime storage/config for source and compiled builds
• codegen formatting features (wrapped arrays, array placement options)
• future-ready algorithm scaffolding (current shipped method: uniform LUT)

Project Layout

• core/: fitting/evaluation math and shared types
• ios/: source loading/parsing (tables + analytic expressions)
• codegen/: C generators, formatting, memory estimates
• app/: GUI, CLI, pipeline, reporting, runtime config/bootstrap
• build/: PyInstaller specs (GUI + CLI)
• tests/: tests and sample data
• fitter.py: source-mode dispatcher entry point

Quick Start

Install dependencies:

python -m pip install -r requirements.txt

The app performs a startup preflight check for runtime dependencies and exits early with a clear message if something is missing.

Running (Source / Python)

Default dispatcher (GUI if no args, CLI if args are present)

python fitter.py
python -m app

Examples using the dispatcher in CLI mode:

python fitter.py tests/test.csv --float --test-count 0
python -m app --func "sin(x) + 0.1*x*x" --xmin -2 --xmax 5

Explicit GUI

python -m app.gui

Explicit CLI

python -m app.cli <path> [flags]
python -m app.cli --func <expr|name> --xmin <x0> --xmax <x1> [flags]

CLI Usage Examples

Table source (CSV / TSV / space-delimited)

python -m app.cli tests/test.csv --linear --float --test-count 0
python -m app.cli tests/test.csv --spline --double --with-slopes

Analytic source

python -m app.cli --func sin --xmin -2 --xmax 5
python -m app.cli --func linear --a 0.7 --b 0.1 --xmin -2 --xmax 5
python -m app.cli --func "sin(x) + 0.1*x*x" --xmin -2 --xmax 5
python -m app.cli --func-file expr.txt --xmin -2 --xmax 5

Built-in analytic names:

• sin, cos, exp, log, linear

Output file

python -m app.cli tests/test.csv --out out/fit_report
python -m app.cli tests/test.csv --out out/fit_report.txt

• If --out has no extension, .txt is appended.
• When --out is set, the report is written to file instead of stdout.

CLI Flags (Current)

Source / interpolation

• --linear / --spline
• <path> for table data
• --func <expr|name>
• --func-file <path>
• --xmin <float> --xmax <float>
• --a <float> --b <float> (for linear)

Error / fit controls

• --abs-err <float>
• --sym-err <float>
• --n-max <int>
• --xmin-shift

Code generation controls

• --float (default)
• --double
• --int16
• --int32
• --with-slopes
• --test-count <int>
• --func-name <str>
• --array-name <str>
• --q-frac <int> / --x-shift <int> / --y-scale <int> (int modes)

Misc

• --plot (requires matplotlib)
• --out <path>

GUI Highlights (Current)

• Table and analytic sources
• Interpolation mode near table-source controls
• Parameter section for error limits, auto-relax, and Optimize xmin
• Fitting tabs (Basic, Advanced, Algorithms) with current uniform LUT controls in Basic
• Codegen section with numeric mode, slopes, test count, names, array placement, fixed-point settings
• Update Code behavior:
  • code/summary are regenerated from cached fit data
  • changing codegen parameters auto-updates output without re-running fitting
• Run progress info shown near the progress bar (estimated N, current N, best-so-far)
• Plot tab with separate X scales for fit/error subplots
• History restore:
  • restores config and rendered outputs immediately
  • restores fit/plot too for newer history entries with cached fit payloads

Runtime Config and Storage

The app uses a bundled default config (app/default_config.json) plus an optional user override file.

Source (Python) runs

• storage root: .fitter_data/ in the repo

Frozen / compiled runs

• storage root: fitter_data/ next to the executable

Stored items

• config.json (user overrides)
• history.json
• outputs/ (default export directory)

Config-driven options (current)

• timing logging on/off (logging.enabled)
• codegen max line length for wrapped arrays
• default array placement
• code preview syntax highlighting toggle

Expression Syntax

• Variable: x
• Operators: + - * / ^ (^ means power)
• Parentheses: ( )
• Constants: pi, e
• Functions include:
  • sin, cos, tan, exp, log, sqrt, abs, pow, min, max
  • asin, acos, atan, floor, ceil, sinh, cosh, tanh

Build (PyInstaller)

Preferred specs (current):

• build/pyinstaller_gui.spec -> FitterGUI
• build/pyinstaller_cli.spec -> FitterCLI

Example:

pyinstaller build/pyinstaller_gui.spec
pyinstaller build/pyinstaller_cli.spec

There is also a legacy build/pyinstaller.spec in the repo.

Tests

Useful test commands:

python -m unittest tests.test_pipeline tests.test_ls_fit tests.test_segment_estimator
python -m unittest tests.test_codegen_formatting tests.test_runtime_bootstrap

Notes

• Current fitting method shipped in the GUI/CLI is uniform LUT.
• The codebase already includes planning documents for future expansion (Plan.md, BenchmarkSpec.md).