modelit 0.1.0__tar.gz → 0.2.2__tar.gz

modelit-0.2.2/.github/workflows/publish.yml

@@ -0,0 +1,39 @@
+name: publish
+
+on:
+  push:
+    tags:
+      - "v*"
+  workflow_dispatch:
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+      - run: python -m pip install --upgrade pip build setuptools-scm
+      - run: python -m build
+      - uses: actions/upload-artifact@v4
+        with:
+          name: dist
+          path: dist/*
+
+  publish:
+    needs: build
+    runs-on: ubuntu-latest
+    environment:
+      name: pypi
+      url: https://pypi.org/p/modelit
+    permissions:
+      id-token: write
+    steps:
+      - uses: actions/download-artifact@v4
+        with:
+          name: dist
+          path: dist
+      - uses: pypa/gh-action-pypi-publish@release/v1

modelit-0.2.2/.github/workflows/test.yml

@@ -0,0 +1,19 @@
+name: test
+
+on:
+  push:
+  pull_request:
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+      - run: python -m pip install --upgrade pip build twine
+      - run: python -m build
+      - run: twine check dist/*
+      - run: pip install dist/*.whl
+      - run: python -c "from modelit import perceptron; print(perceptron.__name__)"

modelit-0.2.2/.gitignore

@@ -0,0 +1,15 @@
+__pycache__/
+*.py[cod]
+*.egg-info/
+build/
+dist/
+.pytest_cache/
+.mypy_cache/
+.ruff_cache/
+.venv/
+
+venv/
+.DS_Store
+code1.py
+screenshots/
+

{modelit-0.1.0/modelit.egg-info → modelit-0.2.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: modelit
-Version: 0.1.0
+Version: 0.2.2
 Summary: Local-first ML starter templates you can print or save
 Author: Yashsmith
 License-Expression: MIT
@@ -20,7 +20,14 @@ Dynamic: license-file
 
 # ModelIt
 
-Tiny local-first ML template printer.
+ModelIt is a tiny Python package for storing your ML/DL boilerplate templates.
+
+## What it does
+
+- `from modelit import perceptron`
+- `perceptron()` prints the full code
+- `perceptron(output="code1.py")` saves it to a file
+- `modelit create perceptron` works from the terminal
 
 ## Install
 
@@ -28,13 +35,15 @@ Tiny local-first ML template printer.
 pip install modelit
 ```
 
-For development from a clone, use:
+For local development:
 
 ```bash
 pip install -e .
 ```
 
-## Run
+## Use it
+
+### Simple Python Run which prints the code.
 
 ```python
 from modelit import perceptron
@@ -42,13 +51,7 @@ from modelit import perceptron
 perceptron()
 ```
 
-Or from CLI:
-
-```bash
-modelit create perceptron
-```
-
-## Save to file
+### Save to file And then Run so the file is created with code in it.
 
 ```python
 from modelit import perceptron
@@ -56,39 +59,57 @@ from modelit import perceptron
 perceptron(output="code1.py")
 ```
 
-Or:
-
-```bash
-modelit create perceptron --output code1.py
-```
-
 Then run:
 
 ```bash
 python3 code1.py
 ```
 
-## Publish
+### CLI
 
-This project is set up for PyPI publishing with GitHub Actions.
+```bash
+modelit create perceptron
+modelit create perceptron --output code1.py
+```
 
-1. Push a tag like `v0.1.0`
-2. GitHub Actions builds and publishes to PyPI
-3. Users install with `pip install modelit`
+## Add a new code
 
-To update later, change the version, add your new templates, and publish a new tag.
+1. Create a folder:
 
-## Add a new code
+```text
+modelit/templates/mycode/
+```
 
-Create:
+2. Add your code file:
 
 ```text
-modelit/templates/<name>/template.py
-modelit/templates/<name>/metadata.json
+modelit/templates/mycode/template.py
 ```
 
-Then the new name is exposed automatically as:
+3. Done. The folder name becomes the function name.
+
+That means this will work automatically:
 
 ```python
-from modelit import <name>
+from modelit import mycode
+
+mycode()
 ```
+
+And this too:
+
+```bash
+modelit create mycode
+modelit create mycode --output mycode.py
+```
+
+## Publish flow
+
+1. Add a new template folder.
+2. `git add .`
+3. `git commit -m "..."`
+4. `git push`
+5. `git tag v0.x.y`
+6. `git push origin v0.x.y`
+
+The version is taken from the git tag, so you do not edit `pyproject.toml` for releases.

modelit-0.2.2/README.md

@@ -0,0 +1,95 @@
+# ModelIt
+
+ModelIt is a tiny Python package for storing your ML/DL boilerplate templates.
+
+## What it does
+
+- `from modelit import perceptron`
+- `perceptron()` prints the full code
+- `perceptron(output="code1.py")` saves it to a file
+- `modelit create perceptron` works from the terminal
+
+## Install
+
+```bash
+pip install modelit
+```
+
+For local development:
+
+```bash
+pip install -e .
+```
+
+## Use it
+
+### Simple Python Run which prints the code.
+
+```python
+from modelit import perceptron
+
+perceptron()
+```
+
+### Save to file And then Run so the file is created with code in it.
+
+```python
+from modelit import perceptron
+
+perceptron(output="code1.py")
+```
+
+Then run:
+
+```bash
+python3 code1.py
+```
+
+### CLI
+
+```bash
+modelit create perceptron
+modelit create perceptron --output code1.py
+```
+
+## Add a new code
+
+1. Create a folder:
+
+```text
+modelit/templates/mycode/
+```
+
+2. Add your code file:
+
+```text
+modelit/templates/mycode/template.py
+```
+
+3. Done. The folder name becomes the function name.
+
+That means this will work automatically:
+
+```python
+from modelit import mycode
+
+mycode()
+```
+
+And this too:
+
+```bash
+modelit create mycode
+modelit create mycode --output mycode.py
+```
+
+## Publish flow
+
+1. Add a new template folder.
+2. `git add .`
+3. `git commit -m "..."`
+4. `git push`
+5. `git tag v0.x.y`
+6. `git push origin v0.x.y`
+
+The version is taken from the git tag, so you do not edit `pyproject.toml` for releases.

{modelit-0.1.0 → modelit-0.2.2}/modelit/registry.py

@@ -5,22 +5,16 @@ from __future__ import annotations
 from dataclasses import dataclass
 from importlib.resources import files
 from pathlib import Path
-import json
-import sys
 
 
 PACKAGE_NAME = "modelit"
 TEMPLATES_DIR = "templates"
 TEMPLATE_FILENAME = "template.py"
-METADATA_FILENAME = "metadata.json"
 
 
 @dataclass(frozen=True)
 class TemplateInfo:
     name: str
-    title: str
-    output_file: str
-    description: str
 
 
 def _templates_root():
@@ -44,21 +38,7 @@ def available_models() -> tuple[str, ...]:
 
 
 def load_metadata(name: str) -> TemplateInfo:
-    metadata_path = _template_dir(name).joinpath(METADATA_FILENAME)
-    data: dict[str, str] = {}
-    if metadata_path.is_file():
-        data = json.loads(metadata_path.read_text(encoding="utf-8"))
-
-    title = data.get("title") or name.replace("_", " ").title()
-    output_file = data.get("output_file") or f"{name}.py"
-    description = data.get("description") or f"Print the {title} template."
-
-    return TemplateInfo(
-        name=name,
-        title=title,
-        output_file=output_file,
-        description=description,
-    )
+    return TemplateInfo(name=name)
 
 
 def load_source(name: str) -> str:
@@ -69,8 +49,9 @@ def load_source(name: str) -> str:
 
 
 def build_template_callable(name: str):
-    info = load_metadata(name)
     source = load_source(name)
+    info = load_metadata(name)
+    output_file = f"{name}.py"
 
     def runner(output: str | None = None) -> None:
         if output:
@@ -82,13 +63,12 @@ def build_template_callable(name: str):
             print(f"Generated {path}")
             return None
 
-        for chunk in source.splitlines(keepends=True):
-            sys.stdout.write(chunk)
+        print(source, end="")
 
     runner.__name__ = name
     runner.__qualname__ = name
     runner.__module__ = "modelit"
-    runner.__doc__ = info.description
-    runner.output_file = info.output_file  # type: ignore[attr-defined]
+    runner.__doc__ = f"Print or save the {name} template."
+    runner.output_file = output_file  # type: ignore[attr-defined]
     runner.template_info = info  # type: ignore[attr-defined]
     return runner

modelit-0.2.2/modelit/templates/backpropogation/template.py

@@ -0,0 +1,483 @@
+# =========================================================
+# BACKPROPAGATION FROM SCRATCH
+# XOR using SGD
+# Xavier Initialization
+# Titanic Dataset using MLPClassifier
+# =========================================================
+
+
+# =========================================================
+# IMPORTS
+# =========================================================
+
+import numpy as np
+import pandas as pd
+
+from sklearn.datasets import load_breast_cancer
+from sklearn.model_selection import train_test_split
+from sklearn.neural_network import MLPClassifier
+from sklearn.preprocessing import StandardScaler
+
+
+# =========================================================
+# ACTIVATION FUNCTION
+# SIGMOID
+# =========================================================
+
+def activation_func(net):
+
+    return 1 / (1 + np.exp(-net))
+
+
+# =========================================================
+# ERROR FUNCTION
+# E = 1/2 (d - o)^2
+# =========================================================
+
+def calc_error(output, det_output):
+
+    return (1 / 2) * (det_output - output) ** 2
+
+
+# =========================================================
+# FORWARD PASS
+#
+# Architecture:
+# 2 Input Neurons
+# 2 Hidden Neurons
+# 1 Output Neuron
+# =========================================================
+
+def forward_pass(x_input, weights, biases):
+
+    # Hidden Neuron h0
+    net_h0 = (
+        (x_input[0] * weights[0]) +
+        (x_input[1] * weights[1]) +
+        biases[0]
+    )
+
+    outh0 = activation_func(net_h0)
+
+    # Hidden Neuron h1
+    net_h1 = (
+        (x_input[0] * weights[2]) +
+        (x_input[1] * weights[3]) +
+        biases[1]
+    )
+
+    outh1 = activation_func(net_h1)
+
+    # Output Neuron
+    net_o0 = (
+        (outh0 * weights[4]) +
+        (outh1 * weights[5]) +
+        biases[2]
+    )
+
+    out_o0 = activation_func(net_o0)
+
+    return outh0, outh1, out_o0
+
+
+# =========================================================
+# BACKWARD PASS
+# =========================================================
+
+def backward_pass(
+    x_input,
+    weights,
+    biases,
+    desired_output,
+    learning_rate,
+    outh0,
+    outh1,
+    out_o0
+):
+
+    # Output layer delta
+    delta_o0 = (
+        (desired_output - out_o0) *
+        out_o0 *
+        (1 - out_o0)
+    )
+
+    # Hidden layer deltas
+    delta_h0 = (
+        delta_o0 *
+        weights[4] *
+        outh0 *
+        (1 - outh0)
+    )
+
+    delta_h1 = (
+        delta_o0 *
+        weights[5] *
+        outh1 *
+        (1 - outh1)
+    )
+
+    # Weight updates
+    dw0 = learning_rate * delta_h0 * x_input[0]
+    dw1 = learning_rate * delta_h0 * x_input[1]
+
+    dw2 = learning_rate * delta_h1 * x_input[0]
+    dw3 = learning_rate * delta_h1 * x_input[1]
+
+    dw4 = learning_rate * delta_o0 * outh0
+    dw5 = learning_rate * delta_o0 * outh1
+
+    dw_values = [dw0, dw1, dw2, dw3, dw4, dw5]
+
+    # Bias updates
+    db0 = learning_rate * delta_h0
+    db1 = learning_rate * delta_h1
+    db2 = learning_rate * delta_o0
+
+    db_values = [db0, db1, db2]
+
+    return dw_values, db_values
+
+
+# =========================================================
+# XOR DATASET
+# =========================================================
+
+x1 = [0, 0]
+x2 = [0, 1]
+x3 = [1, 0]
+x4 = [1, 1]
+
+y = [0, 1, 1, 0]
+
+X = [x1, x2, x3, x4]
+d = y
+
+
+# =========================================================
+# INITIAL WEIGHTS AND BIASES
+# =========================================================
+
+w = [0.5, -0.3, 0.8, 0.2, 0.6, -0.5]
+
+b = [0.2, -0.4, 0.3]
+
+learning_rate_initial = 0.01
+
+epochs = 2000
+
+error_threshold = 0.001
+
+
+# =========================================================
+# TRAINING USING SGD
+# =========================================================
+
+def train_SGD(
+    X_data,
+    y_data,
+    epochs_count,
+    lr_val,
+    initial_weights_list,
+    initial_biases_list
+):
+
+    w1, w2, w3, w4, w5, w6 = (
+        initial_weights_list[0],
+        initial_weights_list[1],
+        initial_weights_list[2],
+        initial_weights_list[3],
+        initial_weights_list[4],
+        initial_weights_list[5]
+    )
+
+    b1, b2, b3 = (
+        initial_biases_list[0],
+        initial_biases_list[1],
+        initial_biases_list[2]
+    )
+
+    for epoch in range(epochs_count):
+
+        total_loss = 0
+
+        for i in range(len(X_data)):
+
+            x_input = X_data[i]
+
+            target = y_data[i]
+
+            current_weights = [w1, w2, w3, w4, w5, w6]
+
+            current_biases = [b1, b2, b3]
+
+            # Forward pass
+            outh0, outh1, out_o0 = forward_pass(
+                x_input,
+                current_weights,
+                current_biases
+            )
+
+            # Loss
+            current_loss = calc_error(out_o0, target)
+
+            total_loss += current_loss
+
+            # Backpropagation
+            dw_values, db_values = backward_pass(
+                x_input,
+                current_weights,
+                current_biases,
+                target,
+                lr_val,
+                outh0,
+                outh1,
+                out_o0
+            )
+
+            # Update weights
+            w1 += dw_values[0]
+            w2 += dw_values[1]
+            w3 += dw_values[2]
+            w4 += dw_values[3]
+            w5 += dw_values[4]
+            w6 += dw_values[5]
+
+            # Update biases
+            b1 += db_values[0]
+            b2 += db_values[1]
+            b3 += db_values[2]
+
+        # Print every 100 epochs
+        if epoch % 100 == 0 or total_loss < error_threshold:
+
+            print(f"\nEpoch {epoch}")
+
+            print(f"Total Error = {total_loss:.6f}")
+
+            output_example = forward_pass(
+                X_data[0],
+                [w1, w2, w3, w4, w5, w6],
+                [b1, b2, b3]
+            )[2]
+
+            print(
+                f"Output for input {X_data[0]} : "
+                f"{output_example:.4f}"
+            )
+
+        # Early stopping
+        if total_loss < error_threshold:
+
+            print(
+                f"\nStopping condition met at epoch "
+                f"{epoch + 1}"
+            )
+
+            print(
+                f"Total Error ({total_loss:.6f}) "
+                f"<= Threshold ({error_threshold:.6f})"
+            )
+
+            break
+
+    updated_weights = [w1, w2, w3, w4, w5, w6]
+
+    updated_biases = [b1, b2, b3]
+
+    return updated_weights, updated_biases
+
+
+# =========================================================
+# TRAIN XOR NETWORK
+# =========================================================
+
+updated_w, updated_b = train_SGD(
+    X,
+    d,
+    epochs,
+    learning_rate_initial,
+    w,
+    b
+)
+
+
+# =========================================================
+# FINAL WEIGHTS
+# =========================================================
+
+print("\nFinal Updated Weights:")
+
+for i, weight in enumerate(updated_w):
+
+    print(f"w{i + 1} = {weight:.4f}")
+
+
+# =========================================================
+# FINAL BIASES
+# =========================================================
+
+print("\nFinal Updated Biases:")
+
+for i, bias in enumerate(updated_b):
+
+    print(f"b{i + 1} = {bias:.4f}")
+
+
+# =========================================================
+# TEST XOR OUTPUTS
+# =========================================================
+
+print("\n=================================================")
+print("XOR Predictions")
+print("=================================================")
+
+for i in range(len(X)):
+
+    _, _, prediction = forward_pass(
+        X[i],
+        updated_w,
+        updated_b
+    )
+
+    print(
+        f"Input: {X[i]} "
+        f"Target: {d[i]} "
+        f"Prediction: {prediction:.4f}"
+    )
+
+
+# =========================================================
+# XAVIER INITIALIZATION
+# =========================================================
+
+print("\n=================================================")
+print("Xavier Initialization")
+print("=================================================")
+
+nin = 2
+nhidden = 2
+nout = 1
+
+
+def xavier(nin, nout):
+
+    limit = np.sqrt(6 / (nin + nout))
+
+    return np.random.uniform(
+        -limit,
+        limit,
+        (nin, nout)
+    )
+
+
+w1_xavier = xavier(nin, nhidden)
+
+w2_xavier = xavier(nhidden, nout)
+
+print("\nWeights between Input -> Hidden:")
+print(w1_xavier)
+
+print("\nWeights between Hidden -> Output:")
+print(w2_xavier)
+
+
+# =========================================================
+# TITANIC DATASET USING MLPClassifier
+# =========================================================
+
+print("\n=================================================")
+print("Breast Cancer Dataset using MLPClassifier")
+print("=================================================")
+
+dataset = load_breast_cancer(as_frame=True)
+df = dataset.frame
+
+print("\nDataset Head:")
+print(df.head())
+
+
+# =========================================================
+# FEATURES AND LABELS
+# =========================================================
+
+X_cancer = df.drop(columns=["target"])
+y_cancer = df["target"]
+
+
+# =========================================================
+# FEATURE SCALING
+# =========================================================
+
+scaler = StandardScaler()
+X_scaled = scaler.fit_transform(X_cancer)
+
+
+# =========================================================
+# TRAIN TEST SPLIT
+# =========================================================
+
+X_train, X_test, y_train, y_test = train_test_split(
+    X_scaled,
+    y_cancer,
+    test_size=0.2,
+    random_state=42
+)
+
+
+# =========================================================
+# MLP MODEL
+# =========================================================
+
+mlp = MLPClassifier(
+    hidden_layer_sizes=(10,),
+    activation="logistic",
+    solver="sgd",
+    learning_rate_init=0.01,
+    random_state=42,
+    max_iter=2000
+)
+
+
+# =========================================================
+# TRAIN MODEL
+# =========================================================
+
+mlp.fit(X_train, y_train)
+
+
+# =========================================================
+# PRINT WEIGHTS
+# =========================================================
+
+print("\nWeights (Coefficients):")
+
+for i, coef in enumerate(mlp.coefs_):
+
+    print(f"\nLayer {i} weights:\n")
+
+    print(coef)
+
+
+# =========================================================
+# PRINT BIASES
+# =========================================================
+
+print("\nBiases (Intercepts):")
+
+for i, intercept in enumerate(mlp.intercepts_):
+
+    print(f"\nLayer {i} biases:\n")
+
+    print(intercept)
+
+
+# =========================================================
+# MODEL ACCURACY
+# =========================================================
+
+accuracy = mlp.score(X_test, y_test)
+
+print("\n=================================================")
+print(f"Model Accuracy: {accuracy * 100:.2f}%")
+print("=================================================")

{modelit-0.1.0 → modelit-0.2.2/modelit.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: modelit
-Version: 0.1.0
+Version: 0.2.2
 Summary: Local-first ML starter templates you can print or save
 Author: Yashsmith
 License-Expression: MIT
@@ -20,7 +20,14 @@ Dynamic: license-file
 
 # ModelIt
 
-Tiny local-first ML template printer.
+ModelIt is a tiny Python package for storing your ML/DL boilerplate templates.
+
+## What it does
+
+- `from modelit import perceptron`
+- `perceptron()` prints the full code
+- `perceptron(output="code1.py")` saves it to a file
+- `modelit create perceptron` works from the terminal
 
 ## Install
 
@@ -28,13 +35,15 @@ Tiny local-first ML template printer.
 pip install modelit
 ```
 
-For development from a clone, use:
+For local development:
 
 ```bash
 pip install -e .
 ```
 
-## Run
+## Use it
+
+### Simple Python Run which prints the code.
 
 ```python
 from modelit import perceptron
@@ -42,13 +51,7 @@ from modelit import perceptron
 perceptron()
 ```
 
-Or from CLI:
-
-```bash
-modelit create perceptron
-```
-
-## Save to file
+### Save to file And then Run so the file is created with code in it.
 
 ```python
 from modelit import perceptron
@@ -56,39 +59,57 @@ from modelit import perceptron
 perceptron(output="code1.py")
 ```
 
-Or:
-
-```bash
-modelit create perceptron --output code1.py
-```
-
 Then run:
 
 ```bash
 python3 code1.py
 ```
 
-## Publish
+### CLI
 
-This project is set up for PyPI publishing with GitHub Actions.
+```bash
+modelit create perceptron
+modelit create perceptron --output code1.py
+```
 
-1. Push a tag like `v0.1.0`
-2. GitHub Actions builds and publishes to PyPI
-3. Users install with `pip install modelit`
+## Add a new code
 
-To update later, change the version, add your new templates, and publish a new tag.
+1. Create a folder:
 
-## Add a new code
+```text
+modelit/templates/mycode/
+```
 
-Create:
+2. Add your code file:
 
 ```text
-modelit/templates/<name>/template.py
-modelit/templates/<name>/metadata.json
+modelit/templates/mycode/template.py
 ```
 
-Then the new name is exposed automatically as:
+3. Done. The folder name becomes the function name.
+
+That means this will work automatically:
 
 ```python
-from modelit import <name>
+from modelit import mycode
+
+mycode()
 ```
+
+And this too:
+
+```bash
+modelit create mycode
+modelit create mycode --output mycode.py
+```
+
+## Publish flow
+
+1. Add a new template folder.
+2. `git add .`
+3. `git commit -m "..."`
+4. `git push`
+5. `git tag v0.x.y`
+6. `git push origin v0.x.y`
+
+The version is taken from the git tag, so you do not edit `pyproject.toml` for releases.

{modelit-0.1.0 → modelit-0.2.2}/modelit.egg-info/SOURCES.txt

@@ -1,7 +1,10 @@
+.gitignore
 LICENSE
 MANIFEST.in
 README.md
 pyproject.toml
+.github/workflows/publish.yml
+.github/workflows/test.yml
 modelit/__init__.py
 modelit/__main__.py
 modelit/cli.py
@@ -11,5 +14,5 @@ modelit.egg-info/SOURCES.txt
 modelit.egg-info/dependency_links.txt
 modelit.egg-info/entry_points.txt
 modelit.egg-info/top_level.txt
-modelit/templates/perceptron/metadata.json
+modelit/templates/backpropogation/template.py
 modelit/templates/perceptron/template.py

{modelit-0.1.0 → modelit-0.2.2}/pyproject.toml

@@ -1,10 +1,10 @@
 [build-system]
-requires = ["setuptools>=77", "wheel"]
+requires = ["setuptools>=77", "wheel", "setuptools-scm>=8"]
 build-backend = "setuptools.build_meta"
 
 [project]
 name = "modelit"
-version = "0.1.0"
+dynamic = ["version"]
 description = "Local-first ML starter templates you can print or save"
 readme = "README.md"
 requires-python = ">=3.10"
@@ -36,4 +36,8 @@ where = ["."]
 include = ["modelit*"]
 
 [tool.setuptools.package-data]
-modelit = ["templates/*/*.py", "templates/*/*.json"]
+modelit = ["templates/*/*.py"]
+
+[tool.setuptools_scm]
+version_scheme = "guess-next-dev"
+local_scheme = "no-local-version"

modelit-0.1.0/README.md

@@ -1,74 +0,0 @@
-# ModelIt
-
-Tiny local-first ML template printer.
-
-## Install
-
-```bash
-pip install modelit
-```
-
-For development from a clone, use:
-
-```bash
-pip install -e .
-```
-
-## Run
-
-```python
-from modelit import perceptron
-
-perceptron()
-```
-
-Or from CLI:
-
-```bash
-modelit create perceptron
-```
-
-## Save to file
-
-```python
-from modelit import perceptron
-
-perceptron(output="code1.py")
-```
-
-Or:
-
-```bash
-modelit create perceptron --output code1.py
-```
-
-Then run:
-
-```bash
-python3 code1.py
-```
-
-## Publish
-
-This project is set up for PyPI publishing with GitHub Actions.
-
-1. Push a tag like `v0.1.0`
-2. GitHub Actions builds and publishes to PyPI
-3. Users install with `pip install modelit`
-
-To update later, change the version, add your new templates, and publish a new tag.
-
-## Add a new code
-
-Create:
-
-```text
-modelit/templates/<name>/template.py
-modelit/templates/<name>/metadata.json
-```
-
-Then the new name is exposed automatically as:
-
-```python
-from modelit import <name>
-```

modelit-0.1.0/modelit/templates/perceptron/metadata.json

@@ -1,6 +0,0 @@
-{
-  "name": "perceptron",
-  "title": "Perceptron Learning Algorithm",
-  "output_file": "perceptron.py",
-  "description": "Print the full perceptron learning algorithm example."
-}