scypyy 0.1.0__tar.gz

scypyy-0.1.0/PKG-INFO

@@ -0,0 +1,6 @@
+Metadata-Version: 2.4
+Name: scypyy
+Version: 0.1.0
+Classifier: Programming Language :: Python :: 3
+Classifier: Operating System :: OS Independent
+Requires-Python: >=3.9

scypyy-0.1.0/pyproject.toml

@@ -0,0 +1,16 @@
+[build-system]
+requires = ["setuptools>=61.0", "wheel"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "scypyy"
+version = "0.1.0"
+description = ""
+requires-python = ">=3.9"
+classifiers = [
+    "Programming Language :: Python :: 3",
+    "Operating System :: OS Independent",
+]
+
+[tool.setuptools]
+packages = ["scypyy"]

scypyy-0.1.0/scypyy/__init__.py

@@ -0,0 +1,69 @@
+"""scypyy — code snippets stored inside the library, accessed as attributes.
+
+All snippet code is embedded directly in this package (see ``_data.py``).
+Nothing is read from disk at runtime. Access a snippet by name::
+
+    import scypyy
+    scypyy.bwp          # shows the stored code in the REPL
+    scypyy.nn_final     # another snippet
+    print(scypyy.bwp)   # prints the code
+    code = scypyy.bwp   # a normal str you can save or exec
+
+Use ``scypyy.available()`` to list every snippet name.
+"""
+
+from __future__ import annotations
+
+from ._data import SNIPPETS
+
+__version__ = "0.1.0"
+
+
+class Code(str):
+    """A ``str`` subclass that displays its raw contents in the REPL.
+
+    A normal string echoes with escaped newlines (``\\n``). This shows the
+    actual code instead, while still behaving like an ordinary string.
+    """
+
+    __slots__ = ()
+
+    def __repr__(self) -> str:  # noqa: D105
+        return str.__str__(self)
+
+
+def available():
+    """Return a sorted list of every stored snippet name."""
+    return sorted(SNIPPETS)
+
+
+def get(name: str) -> Code:
+    """Return the stored code for ``name`` (case-insensitive) as a :class:`Code`."""
+    if name in SNIPPETS:
+        return Code(SNIPPETS[name])
+    for key in SNIPPETS:
+        if key.lower() == name.lower():
+            return Code(SNIPPETS[key])
+    raise KeyError(
+        f"no snippet named {name!r}. Available: {', '.join(available()) or '(none)'}"
+    )
+
+
+def __getattr__(name: str) -> Code:
+    """Resolve ``scypyy.<name>`` to the stored code for ``<name>`` (PEP 562)."""
+    if name.startswith("_"):
+        raise AttributeError(f"module {__name__!r} has no attribute {name!r}")
+    try:
+        return get(name)
+    except KeyError:
+        raise AttributeError(
+            f"module {__name__!r} has no attribute {name!r}. "
+            f"Available snippets: {', '.join(available()) or '(none)'}"
+        ) from None
+
+
+def __dir__():
+    return sorted(set(globals()) | set(SNIPPETS))
+
+
+__all__ = ["available", "get", "Code"]

scypyy-0.1.0/scypyy/_data.py

@@ -0,0 +1,11 @@
+# Auto-generated: code snippets embedded as string constants.
+# Do not read any external files at runtime — everything lives here.
+
+SNIPPETS = {
+    'LGD': '# x1,x2 are the input features\ndef LinearGradientDescent2(x1, x2, x3, y, lr, epochs):\n    w0 = 0.25\n    w1 = 0.65\n    w2 = 0.65\n    w3 = 0.65\n\n    m = float(len(y))\n    cost_hist = []\n\n    for i in range(epochs):\n        h = w0 + w1*x1 + w2*x2 + w3*x3\n        cost = (1/(2*m)) * sum((h - y)**2)\n\n        dw0 = (1/m) * sum(h - y)\n        dw1 = (1/m) * sum((h - y) * x1)\n        dw2 = (1/m) * sum((h - y) * x2)\n        dw3 = (1/m) * sum((h - y) * x3)\n\n        if i % 100 == 0:\n            print(f"Cost: {cost} \\tIteration: {i}")\n\n        w0 = w0 - lr * dw0\n        w1 = w1 - lr * dw1\n        w2 = w2 - lr * dw2\n        w3 = w3 - lr * dw3\n\n        cost_hist.append(cost)\n\n        if len(cost_hist) > 1 and abs(cost_hist[-1] - cost_hist[-2]) < 1e-6:\n            print(f"Converged at iteration {i}")\n            break\n\n    return w0, w1, w2, w3',
+    'activationfunctions': 'def tanH(X):\n    return (np.exp(X)-np.exp(-X))/(np.exp(X)+np.exp(-X))\n\ndef ReLU(X):\n    return np.maximum(0, X)\n\ndef sigmoid(Z):\n    s = 1/(1+np.exp(-Z))\n    return s\n\ndef sigmoid_grad(y):\n    return y * (1 - y)\n\ndef tanH_grad(y):\n    return 1 - y**2\n\ndef ReLU_grad(x):\n    return np.where(x > 0, 1, 0)\n\ndef eReLU(x, alpha=0.01):\n    return np.where(x > 0, x, alpha * (np.exp(x) - 1))\n \ndef eReLU_grad(x, alpha=0.01):\n    return np.where(x > 0, 1, alpha * np.exp(x))',
+    'bwp': 'def BWP(par, cat, X, y):\n\n    w1 = par["w1"]\n    b1 = par["b1"]\n    w2 = par["w2"]\n    b2 = par["b2"]\n    w3 = par["w3"]\n    b3 = par["b3"]\n    w4 = par["w4"]\n    b4 = par["b4"]\n    w5 = par["w5"]\n    b5 = par["b5"]\n    a1 = cat["a1"]\n    a2 = cat["a2"]\n    a3 = cat["a3"]\n    a4 = cat["a4"]\n    a5 = cat["a5"]\n\n    m = y.shape[1]\n    dz5 = a5 - y\n    dw5 = np.dot(dz5, a4.T) / m\n    db5 = np.sum(dz5, axis=1, keepdims=True) / m\n    da4 = np.dot(w5.T, dz5)\n    dz4 = da4 * ReLU_grad(a4)\n    dw4 = np.dot(dz4, a3.T) / m\n    db4 = np.sum(dz4, axis=1, keepdims=True) / m\n    da3 = np.dot(w4.T, dz4)\n    dz3 = da3 * eReLU_grad(a3)\n    dw3 = np.dot(dz3, a2.T) / m\n    db3 = np.sum(dz3, axis=1, keepdims=True) / m\n    da2 = np.dot(w3.T, dz3)\n    dz2 = da2 * ReLU_grad(a2)\n    dw2 = np.dot(dz2, a1.T) / m\n    db2 = np.sum(dz2, axis=1, keepdims=True) / m\n    da1 = np.dot(w2.T, dz2)\n    dz1 = da1 * ReLU_grad(a1)\n    dw1 = np.dot(dz1, X.T) / m\n    db1 = np.sum(dz1, axis=1, keepdims=True) / m\n\n    grades = {"dw1": dw1, "db1": db1, "dw2": dw2, "db2": db2, "dw3": dw3, "db3": db3, "dw4": dw4, "db4": db4, "dw5": dw5, "db5": db5}\n    return grades\n\ngrades = BWP(par,cache,x,y)\n\n\ndef update(par,grades,learning_rate = 0.01):\n    w1 = par[\'w1\']\n    b1 = par[\'b1\']\n    w2 = par[\'w2\']\n    b2 = par[\'b2\']\n    w3 = par[\'w3\']\n    b3 = par[\'b3\']\n    w4 = par[\'w4\']\n    b4 = par[\'b4\']\n    w5 = par[\'w5\']\n    b5 = par[\'b5\']\n    dw1 = grades[\'dw1\']\n    dw2 = grades[\'dw2\']\n    dw3 = grades[\'dw3\']\n    dw4 = grades[\'dw4\']\n    dw5 = grades[\'dw5\']\n    db1 = grades[\'db1\']\n    db2 = grades[\'db2\']\n    db3 = grades[\'db3\']\n    db4 = grades[\'db4\']\n    db5 = grades[\'db5\']\n    w1 = w1 - learning_rate*dw1\n    w2 = w2 - learning_rate*dw2\n    w3 = w3 - learning_rate*dw3\n    w4 = w4 - learning_rate*dw4\n    w5 = w5 - learning_rate*dw5\n    b1 = b1 - learning_rate*db1\n    b2 = b2 - learning_rate*db2\n    b3 = b3 - learning_rate*db3\n    b4 = b4 - learning_rate*db4\n    b5 = b5 - learning_rate*db5\n    par = {"w1":w1,"b1":b1,"w2":w2,"b2":b2,"w3":w3,"b3":b3,"w4":w4,"b4":b4,"w5":w5,"b5":b5}\n    return par\n\nupdate(par,grades)',
+    'fwdp': 'def FWDP(X, par):\n    w1 = par["w1"]\n    b1 = par["b1"]\n    w2 = par["w2"]\n    b2 = par["b2"]\n    w3 = par["w3"]\n    b3 = par["b3"]\n    w4 = par["w4"]\n    b4 = par["b4"]\n    w5 = par["w5"]\n    b5 = par["b5"]\n\n    z1 = np.dot(w1, X) + b1\n    a1 = ReLU(z1)\n    z2 = np.dot(w2, a1) + b2\n    a2 = ReLU(z2)\n    z3 = np.dot(w3, a2) + b3\n    a3 = eReLU(z3)\n    z4 = np.dot(w4, a3) + b4\n    a4 = ReLU(z4)\n    z5 = np.dot(w5, a4) + b5\n    a5 = sigmoid(z5)\n\n    cache = {\n        "z1": z1, "a1": a1,\n        "z2": z2, "a2": a2,\n        "z3": z3, "a3": a3,\n        "z4": z4, "a4": a4,\n        "z5": z5, "a5": a5\n    }\n\n    return a5, cache\n\na5, cache = FWDP(x,par)\nprint(a5)\nprint(cache)\n\ndef compute_cost(a5,y):\n    m = y.shape[1]\n    cost = np.sum(y*np.log(a5)+(1-y)*np.log(1-a5))\n    cost = -cost/m\n    return cost\ncompute_cost(a5,y)',
+    'nn_final': 'def NN(x,y,n_h1,n_h2,n_h3,n_h4,epoch = 10000,print_cost = True):\n    np.random.seed(3)\n    n_x,n_h1,n_h2,n_h3,n_h4, n_y = nodes(x,y,n_h1,n_h2,n_h3,n_h4)\n    par = int_par(n_x,n_h1,n_h2,n_h3,n_h4,n_y)\n    cost_hist=[]\n\n    for i in range(epoch):\n        a4,cat = FWDP(x,par)\n        cost = compute_cost(a5,y)\n        grades = BWP(par,cache,x,y)\n        par = update(par,grades)\n        if print_cost and i%1000 == 0:\n            print(f"cost {cost} \\t iteration: {i}")\n\n        cost_hist.append(cost)\n\n\n    return par\n\n\nNN(x,y,n_h1,n_h2,n_h3,n_h4,epoch = 10000,print_cost = True)',
+    'nnintro': 'import numpy as np\nnp.random.seed(123)\nx = np.random.randn(20,100)    #(nx,m)\ny = np.random.randn(1,100)>0  #(1,m)\nx\ny\n\ndef nodes(x,y,n_h1,n_h2,n_h3,n_h4):\n    n_x = x.shape[0]\n    n_h1 = n_h1\n    n_h2 = n_h2\n    n_h3 = n_h3\n    n_h4 = n_h4\n    n_y = y.shape[0]\n    return n_x,n_h1,n_h2,n_h3,n_h4,n_y\n\n    n_x,n_h1,n_h2,n_h3,n_h4,n_y = nodes(x,y,n1,n2,n3,n4)\n\n    def int_par(n_x,n_h1,n_h2,n_h3,n_h4,n_y):\n    np.random.seed(2)\n    w1 = np.random.randn(n_h1,n_x)*0.01\n    b1 = np.random.randn(n_h1,1)\n    w2 = np.random.randn(n_h2,n_h1)*0.01\n    b2 = np.random.randn(n_h2,1)\n    w3 = np.random.randn(n_h3,n_h2)*0.01\n    b3 = np.random.randn(n_h3,1)\n    w4 = np.random.randn(n_h4,n_h3)*0.01\n    b4 = np.random.randn(n_h4,1)\n    w5 = np.random.randn(n_y,n_h4)*0.01\n    b5 = np.random.randn(n_y,1)\n    par = {"w1":w1,"b1":b1,"w2":w2,"b2":b2,"w3":w3,"b3":b3,"w4":w4,"b4":b4,"w5":w5, "b5":b5}\n    return par\n\n    par = int_par(n_x,n_h1,n_h2,n_h3,n_h4,n_y)',
+}

scypyy-0.1.0/scypyy.egg-info/PKG-INFO

@@ -0,0 +1,6 @@
+Metadata-Version: 2.4
+Name: scypyy
+Version: 0.1.0
+Classifier: Programming Language :: Python :: 3
+Classifier: Operating System :: OS Independent
+Requires-Python: >=3.9

scypyy-0.1.0/scypyy.egg-info/SOURCES.txt

@@ -0,0 +1,7 @@
+pyproject.toml
+scypyy/__init__.py
+scypyy/_data.py
+scypyy.egg-info/PKG-INFO
+scypyy.egg-info/SOURCES.txt
+scypyy.egg-info/dependency_links.txt
+scypyy.egg-info/top_level.txt

scypyy-0.1.0/scypyy.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

scypyy-0.1.0/scypyy.egg-info/top_level.txt

@@ -0,0 +1 @@
+scypyy

scypyy-0.1.0/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+