pip install finished

AD20/AD20.egg-info/PKG-INFO

@@ -0,0 +1,20 @@
+Metadata-Version: 2.1
+Name: AD20
+Version: 0.0.6
+Summary: Automatic Differentiation package
+Home-page: https://github.com/CS207-AD20/cs207-FinalProject
+Author: Lindsey Brown, Xinyue Wang, Kevin Yoon
+Author-email:  
+License: UNKNOWN
+Description: AD20 package by Group 20:
+        Lindsey Brown
+        Xinyue Wang
+        Kevin Yoon
+
+        For documentation, see https://github.com/CS207-AD20/cs207-FinalProject.
+
+Platform: UNKNOWN
+Classifier: Programming Language :: Python :: 3
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Description-Content-Type: text/markdown

AD20/AD20.egg-info/SOURCES.txt

@@ -0,0 +1,13 @@
+README.md
+setup.py
+AD20/ADgraph.py
+AD20/ADmath.py
+AD20/ADnum.py
+AD20/__init__.py
+AD20.egg-info/PKG-INFO
+AD20.egg-info/SOURCES.txt
+AD20.egg-info/dependency_links.txt
+AD20.egg-info/requires.txt
+AD20.egg-info/top_level.txt
+Tests/__init__.py
+Tests/test_AD20.py

AD20/AD20.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

AD20/AD20.egg-info/requires.txt

@@ -0,0 +1,10 @@
+numpy==1.15.2
+pandas==0.23.4
+networkx==2.2
+matplotlib==3.0.2
+pandastable==0.11.0
+scipy==1.1.0
+pytest-timeout==1.2.1
+pytest==3.4.2
+pytest-cov==2.5.1
+pytest-dependency==0.2

AD20/AD20.egg-info/top_level.txt

@@ -0,0 +1,2 @@
+AD20
+Tests

AD20/AD20/ADgraph.py

@@ -1,4 +1,6 @@
 import networkx as nx
+import matplotlib
+matplotlib.use('TkAgg')
 import matplotlib.pyplot as plt
 import matplotlib.patches as mpatches
 import numpy as np

AD20/build/lib/AD20/ADgraph.py

@@ -0,0 +1,228 @@
+import networkx as nx
+import matplotlib
+matplotlib.use('TkAgg')
+import matplotlib.pyplot as plt
+import matplotlib.patches as mpatches
+import numpy as np
+import pandas as pd
+from AD20.ADnum import ADnum
+
+
+def gen_graph(y):
+    """ Function to create a directed graph from an ADnum.
+
+    INPUTS
+    ======
+    y : ADnum
+
+    OUTPUTS
+    =======
+    A networkx digraph
+    """
+    G = nx.DiGraph()
+    d = y.graph
+    if len(d)== 0:
+        G.add_node(y)
+    for key in d:
+        G.add_node(key)
+        neighbors = d[key]
+        for neighbor in neighbors:
+            G.add_edge(key, neighbor[0], label = neighbor[1])
+    return G
+
+def reverse_graph(y):
+    """ Function to create a dictionary containing edges of y reversed.
+
+    INPUTS
+    ======
+    y : ADnum
+
+    OUTPUTS
+    =======
+    A dictionary
+    """
+    d = y.graph
+    parents = {}
+    for key in d:
+        neighbors = d[key]
+        for neighbor in neighbors:
+            if neighbor[0] not in parents:
+                parents[neighbor[0]] = []
+            parents[neighbor[0]].append((key, neighbor[1]))
+    return parents
+
+def get_labels(y):
+    """ Function to generate labels for plotting networkx graph.
+
+    INPUTS
+    ======
+    y : ADnum
+
+    OUTPUTS
+    =======
+    A dictionary of ADnum objects mapped to string labels
+    """
+    parents = reverse_graph(y)
+    total = len(y.graph) - sum([entry.constant for entry in y.graph.keys()])
+    new_names = {}
+    nodes = [y]
+    while len(nodes)>0:
+        node = nodes.pop()
+        if node not in new_names:
+            if node.constant:
+                new_names[node] = str(np.round(node.val, decimals=1))
+            else:
+                new_names[node] = 'X' + str(total)
+                total = total - 1
+            if node in parents:
+                neighbors = parents[node]
+                for neighbor in neighbors:
+                    nodes.append(neighbor[0])
+    return new_names
+
+def get_colors(G, y):
+    """ Function to assign colors to nodes in the graph.
+
+    INPUTS
+    ======
+    G : networkx digraph
+    y : ADnum
+
+    OUTPUTS
+    =======
+    A list of colors for the graph
+    """
+    colors = []
+    parents = reverse_graph(y)
+    for node in G:
+        if node.constant:
+            colors.append('blue')
+        else:
+            if node == y:
+                colors.append('green')
+            else:
+                if node in parents:
+                    colors.append('red')
+                else:
+                    colors.append('magenta')
+    return colors
+
+def get_sizes(G, y, labs):
+    """ Function to assign sizes to nodes in the graph.
+
+    INPUTS
+    ======
+    G : networkx digraph
+    y : ADnum
+    labs : dictionary of graph labels
+
+    OUTPUTS
+    =======
+    A list of sizes for the graph
+    """
+    sizes = []
+    for node in G:
+        label = labs[node]
+        sizes.append(len(label)*200)
+    return sizes
+
+def draw_graph(y):
+    """ Function to draw the graph.
+
+    INPUTS
+    ======
+    y : ADnum
+
+    OUTPUTS
+    =======
+    A plot of the graph
+    """  
+    fig = plt.figure()
+    G = gen_graph(y)
+    edge_labs = nx.get_edge_attributes(G, 'label')
+    pos = nx.spring_layout(G)
+    labs = get_labels(y)
+    nx.draw_networkx(G, pos, labels = labs, node_color = get_colors(G, y), node_size = get_sizes(G, y, labs), font_color= 'white')
+    nx.draw_networkx_edge_labels(G, pos, edge_labels = edge_labs)
+    limits = plt.axis('off')
+    mag_patch = mpatches.Patch(color = 'magenta', label = 'input')
+    red_patch = mpatches.Patch(color = 'red', label = 'intermediate')
+    blue_patch = mpatches.Patch(color = 'blue', label = 'constant')
+    green_patch = mpatches.Patch(color = 'green', label = 'output')
+    plt.legend(handles = [mag_patch, red_patch, blue_patch, green_patch])
+    plt.show()
+    return fig
+
+def gen_table(y):
+    """ Function to generate tables for the ADnum.
+
+    INPUTS
+    ======
+    y : ADnum
+
+    OUTPUTS
+    =======
+    A pandas data frame of the computational traces
+    """
+    parents = reverse_graph(y)
+    labs = get_labels(y)
+    visited = []
+    data = {}
+    data['Trace'] = []
+    data['Operation']=[]
+    data['Value']= []
+    data['Derivative']=[]
+    nodes = [y]
+    while len(nodes)>0:
+        node = nodes.pop()
+        if node not in visited:
+            if node.constant:
+                visited.append(node)
+            else:
+                visited.append(node)
+                data['Trace'].append(labs[node])
+                data['Value'].append(node.val)
+                data['Derivative'].append(node.der)
+                if node in parents:
+                    if len(parents[node]) == 1:
+                        link = parents[node][0][1]+'('+labs[parents[node][0][0]]+')'
+                    else:
+                        link = parents[node][0][1]+'(' +labs[parents[node][0][0]]+ ' , ' + labs[parents[node][1][0]] + ')'
+                    neighbors = parents[node]
+                    for neighbor in neighbors:
+                        nodes.append(neighbor[0])
+                else:
+                    link = 'input'
+                data['Operation'].append(link)
+    result = pd.DataFrame.from_dict(data)
+    result2 = result.sort_values('Trace')
+    resultorder = result2[['Trace', 'Operation', 'Value', 'Derivative']]  
+    return resultorder
+
+def plot_ADnum(x, xmin = -10, xmax = 10):
+    '''Function to plot f and its derivative for single variable input
+
+    INPUTS
+    ======
+    x : ADnum
+    xmin : starting value of input
+    xmax : ending value of input
+
+    OUTPUTS
+    =======
+    A plot of x evaluated from xmin to xmax and its derivative
+    '''
+    vals = np.linspace(xmin, xmax, 100)
+    evals = [x(ADnum(value, der=1)).val for value in vals]
+    ders = [x(ADnum(value, der=1)).der for value in vals]
+    fig = plt.figure()
+    plt.plot(vals, evals, label = 'f', linewidth = 2)
+    plt.plot(vals, ders, label = 'df/dx', linewidth = 2)
+    plt.legend(fontsize = 20)
+    plt.xlabel('x', fontsize = 20)
+    plt.ylabel('f', fontsize = 20)
+    plt.xticks(fontsize = 12)
+    plt.yticks(fontsize = 12)
+    return fig
+
+