analysis

analysis_prod_network.py

@@ -0,0 +1,176 @@
+import pandas as pd
+import numpy as np
+import matplotlib.pyplot as plt
+import networkx as nx
+import math
+
+plt.rcParams['font.sans-serif'] = 'SimHei'
+
+count_prod = pd.read_csv("analysis\\count_prod.csv")
+print(count_prod)
+
+# category
+print(count_prod.describe())
+
+# pie
+count_prod_trim = count_prod[count_prod['count'] > 50]
+plt.pie(count_prod_trim['count'], labels=count_prod_trim['Name'])
+plt.savefig("analysis\\count_prod_pie")
+plt.close()
+
+# prod_networkx
+BomNodes = pd.read_csv('BomNodes.csv', index_col=0)
+BomNodes.set_index('Code', inplace=True)
+BomCateNet = pd.read_csv('BomCateNet.csv', index_col=0)
+BomCateNet.fillna(0, inplace=True)
+
+G = nx.from_pandas_adjacency(BomCateNet.T, create_using=nx.MultiDiGraph())
+
+labels_dict = {}
+for code in G.nodes:
+    node_attr = BomNodes.loc[code].to_dict()
+    index_list = count_prod[count_prod['id_product'] == code].index.tolist()
+    index = index_list[0] if len(index_list) == 1 else -1
+    node_attr['count'] = count_prod['count'].get(index, 0)
+    node_attr['node_size'] = 5 * count_prod['count'].get(index, 0)
+    node_attr['node_color'] = count_prod['count'].get(index, 0)
+    labels_dict[code] = node_attr
+nx.set_node_attributes(G, labels_dict)
+# print(labels_dict)
+
+pos = nx.nx_agraph.graphviz_layout(G, prog="twopi", args="")
+dict_node_name = nx.get_node_attributes(G, 'Name')
+node_labels = {}
+for node in nx.nodes(G):
+    node_labels[node] = f"{node} {str(dict_node_name[node])}"
+    # node_labels[node] = f"{str(dict_node_name[node])}"
+colors = list(nx.get_node_attributes(G, 'node_color').values())
+vmin = min(colors)
+vmax = max(colors)
+cmap = plt.cm.Blues
+fig = plt.figure(figsize=(10, 10), dpi=300)
+nx.draw(G,
+        pos,
+        node_size=list(nx.get_node_attributes(G, 'node_size').values()),
+        labels=node_labels,
+        font_size=6,
+        node_color=colors,
+        cmap=cmap,
+        vmin=vmin,
+        vmax=vmax,
+        edge_color='grey')
+sm = plt.cm.ScalarMappable(cmap=cmap, norm=plt.Normalize(vmin=vmin, vmax=vmax))
+sm._A = []
+position = fig.add_axes([0.01, 0.05, 0.01, 0.3])
+plt.colorbar(sm, fraction=0.01, cax=position)
+# plt.savefig("analysis\\count_prod_network")
+plt.close()
+
+# dcp_prod
+count_dcp = pd.read_csv("analysis\\count_dcp.csv",
+                        dtype={
+                            'up_id_firm': str,
+                            'down_id_firm': str
+                        })
+count_dcp_prod = count_dcp.groupby(['up_id_product','up_name_product', 'down_id_product', 'down_name_product'])['count'].sum()
+count_dcp_prod = count_dcp_prod.reset_index()
+count_dcp_prod.sort_values('count', inplace=True, ascending=False)
+count_dcp_prod.to_csv('analysis\\count_dcp_prod.csv',
+                       index=False,
+                       encoding='utf-8-sig')
+count_dcp_prod = count_dcp_prod[count_dcp_prod['count'] > 2]
+# print(count_dcp_prod)
+
+list_prod = count_dcp_prod['up_id_product'].tolist(
+) + count_dcp['down_id_product'].tolist()
+list_prod = list(set(list_prod))
+
+# init graph bom
+
+BomNodes = pd.read_csv('BomNodes.csv', index_col=0)
+BomNodes.set_index('Code', inplace=True)
+
+g_bom = nx.MultiDiGraph()
+g_bom.add_nodes_from(list_prod)
+
+bom_labels_dict = {}
+for code in list_prod:
+    dct_attr = BomNodes.loc[code].to_dict()
+    bom_labels_dict[code] = dct_attr
+nx.set_node_attributes(g_bom, bom_labels_dict)
+
+
+count_max = count_dcp_prod['count'].max()
+count_min = count_dcp_prod['count'].min()
+k = 5 / (count_max - count_min)
+for _, row in count_dcp_prod.iterrows():
+    # print(row)
+    lst_add_edge = [(
+        row['up_id_product'],
+        row['down_id_product'],
+        {
+            'count': row['count']
+        })]
+    g_bom.add_edges_from(lst_add_edge)
+
+# dcp_networkx
+pos = nx.nx_agraph.graphviz_layout(g_bom, prog="dot", args="")
+node_labels = nx.get_node_attributes(g_bom, 'Name')
+temp = {}
+for key, value in node_labels.items():
+    temp[key] = key + " " + value
+node_labels = temp
+colors = nx.get_edge_attributes(g_bom, "count")
+colors = [w for (n1, n2, _), w in colors.items()]
+vmin = min(colors)
+vmax = max(colors)
+cmap = plt.cm.Blues
+
+# dct_row = {}
+# for node, p in pos.items():
+#     if p[1] not in dct_row.keys():
+#         dct_row[p[1]] = {node: p}
+#     else:
+#         dct_row[p[1]][node] = p
+# dct_row = dict(sorted(dct_row.items(), key=lambda d: d[0], reverse=True))
+# dct_up = dct_row[max(dct_row.keys())]
+# dct_up = dict(sorted(dct_up.items(), key=lambda d: d[1][0], reverse=True))
+# h = list(dct_row.keys())[0] - list(dct_row.keys())[1]
+# n = len(dct_up.items())
+# arr_h = np.linspace(list(dct_row.keys())[0]-h/2, list(dct_row.keys())[0]+2*h, num=n)
+# dct_up_new = {}
+# for index, (node, p) in enumerate(dct_up.items()):
+#     dct_up_new[node] = (p[0], arr_h[index])
+# pos_new = {}
+# for row, dct in dct_row.items():
+#     if row == list(dct_row.keys())[0]:
+#         pos_new.update(dct_up_new)
+#     else:
+#         pos_new.update(dct)
+pos_new ={}
+for node, p in pos.items():
+    pos_new[node] = (p[1], p[0])
+
+fig = plt.figure(figsize=(6, 10), dpi=300)
+# plt.subplots_adjust(right=0.7)
+nx.draw(g_bom,
+        pos_new,
+        node_size=50,
+        labels=node_labels,
+        font_size=6,
+        width = 1.5,
+        edge_color=colors,
+        edge_cmap=cmap,
+        edge_vmin=vmin,
+        edge_vmax=vmax)
+plt.axis('off')
+axis = plt.gca()
+axis.set_xlim([1.2*x for x in axis.get_xlim()])
+axis.set_ylim([1.2*y for y in axis.get_ylim()])
+
+sm = plt.cm.ScalarMappable(cmap=cmap, norm=plt.Normalize(vmin=vmin, vmax=vmax))
+sm._A = []
+position=fig.add_axes([0.1, 0.4, 0.01, 0.2])
+plt.colorbar(sm, fraction=0.01, cax=position)
+# plt.savefig("analysis\\count_dcp_prod_network")
+plt.close()