Updated Directory Structure

SANDBOX/meta_exploit.py

@@ -1,139 +1,139 @@
-import numpy as np
-import vish_graphs as vg
-import pandas as pd
-import torch
-import torch.nn as nn
-import torch.optim as optim
-from torch.utils.data import Dataset, DataLoader
-import core_rec as cs
-import matplotlib.pyplot as plt
-from matplotlib.table import Table
-
-
-# Load the CSV file into a DataFrame
-adj_matrix = np.loadtxt('SANDBOX/adj.csv', delimiter=",")
-# wgt_matrix = np.loadtxt('SANDBOX/label.csv', delimiter=",")
-
-wgt_matrix = []
-for i in range(1, 11):
-    weight_matrix1 = np.loadtxt(f'SANDBOX/delete/label_{i}.csv', delimiter=",")
-    wgt_matrix.append(weight_matrix1)
-
-
-# Load node labels
-df = pd.read_csv("SANDBOX/labelele.csv")
-col = df.values.flatten()
-node_labels = {i: label for i, label in enumerate(col)}
-
-# Find top nodes
-top_nodes = vg.find_top_nodes(adj_matrix, 4)
-
-# ML
-# Convert adjacency matrix to dataset
-graph_dataset = cs.GraphDataset(adj_matrix)
-batch_size=3
-var=1.0
-
-#picked form here  
-
-
-# Predict recommendations for a specific node
-# node_index =4    # target node
-# recommended_nodes = cs.predict(model, adj_matrix, node_index, top_k=1, threshold=0.5)
-
-# # Convert recommended node indices to labels
-# recommended_labels = [node_labels[node] for node in recommended_nodes]
-# print(f"Recommended nodes for node {node_labels[node_index]}: {recommended_labels}")
-# torch.save(model.state_dict(), 'massive_surya_e10k.pth')
-
-# print("\n")
-# if Warning:
-#     print(f"Actually Multi-head Attn was not properly being distribuited amongst {num_heads} heads, it was being distribuited amongst {num_heads-delta} heads")
-#     print(f"Hence the embedding dimension was reduced by {delta} to make it divisible by the number of heads")
-        # Define model parameters
-num_layers = 1
-d_model = 128 #embedding dimension
-num_heads=2
-
-        # decoration for warning 
-delta=0
-Warning = False
-if(d_model % num_heads != 0):
-    Warning = True
-    delta = d_model % num_heads
-    d_model = d_model - delta
-else:
-    pass
-
-
-d_feedforward = 512
-input_dim = len(adj_matrix[0])
-num_weights = 10
-# Initialize model, loss function, and optimizer
-model = cs.GraphTransformer(num_layers, d_model, num_heads, d_feedforward, input_dim, num_weights)
-# model = cs.GraphTransformer(num_layers, d_model, num_heads, d_feedforward, input_dim, use_weights=True)
-criterion = nn.MSELoss()
-optimizer = optim.Adam(model.parameters(), lr=0.0001)
-top_nodes = vg.find_top_nodes(adj_matrix, num_nodes=5)
-
-# Train the model
-num_epochs = 100   
-batch_idx=[1,2,3,5,8,10]
-threshold_idx=[0.5,0.7,0.9,1.0]
-rl=['Vishesh', 'Shrestha', 'Biswajeet', 'Priyanka', 'Poonam', 'Adhiraaj', 'Yash', 'Sachin', 'Vinayak', 'Kranti', 'Sai']
-for m in range(11):
-    print(f"Recommending for : {rl[m]} ")
-    for i in range(len(threshold_idx)):
-        print(f"{threshold_idx[i]} : ")
-        for k in range(len(batch_idx)):
-            data_loader = DataLoader(graph_dataset, batch_size=batch_idx[k], shuffle=True)
-            cs.train_model(model, data_loader, criterion, optimizer, num_epochs)
-            node_index =m   # target node
-            recommended_nodes = cs.predict(model, adj_matrix, node_index, top_k=3, threshold=threshold_idx[i])
-            # Convert recommended node indices to labels
-            recommended_labels = [node_labels[node] for node in recommended_nodes]
-            # print(f"Recommended nodes for node {node_labels[node_index]}: {recommended_labels}")
-            # print(f"Recommended nodes for vishesh of thr {threshold_idx[i]} and batchsize {batch_idx[k]} : {recommended_labels}")
-            print(f" {batch_idx[k]} : {recommended_labels}")
-
-
-# Load node labels from the CSV file
-df_labels = pd.read_csv("SANDBOX/labelele.csv")
-node_labels = df_labels["Names"].tolist()
-
-
-# Assuming you have the recommended nodes for each node index stored in a list of lists
-recommended_nodes_list = []
-for i in range(11):
-    recommended_nodes = cs.predict(model, adj_matrix, 0, top_k=3, threshold=i)
-    recommended_nodes_list.append(recommended_nodes)
-
-
-
-
-
-# # Create a scatter plot with labels for node indexes
-# plt.figure(figsize=(10, 8))
+# import numpy as np
+# import vish_graphs as vg
+# import pandas as pd
+# import torch
+# import torch.nn as nn
+# import torch.optim as optim
+# from torch.utils.data import Dataset, DataLoader
+# import core_rec as cs
+# import matplotlib.pyplot as plt
+# from matplotlib.table import Table
+
+
+# # Load the CSV file into a DataFrame
+# adj_matrix = np.loadtxt('SANDBOX/adj.csv', delimiter=",")
+# # wgt_matrix = np.loadtxt('SANDBOX/label.csv', delimiter=",")
+
+# wgt_matrix = []
+# for i in range(1, 11):
+#     weight_matrix1 = np.loadtxt(f'SANDBOX/delete/label_{i}.csv', delimiter=",")
+#     wgt_matrix.append(weight_matrix1)
+
+
+# # Load node labels
+# df = pd.read_csv("SANDBOX/labelele.csv")
+# col = df.values.flatten()
+# node_labels = {i: label for i, label in enumerate(col)}
+
+# # Find top nodes
+# top_nodes = vg.find_top_nodes(adj_matrix, 4)
+
+# # ML
+# # Convert adjacency matrix to dataset
+# graph_dataset = cs.GraphDataset(adj_matrix)
+# batch_size=3
+# var=1.0
+
+# #picked form here  
+
+
+# # Predict recommendations for a specific node
+# # node_index =4    # target node
+# # recommended_nodes = cs.predict(model, adj_matrix, node_index, top_k=1, threshold=0.5)
+
+# # # Convert recommended node indices to labels
+# # recommended_labels = [node_labels[node] for node in recommended_nodes]
+# # print(f"Recommended nodes for node {node_labels[node_index]}: {recommended_labels}")
+# # torch.save(model.state_dict(), 'massive_surya_e10k.pth')
+
+# # print("\n")
+# # if Warning:
+# #     print(f"Actually Multi-head Attn was not properly being distribuited amongst {num_heads} heads, it was being distribuited amongst {num_heads-delta} heads")
+# #     print(f"Hence the embedding dimension was reduced by {delta} to make it divisible by the number of heads")
+#         # Define model parameters
+# num_layers = 1
+# d_model = 128 #embedding dimension
+# num_heads=2
+
+#         # decoration for warning 
+# delta=0
+# Warning = False
+# if(d_model % num_heads != 0):
+#     Warning = True
+#     delta = d_model % num_heads
+#     d_model = d_model - delta
+# else:
+#     pass
+
+
+# d_feedforward = 512
+# input_dim = len(adj_matrix[0])
+# num_weights = 10
+# # Initialize model, loss function, and optimizer
+# model = cs.GraphTransformer(num_layers, d_model, num_heads, d_feedforward, input_dim, num_weights)
+# # model = cs.GraphTransformer(num_layers, d_model, num_heads, d_feedforward, input_dim, use_weights=True)
+# criterion = nn.MSELoss()
+# optimizer = optim.Adam(model.parameters(), lr=0.0001)
+# top_nodes = vg.find_top_nodes(adj_matrix, num_nodes=5)
+
+# # Train the model
+# num_epochs = 100   
+# batch_idx=[1,2,3,5,8,10]
+# threshold_idx=[0.5,0.7,0.9,1.0]
+# rl=['Vishesh', 'Shrestha', 'Biswajeet', 'Priyanka', 'Poonam', 'Adhiraaj', 'Yash', 'Sachin', 'Vinayak', 'Kranti', 'Sai']
+# for m in range(11):
+#     print(f"Recommending for : {rl[m]} ")
+#     for i in range(len(threshold_idx)):
+#         print(f"{threshold_idx[i]} : ")
+#         for k in range(len(batch_idx)):
+#             data_loader = DataLoader(graph_dataset, batch_size=batch_idx[k], shuffle=True)
+#             cs.train_model(model, data_loader, criterion, optimizer, num_epochs)
+#             node_index =m   # target node
+#             recommended_nodes = cs.predict(model, adj_matrix, node_index, top_k=3, threshold=threshold_idx[i])
+#             # Convert recommended node indices to labels
+#             recommended_labels = [node_labels[node] for node in recommended_nodes]
+#             # print(f"Recommended nodes for node {node_labels[node_index]}: {recommended_labels}")
+#             # print(f"Recommended nodes for vishesh of thr {threshold_idx[i]} and batchsize {batch_idx[k]} : {recommended_labels}")
+#             print(f" {batch_idx[k]} : {recommended_labels}")
+
+
+# # Load node labels from the CSV file
+# df_labels = pd.read_csv("SANDBOX/labelele.csv")
+# node_labels = df_labels["Names"].tolist()
+
+
+# # Assuming you have the recommended nodes for each node index stored in a list of lists
+# recommended_nodes_list = []
 # for i in range(11):
-#     x = [i] * len(recommended_nodes_list[i])
-#     y = recommended_nodes_list[i]
-#     plt.scatter(x, y, label=f"Node {i}")
-#     for j, txt in enumerate(y):
-#         plt.annotate(node_labels[y[j]], (x[j], y[j]), textcoords="offset points", xytext=(0,10), ha='center')
-
-# plt.xticks(range(11), node_labels, rotation=45)
-# plt.yticks(range(11), node_labels)
-# plt.xlabel('Node Index / Names')
-# plt.ylabel('Recommended Node Index / Names')
-# plt.title('Recommendations for Node Index')
-
-# # Create a legend table with batch size and threshold details
-# legend_data = [['Batch Size', 'Threshold'],
-#                [batch_size , var]]
-# table = plt.table(cellText=legend_data, loc='upper left', cellLoc='center', colWidths=[0.1, 0.1])
-# table.auto_set_font_size(False)
-# table.set_fontsize(12)
-# table.scale(1.5, 1.5)
-# 
-# plt.grid(True)
-# plt.show()
+#     recommended_nodes = cs.predict(model, adj_matrix, 0, top_k=3, threshold=i)
+#     recommended_nodes_list.append(recommended_nodes)
+
+
+
+
+
+# # # Create a scatter plot with labels for node indexes
+# # plt.figure(figsize=(10, 8))
+# # for i in range(11):
+# #     x = [i] * len(recommended_nodes_list[i])
+# #     y = recommended_nodes_list[i]
+# #     plt.scatter(x, y, label=f"Node {i}")
+# #     for j, txt in enumerate(y):
+# #         plt.annotate(node_labels[y[j]], (x[j], y[j]), textcoords="offset points", xytext=(0,10), ha='center')
+
+# # plt.xticks(range(11), node_labels, rotation=45)
+# # plt.yticks(range(11), node_labels)
+# # plt.xlabel('Node Index / Names')
+# # plt.ylabel('Recommended Node Index / Names')
+# # plt.title('Recommendations for Node Index')
+
+# # # Create a legend table with batch size and threshold details
+# # legend_data = [['Batch Size', 'Threshold'],
+# #                [batch_size , var]]
+# # table = plt.table(cellText=legend_data, loc='upper left', cellLoc='center', colWidths=[0.1, 0.1])
+# # table.auto_set_font_size(False)
+# # table.set_fontsize(12)
+# # table.scale(1.5, 1.5)
+# # 
+# # plt.grid(True)
+# # plt.show()

SANDBOX/test.py

@@ -16,7 +16,7 @@
 import numpy as np
 # import vishgraph as vg
 import core_rec as cs
-
+import vish_graphs as vg
 # for trainig
 import torch
 import torch.nn as nn