aslan-ng commited on
Commit
f7ee56e
·
verified ·
1 Parent(s): ed428ee

Create data.py

Browse files
Files changed (1) hide show
  1. data.py +40 -0
data.py ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Imports
2
+
3
+ from datasets import load_dataset
4
+
5
+ from settings import HF_TOKEN
6
+
7
+
8
+ # Load Data
9
+
10
+ NUMERIC_PROFILE = ["Laser Cutting", "Wood Working", "Wood CNC", "Metal Machining", "Metal CNC", "3D Printer", "Welding", "Electronics"]
11
+
12
+ login(HF_TOKEN)
13
+
14
+ def load_data_from_huggingface():
15
+ """
16
+ Loads data from HuggingFace.
17
+ """
18
+ # Staff (People)
19
+ ds_staff = load_dataset(REPO_ID_TECHSPARK_STAFF)
20
+ staff_df = ds_staff["train"].to_pandas()
21
+
22
+ # Courses
23
+ ds_courses = load_dataset(REPO_ID_TECHSPARK_COURSES)
24
+ courses_df = ds_courses["train"].to_pandas()
25
+
26
+ # Tools
27
+ ds_tools = load_dataset(REPO_ID_TECHSPARK_TOOLS)
28
+ tools_df = ds_tools["train"].to_pandas()
29
+
30
+ # Map Nodes
31
+ ds_nodes = load_dataset(REPO_ID_TECHSPARK_MAP_NODES)
32
+ nodes_df = ds_nodes["train"].to_pandas()
33
+
34
+ # Map Edges
35
+ ds_edges = load_dataset(REPO_ID_TECHSPARK_MAP_EDGES)
36
+ edges_df = ds_edges["train"].to_pandas()
37
+
38
+ return staff_df, courses_df, tools_df, nodes_df, edges_df
39
+
40
+ staff_df, courses_df, tools_df, nodes_df, edges_df = load_data_from_huggingface()