FSE-2025-Understanding-Debugging-as-Episodes/generate_sankey_diagram.py at main · AI-4-SE/FSE-2025-Understanding-Debugging-as-Episodes · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
import os
import pandas as pd
import holoviews as hv
from holoviews import opts

hv.extension('matplotlib')

episoden_colors = [
    "#8dd3c7",
    "#ffffb3",
    "#bebada",
    "#fb8072",
    "#80b1d3",
    "#fdb462",
    "#b3de69",
    "#fccde5",
    "#d9d9d9",
    "#bc80bd",
    "#ccebc5",
    "#ffed6f"
]

strategies = {'follow control flow': "Follow control flow",
              'follow data flow': 'Follow data flow',
              'configuration': "Program comprehension",
              'hotspot': "Program comprehension", 'accept/reject hypothesis': "Scientific strategy",
              'scanning': "Inspect source code", 'planning': "Planning",
              'follow intuition': "Intuition", 'code reading': "Inspect source code",
              'state/test hypothesis': "Scientific strategy",
              'bug report': "Program comprehension", 'simulate execution': "Simulation"}


def read_data(path):
    df = pd.read_csv(os.path.join(path, "actions-by-participant.csv"))
    to_replace = {'Cause-effect chain starts': None,
                  'Changes location in code/graph because of a method-call': "follow control flow",
                  'Changes location in code/graph because of a variable': "follow data flow",
                  'Configuration': "configuration",
                  'Hotspot': "hotspot",
                  'Open Codewindow (from Code Window)': None,
                  'Open Codewindow (from Graph)': None,
                  'Open Method inside window (click on graph)': None,
                  'Open Method inside window (click on window)': None,
                  'P ist lost und braucht Hilfe': None,
                  'Player Movement': None,
                  'Reposition Windows': None,
                  'accept hypothesis': "accept/reject hypothesis",
                  'browsing/scrolling/scanning': "scanning",
                  'deducting from prior knowledge': None,
                  'expresses a plan or series of tasks that they want to do': "planning",
                  'expresses insight/statement/assumption\nbased on a feeling/experience/intuition': "follow intuition",
                  'plan successfully executed': None,
                  'reject hypothesis': "accept/reject hypothesis",
                  'specific code area, detailed reading': "code reading",
                  'state & test hypothesis': "state/test hypothesis",
                  'inducting from code bases': None,
                  'Bug report': "bug report",
                  'Close Codewindow': None,
                  'problem formulation': None,
                  'browsing/scrolling/scanning\n(attention on code window)': "scanning",
                  'simulates execution': "simulate execution",
                  'browsing/scrolling/scanning\n(attention on code window)\n(some expression about source code)': "scanning",
                  'Cause-effect chain end': None}

    df["Action"] = df["Action"].replace(to_replace)

    df = df[[True if a else False for a in df["Action"]]]
    return df


def plot_sankey(df, out_path):
    strategies_actions = [{"source": v, "target": k,
                           "value": sum(df[df["Action"] == k]["Marker"]) if sum(df[df["Action"] == k]["Marker"]) else 1}
                          for k, v in strategies.items()]
    actions_participant = [{"source": r["Action"], "target": r["Participant"], "value": r["Marker"]} for _, r in
                           df.iterrows()]
    unused = [{"source": action, "target": "unused", "value": 1} for action in
              pd.DataFrame(strategies_actions)["target"] if
              action not in list(pd.DataFrame(actions_participant)["source"])]

    nodes_gray = list(set(strategies.keys())) + list(set(strategies.values()))
    nodes_colors = ["P" + str(x) for x in range(1, 13)]
    color_mapping = {s: "#d1cdcd" for s in nodes_gray} | {nodes_colors[i]: episoden_colors[i] for i in
                                                          range(len(nodes_colors))}
    # print(color_mapping)

    edges = pd.DataFrame(strategies_actions + actions_participant + unused)
    sankey = hv.Sankey(edges)
    sankey.opts(
        opts.Sankey(
            edge_color='target',
            cmap=color_mapping,
            node_color="index",
            label_position='left',
            fig_inches=(20, 3),
            # node_padding=20,
        )
    )

    hv.save(sankey, os.path.join(out_path, 'sankey_plot_overview.pdf'), fmt='pdf', dpi=300)
    return


def gen_sankey():
    path = './data/user study'
    out_path = './output'
    df = read_data(path)
    plot_sankey(df, out_path)
    return