STRLCPY/451-CorporateRiskMiner

Initial code
Peter Zatka-Haas committed 2 years ago

9be656be

1 parent 419867ce

■ ■ ■ ■ ■ ■

src/app.py

1	+	import streamlit as st
2	+	import pandas as pd
3	+	from streamlit_agraph import agraph, Config
4	+	from utils import build_agraph_components, get_edges_df, get_subgraph_df, get_nodes_df
5	+
6	+
7	+	st.set_page_config(layout="wide")
8	+
9	+
10	+	SLIDER_MIN = 1
11	+	SLIDER_MAX = 100
12	+	SLIDER_DEFAULT = 50
13	+
14	+	nodes = get_nodes_df()
15	+	edges = get_edges_df()
16	+	subgraphs = get_subgraph_df()
17	+
18	+	with st.sidebar:
19	+	st.title("Corporate risks")
20	+
21	+	weight_chains = st.slider(
22	+	"Long ownership chains",
23	+	min_value=SLIDER_MIN,
24	+	max_value=SLIDER_MAX,
25	+	value=SLIDER_DEFAULT,
26	+	)
27	+	weight_cyclic = st.slider(
28	+	"Cyclic ownership",
29	+	min_value=SLIDER_MIN,
30	+	max_value=SLIDER_MAX,
31	+	value=SLIDER_DEFAULT,
32	+	)
33	+	weight_psc_haven = st.slider(
34	+	"Persons of significant control associated with tax havens",
35	+	min_value=SLIDER_MIN,
36	+	max_value=SLIDER_MAX,
37	+	value=SLIDER_DEFAULT,
38	+	)
39	+	weight_pep = st.slider(
40	+	"Officers/PSCs are politically exposed",
41	+	min_value=SLIDER_MIN,
42	+	max_value=SLIDER_MAX,
43	+	value=SLIDER_DEFAULT,
44	+	)
45	+	weight_sanctions = st.slider(
46	+	"Officers/PSCs/Companies are sanctioned",
47	+	min_value=SLIDER_MIN,
48	+	max_value=SLIDER_MAX,
49	+	value=SLIDER_DEFAULT,
50	+	)
51	+	weight_disqualified = st.slider(
52	+	"Officers are disqualified directors",
53	+	min_value=SLIDER_MIN,
54	+	max_value=SLIDER_MAX,
55	+	value=SLIDER_DEFAULT,
56	+	)
57	+
58	+	custom_names_a = st.multiselect(
59	+	label="Custom persons of interest",
60	+	options=nodes["node_id"],
61	+	default=None,
62	+	)
63	+	custom_names_b = st.file_uploader(label="Custom persons of interest", type="csv")
64	+
65	+	go = st.button("Go")
66	+
67	+
68	+	with st.container():
69	+	st.write(subgraphs)
70	+
71	+	selected_subgraph_hash = st.selectbox(
72	+	label="Select subgraph to explore", options=subgraphs.index
73	+	)
74	+
75	+	nodes_selected = nodes.loc[nodes["subgraph_hash"] == selected_subgraph_hash]
76	+	edges_selected = edges.loc[edges["subgraph_hash"] == selected_subgraph_hash]
77	+
78	+	with st.container():
79	+
80	+	col1, col2 = st.columns(2)
81	+
82	+	with col1:
83	+	(node_objects, edge_objects) = build_agraph_components(
84	+	nodes_selected, edges_selected
85	+	)
86	+	agraph(
87	+	nodes=node_objects,
88	+	edges=edge_objects,
89	+	config=Config(
90	+	width=500,
91	+	height=500,
92	+	),
93	+	)
94	+
95	+	with col2:
96	+	st.write(nodes_selected)
97	+

■ ■ ■ ■ ■ ■

src/utils.py

1	+	import streamlit as st
2	+	from streamlit_agraph import Node, Edge
3	+	import json
4	+	import pandas as pd
5	+
6	+	NODE_COLOUR_PERSON = "#4684B2"
7	+	NODE_COLOUR_COMPANY = "#46B247"
8	+
9	+
10	+	@st.cache()
11	+	def get_subgraph_df():
12	+	return pd.read_csv("subgraphs.csv", index_col="subgraph_hash")
13	+
14	+
15	+	@st.cache()
16	+	def get_nodes_df():
17	+	return pd.read_csv("nodes.csv")
18	+
19	+
20	+	@st.cache()
21	+	def get_edges_df():
22	+	return pd.read_csv("edges.csv")
23	+
24	+
25	+	@st.cache()
26	+	def build_agraph_components(
27	+	nodes,
28	+	edges,
29	+	node_colour_person=NODE_COLOUR_PERSON,
30	+	node_colour_company=NODE_COLOUR_COMPANY,
31	+	):
32	+	"""Create agraph object from node and edge list"""
33	+
34	+	node_objects = []
35	+	edge_objects = []
36	+
37	+	for _, row in nodes.iterrows():
38	+	node_metadata = json.loads(row["node_metadata"])
39	+	node_objects.append(
40	+	Node(
41	+	id=row["node_id"],
42	+	label=node_metadata["name"],
43	+	size=25,
44	+	color=node_colour_person
45	+	if row["is_person"] == 1
46	+	else node_colour_company,
47	+	)
48	+	)
49	+
50	+	for _, row in edges.iterrows():
51	+	edge_objects.append(
52	+	Edge(
53	+	source=row["source"],
54	+	label=row["type"],
55	+	target=row["target"],
56	+	)
57	+	)
58	+
59	+	return (node_objects, edge_objects)
60	+

Initial code