Skip to content

Commit a2b06c2

Browse files
JohTJohT
committed
Introduce anomaly detection pipeline with tuned clustering
1 parent a36293f commit a2b06c2

24 files changed

+2450
-0
lines changed

‎domains/anomaly-detection/anomalyDetectionPipeline.sh‎

Lines changed: 181 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,181 @@
1+
#!/usr/bin/env bash
2+
3+
# Pipeline that coordinates anomaly detection using the Graph Data Science Library of Neo4j.
4+
# It requires an already running Neo4j graph database with already scanned and analyzed artifacts.
5+
# The results will be written into the sub directory reports/anomaly-detection.
6+
7+
# Note that "scripts/prepareAnalysis.sh" is required to run prior to this script.
8+
9+
# Requires executeQueryFunctions.sh, projectionFunctions.sh, cleanupAfterReportGeneration.sh
10+
11+
# Fail on any error ("-e" = exit on first error, "-o pipefail" exist on errors within piped commands)
12+
set -o errexit -o pipefail
13+
14+
# Overrideable Constants (defaults also defined in sub scripts)
15+
REPORTS_DIRECTORY=${REPORTS_DIRECTORY:-"reports"}
16+
17+
## Get this "scripts/reports" directory if not already set
18+
# Even if $BASH_SOURCE is made for Bourne-like shells it is also supported by others and therefore here the preferred solution.
19+
# CDPATH reduces the scope of the cd command to potentially prevent unintended directory changes.
20+
# This way non-standard tools like readlink aren't needed.
21+
ANOMALY_DETECTION_SCRIPT_DIR=${ANOMALY_DETECTION_SCRIPT_DIR:-$(CDPATH=. cd -- "$(dirname -- "${BASH_SOURCE[0]}")" && pwd -P)}
22+
echo "anomalyDetectionPipeline: ANOMALY_DETECTION_SCRIPT_DIR=${ANOMALY_DETECTION_SCRIPT_DIR}"
23+
# Get the "scripts" directory by taking the path of this script and going one directory up.
24+
SCRIPTS_DIR=${SCRIPTS_DIR:-"${ANOMALY_DETECTION_SCRIPT_DIR}/../../scripts"} # Repository directory containing the shell scripts
25+
# Get the "cypher" query directory for gathering features.
26+
ANOMALY_DETECTION_FEATURE_CYPHER_DIR=${ANOMALY_DETECTION_FEATURE_CYPHER_DIR:-"${ANOMALY_DETECTION_SCRIPT_DIR}/features"}
27+
ANOMALY_DETECTION_QUERY_CYPHER_DIR=${ANOMALY_DETECTION_QUERY_CYPHER_DIR:-"${ANOMALY_DETECTION_SCRIPT_DIR}/queries"}
28+
29+
# Function to display script usage
30+
usage() {
31+
echo -e "${COLOR_ERROR}" >&2
32+
echo "Usage: $0 [--verbose]" >&2
33+
echo -e "${COLOR_DEFAULT}" >&2
34+
exit 1
35+
}
36+
37+
# Default values
38+
verboseMode="" # either "" or "--verbose"
39+
40+
# Parse command line arguments
41+
while [[ $# -gt 0 ]]; do
42+
key="$1"
43+
value="${2}"
44+
45+
case ${key} in
46+
--verbose)
47+
verboseMode="--verbose"
48+
;;
49+
*)
50+
echo -e "${COLOR_ERROR}anomalyDetectionPipeline: Error: Unknown option: ${key}${COLOR_DEFAULT}" >&2
51+
usage
52+
;;
53+
esac
54+
shift || true # ignore error when there are no more arguments
55+
done
56+
57+
# Define functions to execute a cypher query from within a given file (first and only argument) like "execute_cypher"
58+
source "${SCRIPTS_DIR}/executeQueryFunctions.sh"
59+
60+
# Define functions to create and delete Graph Projections like "createUndirectedDependencyProjection"
61+
source "${SCRIPTS_DIR}/projectionFunctions.sh"
62+
63+
# Query or recalculate features.
64+
#
65+
# Required Parameters:
66+
# - projection_name=...
67+
# Name prefix for the in-memory projection name. Example: "package-anomaly-detection"
68+
# - projection_node_label=...
69+
# Label of the nodes that will be used for the projection. Example: "Package"
70+
# - projection_weight_property=...
71+
# Name of the node property that contains the dependency weight. Example: "weight"
72+
anomaly_detection_features() {
73+
# Determine the Betweenness centrality (with the directed graph projection) if not already done
74+
execute_cypher_queries_until_results "${ANOMALY_DETECTION_FEATURE_CYPHER_DIR}/AnomalyDetectionFeature-Betweenness-Exists.cypher" \
75+
"${ANOMALY_DETECTION_FEATURE_CYPHER_DIR}/AnomalyDetectionFeature-Betweenness-Write.cypher" "${@}"
76+
# Determine the local clustering coefficient if not already done
77+
execute_cypher_queries_until_results "${ANOMALY_DETECTION_FEATURE_CYPHER_DIR}/AnomalyDetectionFeature-LocalClusteringCoefficient-Exists.cypher" \
78+
"${ANOMALY_DETECTION_FEATURE_CYPHER_DIR}/AnomalyDetectionFeature-LocalClusteringCoefficient-Write.cypher" "${@}"
79+
# Determine the page rank if not already done
80+
execute_cypher_queries_until_results "${ANOMALY_DETECTION_FEATURE_CYPHER_DIR}/AnomalyDetectionFeature-PageRank-Exists.cypher" \
81+
"${ANOMALY_DETECTION_FEATURE_CYPHER_DIR}/AnomalyDetectionFeature-PageRank-Write.cypher" "${@}"
82+
# Determine the article rank if not already done
83+
execute_cypher_queries_until_results "${ANOMALY_DETECTION_FEATURE_CYPHER_DIR}/AnomalyDetectionFeature-ArticleRank-Exists.cypher" \
84+
"${ANOMALY_DETECTION_FEATURE_CYPHER_DIR}/AnomalyDetectionFeature-PageRank-Write.cypher" "${@}"
85+
}
86+
# Run queries to find anomalies in the graph.
87+
#
88+
# Required Parameters:
89+
# - projection_node_label=...
90+
# Label of the nodes that will be used for the projection. Example: "Package"
91+
anomaly_detection_queries() {
92+
local nodeLabel
93+
nodeLabel=$( extractQueryParameter "projection_node_label" "${@}" )
94+
95+
execute_cypher "${ANOMALY_DETECTION_QUERY_CYPHER_DIR}/AnomalyDetectionPotentialImbalancedRoles.cypher" "${@}" > "${FULL_REPORT_DIRECTORY}/${nodeLabel}AnomalyDetection_PotentialImbalancedRoles.csv"
96+
execute_cypher "${ANOMALY_DETECTION_QUERY_CYPHER_DIR}/AnomalyDetectionPotentialOverEngineerOrIsolated.cypher" "${@}" > "${FULL_REPORT_DIRECTORY}/${nodeLabel}AnomalyDetection_PotentialOverEngineerOrIsolated.csv"
97+
98+
execute_cypher "${ANOMALY_DETECTION_QUERY_CYPHER_DIR}/AnomalyDetectionHiddenBridgeNodes.cypher" "${@}" > "${FULL_REPORT_DIRECTORY}/${nodeLabel}AnomalyDetection_HiddenBridgeNodes.csv"
99+
execute_cypher "${ANOMALY_DETECTION_QUERY_CYPHER_DIR}/AnomalyDetectionPopularBottlenecks.cypher" "${@}" > "${FULL_REPORT_DIRECTORY}/${nodeLabel}AnomalyDetection_PopularBottlenecks.csv"
100+
execute_cypher "${ANOMALY_DETECTION_QUERY_CYPHER_DIR}/AnomalyDetectionSilentCoordinators.cypher" "${@}" > "${FULL_REPORT_DIRECTORY}/${nodeLabel}AnomalyDetection_SilentCoordinators.csv"
101+
execute_cypher "${ANOMALY_DETECTION_QUERY_CYPHER_DIR}/AnomalyDetectionOverReferencesUtilities.cypher" "${@}" > "${FULL_REPORT_DIRECTORY}/${nodeLabel}AnomalyDetection_OverReferencesUtilities.csv"
102+
execute_cypher "${ANOMALY_DETECTION_QUERY_CYPHER_DIR}/AnomalyDetectionFragileStructuralBridges.cypher" "${@}" > "${FULL_REPORT_DIRECTORY}/${nodeLabel}AnomalyDetection_FragileStructuralBridges.csv"
103+
execute_cypher "${ANOMALY_DETECTION_QUERY_CYPHER_DIR}/AnomalyDetectionDependencyHungryOrchestrators.cypher" "${@}" > "${FULL_REPORT_DIRECTORY}/${nodeLabel}AnomalyDetection_DependencyHungryOrchestrators.csv"
104+
execute_cypher "${ANOMALY_DETECTION_QUERY_CYPHER_DIR}/AnomalyDetectionUnexpectedCentralNodes.cypher" "${@}" > "${FULL_REPORT_DIRECTORY}/${nodeLabel}AnomalyDetection_UnexpectedCentralNodes.csv"
105+
}
106+
107+
# Run the anomaly detection pipeline.
108+
#
109+
# Required Parameters:
110+
# - projection_name=...
111+
# Name prefix for the in-memory projection name. Example: "package-anomaly-detection"
112+
# - projection_node_label=...
113+
# Label of the nodes that will be used for the projection. Example: "Package"
114+
# - projection_weight_property=...
115+
# Name of the node property that contains the dependency weight. Example: "weight"
116+
anomaly_detection_pipeline() {
117+
time anomaly_detection_features "${@}"
118+
time anomaly_detection_queries "${@}"
119+
# Get tuned Leiden communities as a reference to tune clustering
120+
time "${ANOMALY_DETECTION_SCRIPT_DIR}/tunedLeidenCommunityDetection.py" "${@}" ${verboseMode}
121+
# Tuned Fast Random Projection and tuned HDBSCAN clustering
122+
time "${ANOMALY_DETECTION_SCRIPT_DIR}/tunedNodeEmbeddingClustering.py" "${@}" ${verboseMode}
123+
124+
# Query Results: Output all collected features into a CSV file.
125+
local nodeLabel
126+
nodeLabel=$( extractQueryParameter "projection_node_label" "${@}" )
127+
execute_cypher "${ANOMALY_DETECTION_FEATURE_CYPHER_DIR}/AnomalyDetectionFeatures.cypher" "${@}" > "${FULL_REPORT_DIRECTORY}/${nodeLabel}AnomalyDetectionFeatures.csv"
128+
}
129+
130+
# Create report directory
131+
REPORT_NAME="anomaly-detection"
132+
FULL_REPORT_DIRECTORY="${REPORTS_DIRECTORY}/${REPORT_NAME}"
133+
mkdir -p "${FULL_REPORT_DIRECTORY}"
134+
135+
# Query Parameter key pairs for projection and algorithm side
136+
PROJECTION_NAME="dependencies_projection"
137+
ALGORITHM_PROJECTION="projection_name"
138+
139+
PROJECTION_NODE="dependencies_projection_node"
140+
ALGORITHM_NODE="projection_node_label"
141+
142+
PROJECTION_WEIGHT="dependencies_projection_weight_property"
143+
ALGORITHM_WEIGHT="projection_weight_property"
144+
145+
# Code independent algorithm parameters
146+
COMMUNITY_PROPERTY="community_property=communityLeidenIdTuned"
147+
148+
# -- Java Artifact Node Embeddings -------------------------------
149+
150+
if createUndirectedDependencyProjection "${PROJECTION_NAME}=artifact-anomaly-detection" "${PROJECTION_NODE}=Artifact" "${PROJECTION_WEIGHT}=weight"; then
151+
createDirectedDependencyProjection "${PROJECTION_NAME}=artifact-anomaly-detection-directed" "${PROJECTION_NODE}=Artifact" "${PROJECTION_WEIGHT}=weight"
152+
anomaly_detection_pipeline "${ALGORITHM_PROJECTION}=artifact-anomaly-detection" "${ALGORITHM_NODE}=Artifact" "${ALGORITHM_WEIGHT}=weight" "${COMMUNITY_PROPERTY}"
153+
fi
154+
155+
# -- Java Package Node Embeddings --------------------------------
156+
157+
if createUndirectedDependencyProjection "${PROJECTION_NAME}=package-anomaly-detection" "${PROJECTION_NODE}=Package" "${PROJECTION_WEIGHT}=weight25PercentInterfaces"; then
158+
createDirectedDependencyProjection "${PROJECTION_NAME}=package-anomaly-detection-directed" "${PROJECTION_NODE}=Package" "${PROJECTION_WEIGHT}=weight25PercentInterfaces"
159+
anomaly_detection_pipeline "${ALGORITHM_PROJECTION}=package-anomaly-detection" "${ALGORITHM_NODE}=Package" "${ALGORITHM_WEIGHT}=weight25PercentInterfaces" "${COMMUNITY_PROPERTY}"
160+
fi
161+
162+
# -- Java Type Node Embeddings -----------------------------------
163+
164+
if createUndirectedJavaTypeDependencyProjection "${PROJECTION_NAME}=type-anomaly-detection"; then
165+
createDirectedJavaTypeDependencyProjection "${PROJECTION_NAME}=type-anomaly-detection-directed"
166+
anomaly_detection_pipeline "${ALGORITHM_PROJECTION}=type-anomaly-detection" "${ALGORITHM_NODE}=Type" "${ALGORITHM_WEIGHT}=weight" "${COMMUNITY_PROPERTY}"
167+
fi
168+
169+
# -- Typescript Module Node Embeddings ---------------------------
170+
171+
if createUndirectedDependencyProjection "${PROJECTION_NAME}=typescript-module-embedding" "${PROJECTION_NODE}=Module" "${PROJECTION_WEIGHT}=lowCouplingElement25PercentWeight"; then
172+
createDirectedDependencyProjection "${PROJECTION_NAME}=typescript-module-embedding-directed" "${PROJECTION_NODE}=Module" "${PROJECTION_WEIGHT}=lowCouplingElement25PercentWeight"
173+
anomaly_detection_pipeline "${ALGORITHM_PROJECTION}=typescript-module-embedding" "${ALGORITHM_NODE}=Module" "${ALGORITHM_WEIGHT}=lowCouplingElement25PercentWeight" "${COMMUNITY_PROPERTY}"
174+
fi
175+
176+
# ---------------------------------------------------------------
177+
178+
# Clean-up after report generation. Empty reports will be deleted.
179+
source "${SCRIPTS_DIR}/cleanupAfterReportGeneration.sh" "${FULL_REPORT_DIRECTORY}"
180+
181+
echo "anomalyDetectionPipeline: $(date +'%Y-%m-%dT%H:%M:%S%z') Successfully finished."

0 commit comments

Comments
 (0)