Software Dependency Analysis with Graph Databases (Neo4J)

Talk with Philipp Haußleiter at GraphConnect 2015.

Transcript

1. Repo Census Software Dependency Analysis with Graph Databases Philipp Haußleiter & Oliver Tigges
2. Agenda Software Dependency Management Repo Census: Use Cases and Demo Graph Model for Software Dependencies Queries and Cypher
3. Software Dependencies
4. Declaration of Dependencies <project> <modelVersion>4.0.0</modelVersion> <groupId>com.innoq.census</groupId> <artifactId>census-webapp</artifactId> <version>1.0</version> <name>repo-census-webapp</name> <dependencies> <dependency> <groupId>javax.servlet</groupId> <artifactId>javax.servlet-api</artifactId> <version>3.1.0</version> </dependency> <dependency> <groupId>org.springframework</groupId> <artifactId>org.springframework</artifactId> <version>4.1.3.RELEASE</version> </dependency> <dependency> <groupId>org.hibernate</groupId> <artifactId>hibernate-core</artifactId> <version>4.3.9.Final</version> </dependency> </dependencies> </project>
5. Transitive Dependencies
6. Complexity explosion
7. Repo Census
8. Use Cases Keep overview and be aware of changes Estimate impacts of software updates License issues Security issues
9. System Architecture Akka (Worker) Play2 (Web UI) Query Result neo4j (GraphDB) Add Artifact Process Artifact Job Request Artifact Job Add Artifact Job Eclipse Aether Client … Process Artifact Request Artifact (Maven) Repository … activeMQ (Queue)
10. Demo Short introduction of the web app Goto to Maven Census
11. Graph Model
12. Transfer to Neo4j Node Elements to handle: Dependencies License Version GroupId ArtifactId
13. Conceptual model <project> <parent> <groupId>com.explodingpixels</groupId> <artifactId>macwidgets-maven-modul</artifactId> <version>0.9.6-SNAPSHOT</version> </parent> <modelVersion>4.0.0</modelVersion> <groupId>com.explodingpixels</groupId> <artifactId>macwidgets-demo</artifactId> <name>macwidgets-demo</name> <licenses> <license> <name>GNU Lesser General Public License</name> <url>http://www.gnu.org/licenses/lgpl.html</url </license> </licenses> <dependencies> <dependency> <groupId>com.explodingpixels</groupId> <artifactId>macwidgets</artifactId> <version>0.9.6-SNAPSHOT</version> </dependency> </dependencies> </project>
14. Maven Graph Model
15. Maven Graph Model
16. See model in action Goto to Neo console
17. Cypher Queries
18. Search Find by unique version ID MATCH (v:Version {uId: 'com.innoq.census:census-webapp:jar:1.0'}) RETURN v.uId as UID, v:ToDo AS todo, v:Processing AS processing Search with regular expression MATCH (v:Version) WHERE v.uId =~ 'com.innoq.*' RETURN v.uId as UID
19. Get license info MATCH (v:Version {uId: {0}})-[:HAS_LICENSE]->(l:LicenseInfo) RETURN l.name AS lic_name, l.url AS url ORDER BY lic_name
20. Show other versions MATCH (v:Version)-[:VERSION_OF]->(ae:ArtifactElement) <-[:VERSION_OF]-(o:Version) WHERE v.uId = {0} RETURN o.uId AS id ORDER BY id
21. Show direct dependencies MATCH (v:Version {uId: {0}}) OPTIONAL MATCH v-[dr:DEPENDS_ON]->other RETURN other.uId AS dep, dr.scope AS scope ORDER BY dep
22. Show transitive dependencies MATCH (v:Version) WHERE v.uId =~ 'com.innoq.*1.0' OPTIONAL MATCH (v)-[r:DEPENDS_ON_COMPILE|DEPENDS_ON_RUNTIME*1..10]->other WITH distinct other as deps RETURN deps.uId
23. Transitive dependency paths MATCH (v:Version) WHERE v.uId =~ 'com.innoq.*1.0' OPTIONAL MATCH p = (v-[r:DEPENDS_ON_COMPILE|DEPENDS_ON_RUNTIME*1..10]->oth WITH tail ( extract (d IN nodes(p) | d.uId) ) as deps RETURN deps
24. License overview of all deps MATCH (v:Version) WHERE v.uId =~ 'com.innoq.*1.0' OPTIONAL MATCH v-[r:DEPENDS_ON_COMPILE|DEPENDS_ON_RUNTIME*1..5]->other -[:HAS_LICENSE]->(l:LicenseInfo) RETURN DISTINCT l.name as License, l.url as URL ORDER BY l.name
25. Changes between versions MATCH (orig:Version {uId: {0}}), (other:Version {uId: {1}}) OPTIONAL MATCH (orig)-[:DEPENDS_ON]->(common:Version)<-[:DEPENDS_ON]-(other) WITH orig, other, collect(DISTINCT common.uId) AS common OPTIONAL MATCH (orig)-[:DEPENDS_ON]->(origVersion:Version)-[:VERSION_OF]->(commonArtifact:ArtifactElement), (other)-[:DEPENDS_ON]->(otherVersion:Version)-[:VERSION_OF]->(commonArtifact) WITH orig, other, common, collect({origVersion: origVersion.uId, otherVersion: otherVersion.uId}) AS commonArtifacts OPTIONAL MATCH (orig)-[:DEPENDS_ON]->(origOnly:Version)-[:VERSION_OF]->(a:ArtifactElement) WHERE NOT (other)-[:DEPENDS_ON]->(:Version)-[:VERSION_OF]->(a) WITH orig, other, common, commonArtifacts, collect(DISTINCT origOnly.uId) AS origOnly OPTIONAL MATCH (other)-[:DEPENDS_ON]->(otherOnly:Version)-[:VERSION_OF]->(a:ArtifactElement) WHERE NOT (orig)-[:DEPENDS_ON]->(:Version)-[:VERSION_OF]->(a) RETURN common, commonArtifacts, origOnly, collect(DISTINCT otherOnly.uId) AS otherOnly
26. Changes between versions
27. Project Status
28. Limits Early Stage Quantity structures Limits of the System
29. Outlook
30. Planned features Support for Multiple Repositories Additional Attributes: Organization, Developers, Properties, Repositories, Plugins Support for (Multi-)User Project Upload (Pom) Social Login (Github/Twitter/...)
31. Integration & Cooperation Support other Repositories: Gems, RPMs, NPMs, ... Integration in build process Integration of jQAssistant
32. Availability Project Page How to get, test and contribute SaaS Appliance (OVA)
33. Q&A
34. Thank you! Philipp Haußleiter philipp.haussleiter@innoq.com @phaus Oliver Tigges oliver.tigges@innoq.com @otigges