How gridappsd/cimhub works
Main Components
• Converter Engine – This encapsulates the conversion modules that translate between OpenDSS, CIM, GridLAB-D, and CSV formats. There are separate conversion scripts in Java (found under “cimhub/src” and Maven-based builds) and in Python (in “src_python/cimhub”).
• Input/Output Modules – Multiple directories (for IEEE test cases, BES, OEDI, CPYDAR, ecP, etc.) provide inputs (raw or test circuit files) and expect outputs (converted files in various formats). “example”, “tutorial”, and “tests” are key directories that receive/process these files.
• Command-line/Script Interfaces – There are various shell scripts, batch files (e.g., start.sh, stop.sh, onestep.py), and Docker-related files (e.g., Dockerfile, docker-compose.yml) which show that the system supports a CLI-based workflow as well as container-based deployment.
• External Dependencies/Third-Party Tools – Include references to external converters such as OpenDSSCmd and GridLAB-D, as well as dependencies like SparkQLWrapper (for Python) and Apache Maven for building Java code.
Relationships and Data Flow
• Bidirectional interactions between the converter modules and the Blazegraph triple-store (using SPARQL queries for both input retrieval and database population).
• For Java-based processing (the “cimhub/src” module) and Python-based processing (in “src_python/cimhub”), indicate that they use their respective dependencies to perform operations like CIM import, conversion, and export.
• From the converter engine outputs, show separate flows to external tools like GridLAB-D and OpenDSS (also to CSV generation) which are validated through automated test suites (shown under the “tests” directory) and example pipelines.
• If desired, highlight that the Docker orchestration binds volumes (e.g., “/data” for persistent storage and “/platform” for host integration) and encapsulates the runtime environment for the system.
Architectural Patterns/Principles
• Modularity – The system is segmented into various conversion, testing, input handling, and output generation components, each developed in separate directories or even languages (Java vs. Python).
• Separation of Concerns – The design naturally separates data storage (Blazegraph), conversion logic (CIMHub Java/Python converters), and utility/testing scripts.
• Docker Integration – Illustrate the containerization strategy to encapsulate dependencies and ease deployment across different platforms (Linux, Windows, Mac).
Details on Technologies and Frameworks
• Java (OpenJDK 11, Apache Maven, and associated build scripts such as build.gradle, gradlew) for core conversion modules.
• Python (version 3.8+ along with SPARQLWrapper and other libraries available in “src_python/cimhub”) for scripting and automation.
• Blazegraph as the triple-store for CIM data.
• Docker for containerizing the runtime environment.
• External tools like OpenDSSCmd and GridLAB-D which interact with the converted outputs.
• Git LFS (for handling large CIM and test case files).
Recommended Elements
• Each component should be clearly labeled.
• Optionally, annotate version information or dependencies (e.g., Java 11, Python 3.8+), and include Docker as an encapsulation layer if the deployment architecture is intended to cover that aspect.