CIMgo is a Go library and CLI for working with CGMES/CIM electrical grid data. It parses CGMES XML instance files, validates them against SHACL and SPARQL based rules, and generates typed Go structs and Protobuf definitions from CIM RDF schemas.

cimcli is a command-line tool for working with CGMES (Common Grid Model Exchange Standard) XML files.

Pre-built binaries are attached to each GitHub release:

Linux — make the binary executable after downloading:

chmod +x cimcli-linux-amd64

Windows — the .exe can be run directly from PowerShell or CMD.

Validates CGMES XML instance files against CGMES SHACL and SPARQL rules. Profiles, solved/not-solved state, and EQBD base voltage IDs are detected automatically from the file headers.

• Runs over 4,000 generated checks derived from standard ENTSO-E SHACL files.
• Supports EQ, SSH, TP, SV, DL, DY, SC, GL, OP, and EQBD profiles.
• Rule silencing: dl:DiagramObject.IdentifiedObject-valueType and sv:SvStatus.ConductingEquipment-valueType are silenced by default; additional rules can be suppressed via -silence.
• Human-readable text and JSON output formats.

cimcli validate [options] <xml-file1> [<xml-file2> ...]

Exit code is 0 when no sh:Violation-severity findings are present, 1 otherwise.

Merges one or more CGMES XML files and outputs the combined dataset as JSON.

cimcli convert <xml-file1> [<xml-file2> ...]

Validate a full model:

# Linux
./cimcli-linux-amd64 validate EQ.xml SSH.xml TP.xml SV.xml

# Windows
cimcli-windows-amd64.exe validate EQ.xml SSH.xml TP.xml SV.xml

Validate specific profiles:

./cimcli-linux-amd64 validate -profile EQ,SSH,TP,SV,DL PST_Type1_*.xml

Validate ENTSO-E test configurations:

./cimcli-linux-amd64 validate -profile EQ,SSH,TP,SV,DL \
  CGMES-Test-Configurations/v3.0/PST/PST_PhaseTapChangerLinear_Type1/*.xml

Output violations as JSON:

./cimcli-linux-amd64 validate -json \
  20210401T1730Z_1D_BE_EQ_1.xml \
  20210401T1730Z_1D_BE_SSH_1.xml \
  20210401T1730Z_1D_BE_TP_1.xml \
  20210401T1730Z_1D_BE_SV_1.xml

Convert files to JSON:

./cimcli-linux-amd64 convert EQ.xml SSH.xml > dataset.json

Make sure that you have cloned the repo recursively to include the CGMES schema files from ENTSO-E

git clone --recurse-submodules [...]

or clone the submodule in a second step

git submodule update --init --recursive

Ensure that GOPATH is set and included in your PATH.

For the protobuf code generation, you also require the proto compiler

sudo apt-get install -y protobuf-compiler
# install tools from mod file
go get tool

# Generate all code (must run before build after schema changes)
go generate ./...

# Build
go build -v ./...

# Run all tests
go test -v ./...

# Run a single test
go test -v ./path/to/package -run TestName

Benchmarks covering end-to-end validation and per-profile breakdown live in validation/cgmes_config_test.go.

Rank profiles by wall time (quick, no pprof overhead):

go test -run='^$' \
    -bench='BenchmarkRealGridValidate(EQ|SSH|TP|SV|Common)$' \
    -benchtime=3x ./validation/

Full pipeline with CPU and memory profiles (RealGrid, ~115 MB, 4 profiles):

go test -run='^$' \
    -bench=BenchmarkRealGridValidation \
    -benchtime=3x \
    -cpuprofile=cpu.prof \
    -memprofile=mem.prof \
    ./validation/

Inspect results:

go tool pprof -http=:6060 cpu.prof   # flame graph + top + source view
go tool pprof -http=:6061 mem.prof   # set Sample dropdown to alloc_space

In the flame graph, the widest bands are the hottest call stacks. The Top view's flat column shows a function's own time; cum includes its callees. Click any function in Top to open the annotated Source view.

The code generation process follows these main steps:

1. Schema Loading: The tool begins by finding and parsing the relevant CIM RDF schema or TTL SHACL files based on the specified version and profiles.
2. Schema Processing: It processes the parsed RDF or SHACL data into an internal, language-agnostic representation of CIM classes, properties, datatypes, their relationships and rules.
3. Code Generation: Using Go's text/template engine, it feeds the internal representation into language-specific templates (lang-templates/*.tmpl) to generate the final source code files.

cmd/shaclgen translates each constraint in the CGMES SHACL Turtle files into a Go Check<...> function under shaclgen/. The generated checks are wired into per-profile orchestrators and aggregated by shaclgen.ValidateAllGeneratedProfiles.

For a complete validation pass including both generated and hand-written SPARQL rules, use validation.ValidateAllProfiles.

Across 73 profiles there are 9153 constraints total. Of these, 4184 generate code and 4969 are skipped: 4933 are structurally satisfied by the Go type system (generating a check would never fire or would produce false positives) and 36 cannot be conducted due to upstream SHACL TTL defects.

Before code generation, shaclimport.SimplifyFileResults normalises each property shape's constraint list. Rules are applied in order; a constraint that matches a rule is either dropped or rewritten and is not passed on to cmd/shaclgen.

Complex constraints defined using sh:sparql in the CGMES SHACL files are not automatically generated by cmd/shaclgen. These are instead implemented as hand-written Go functions in the validation/ package and wired into the profile validators.

Across all profiles, there are approximately 200 SPARQL-based constraints. All active constraints are currently implemented, covering 100% of the active validation requirements defined in the CGMES standard.

Each manual validation rule is standardized with a header that provides traceability to the source profile and rule ID:

// CheckSynchronousMachineAggregate implements eq452:SynchronousMachine-aggregate
// Profile: 61970-452_Equipment-AP-Con-Complex
// Origin: Derived from a SPARQL constraint.

CIMdesk implements two categories of checks that are not encoded in the CGMES SHACL TTL files and are therefore not covered by cmd/shaclgen or the hand-written SPARQL rules.

These carry a CGMES rule ID but are defined in the conformance document rather than the SHACL TTL files:

These have no CGMES rule ID and appear to be CIMdesk's own heuristics. They are outside the scope of the CGMES SHACL standard: