output folder structure

All seqgra results are stored in the seqgra output directory (OUTPUT_DIR, see output argument -o, --output-dir in Command line utilities). The organization of the output directory relies heavily on grammar, model, and evaluator IDs.

• {OUTPUT_DIR}/defs/data contains the data definition XML files generated by seqgrae.

• {OUTPUT_DIR}/defs/model contains the model definition XML files generated by seqgrae.

• {OUTPUT_DIR}/input/{GRAMMAR ID} contains data synthesized by GRAMMAR ID

• {OUTPUT_DIR}/input/{DATA_FOLDER} contains experimental data of experiment DATA_FOLDER

• {OUTPUT_DIR}/models/{GRAMMAR ID}/{MODEL ID} contains the model with MODEL ID, trained on {OUTPUT_DIR}/input/{GRAMMAR ID}.

• {OUTPUT_DIR}/evaluation/{GRAMMAR ID}/{MODEL ID}/{EVALUATOR ID} contains the output of evaluator with EVALUATOR ID from the model stored at {OUTPUT_DIR}/models/{GRAMMAR ID}/{MODEL ID}.

• {OUTPUT_DIR}/analyses/{ANALYSIS ID}.sh is the analysis shell script generated by seqgrae.

• {OUTPUT_DIR}/model-comparisons/{ANALYSIS ID} contains the comparator output of ANALYSIS ID.

• {OUTPUT_DIR}/attribution/{ANALYSIS ID} contains the feature-importance evaluator outputs from cross-model attribution analysis with ANALYSIS ID.

{OUTPUT_DIR}
|-- defs
|   |-- data
|   +-- model
|-- input
|   |-- {GRAMMAR ID 1}
|   |-- {GRAMMAR ID 2}
|   +-- {GRAMMAR ID n}
|-- models
|   |-- {GRAMMAR ID 1}
|   |   |-- {MODEL ID 1}
|   |   |-- {MODEL ID 2}
|   |   +-- {MODEL ID n}
|   |-- {GRAMMAR ID 2}
|   |   |-- {MODEL ID 1}
|   |   |-- {MODEL ID 2}
|   |   +-- {MODEL ID n}
|   +-- {GRAMMAR ID n}
|       |-- {MODEL ID 1}
|       |-- {MODEL ID 2}
|       +-- {MODEL ID n}
|-- evaluation
|   |-- {GRAMMAR ID 1}
|   |   |-- {MODEL ID 1}
|   |   |   |-- {EVALUATOR ID 1}
|   |   |   |-- {EVALUATOR ID 2}
|   |   |   |-- {EVALUATOR ID n}
|   |   |-- {MODEL ID 2}
|   |   |   |-- {EVALUATOR ID 1}
|   |   |   |-- {EVALUATOR ID 2}
|   |   |   |-- {EVALUATOR ID n}
|   |   +-- {MODEL ID n}
|   |       |-- {EVALUATOR ID 1}
|   |       |-- {EVALUATOR ID 2}
|   |       +-- {EVALUATOR ID n}
|   |-- {GRAMMAR ID 2}
|   |   |-- {MODEL ID 1}
|   |   |   |-- {EVALUATOR ID 1}
|   |   |   |-- {EVALUATOR ID 2}
|   |   |   |-- {EVALUATOR ID n}
|   |   |-- {MODEL ID 2}
|   |   |   |-- {EVALUATOR ID 1}
|   |   |   |-- {EVALUATOR ID 2}
|   |   |   |-- {EVALUATOR ID n}
|   |   +-- {MODEL ID n}
|   |       |-- {EVALUATOR ID 1}
|   |       |-- {EVALUATOR ID 2}
|   |       +-- {EVALUATOR ID n}
|   +-- {GRAMMAR ID n}
|       |-- {MODEL ID 1}
|       |   |-- {EVALUATOR ID 1}
|       |   |-- {EVALUATOR ID 2}
|       |   |-- {EVALUATOR ID n}
|       |-- {MODEL ID 2}
|       |   |-- {EVALUATOR ID 1}
|       |   |-- {EVALUATOR ID 2}
|       |   |-- {EVALUATOR ID n}
|       +-- {MODEL ID n}
|           |-- {EVALUATOR ID 1}
|           |-- {EVALUATOR ID 2}
|           +-- {EVALUATOR ID n}
|-- analyses
|-- model-comparisons
|   |-- {ANALYSIS ID 1}
|   |-- {ANALYSIS ID 1}
|   +-- {ANALYSIS ID n}
+-- attribution
    |-- {ANALYSIS ID 1}
    |-- {ANALYSIS ID 1}
    +-- {ANALYSIS ID n}