Outputs

7) Outputs

KiMecO writes both text outputs and SQLite databases.

7.1 Text output files

File	Location	Content
KiMecO.log	Run launch directory (where kmo input.json is started)	Main run log. Contains progress messages, validation and initialization status, timing information, warnings, and errors (errors now include the full Python traceback). Existing log files are rotated with a date suffix when a new run starts with the same log filename.
score_info.txt	KMO_Project/ (or your configured project_name)	Per-generation score summary for GOAT tracking. First line is a header (ITER, BEST SCORE, GOAT AVERAGE), followed by one row per generation with the best score and mean GOAT score.
goats.txt	KMO_Project/	GOAT membership history by generation. Each line corresponds to one generation and stores model tokens as gen_id_model_id pairs (for example 3_42), representing the selected best models used as the GOAT ensemble for that generation.
GA_rates.out	KMO_Project/	Rate-statistics report written at GA convergence. Contains the selected model count after max_score filtering, then pressure/temperature-resolved tables of reaction rates (organized by PES), with geometric-mean and geometric-standard-deviation summaries across eligible models.
{name}P{slot:02d}.out	Per-job work directory	Final MESS rate-coefficient output for a PES slot.
_{name}P{slot:02d}.out	Per-job work directory	Preserved MESS pass-1 output, written only when use_automech=true. It is the first-pass result kept before the automatic WellExtension well-lumping produces the extended pass-2 input whose result becomes {name}P{slot:02d}.out.

7.2 Databases (run state and results)

Three SQLite databases are created in KMO_Project/ and together contain the accumulated data necessary for restart and analysis. The databases contain one table per generation/iteration with names such as G0001 for the first generation of a GA, NM0001 for the first Nelder-Mead simplex and NMS0001 for the first Nelder-Mead Swarm simplexes.

Database	File name	Content
SOP_DB	KMO_DB_SOP.db	Contains the set of parameters (SOPs) for all models. Each row corresponds to one model and stores the full SOP parameter vector used to build kinetics/simulations. The columns are dynamically built depending on the system's parameters. Also contains SOP metadata (__sop_item_pes_ids) that maps SOP items to PES IDs to enforce restart consistency.
KIN_DB	KMO_DB_KIN.db	Contain the rate coefficients. Rows store pressure (P), temperature (T), model kinetics id (kin_id) , pes_id, from_name, to_name, and rate coefficient k. Tables are organized per generation/iterations as described above.
SIM_DB	KMO_DB_SIM.db	Simulation outputs for each model and experiment pair. Rows are keyed by (mdl_id, experiment_id) and store a binary result blob (Feather-encoded profile data). Currently, the content of the decoded result includes time and species profiles used for scoring and postprocessing. However, in the future, it will depend on the specific experimental data provided for the scoring.

Postprocessing writes into these same databases. Extrapolation no longer creates separate PP_DB_KIN.db / PP_DB_SIM.db files. Extrapolated rate coefficients are appended into the existing KIN tables for the originating generation (reusing any (P, T) already computed), and extrapolated simulations are appended into the existing SIM tables with banded experiment ids (offset past the original run's experiments) so nothing is overwritten. In the kmoui dashboard these appear from the unified SIM database and are labelled Extrapolated (band b) — <experiment metadata>.

The content of these databases is accessible from the Databases tab of the analysis GUI (kmoui): select one of the .db files in the run folder, pick a table to display it as a dataframe, and use the copy button to copy the table to the clipboard (tab separated) for pasting into a spreadsheet such as Excel.