Record provenance for an Agentic MIKE+ step via the mike-plus MCP server — a manifest.json tying the model copy, inputs, edits, run QA status, result files and figures together, each file with a sha256. Use at the end of every run/compare/plot task (successful or failed) so the work can be audited, re-compared or reproduced later. No license required.
インストールする前に、エージェントに与えられる指示の中身を確認できます。
One JSON per case, written last. It is what makes "we ran X and got Y" checkable a month later, and it is the record a platform layer (SWMM / MIKE+ / LSTM side by side) can index.
mike_manifest_write — {out_json, case?, model?, inputs?[], results?[], figures?[], run?, edits?[], tool_calls?[], notes?} → writes the manifest and returns {manifest, missing_files[], n_files, run_status}.
model: the .sqlite COPY that was run.inputs: rain .dfs0 (incl. ones made by mike_rain_to_dfs0), CSVs, hotstart files.results: the result_files returned by mike_run (or the two res1d of a compare).figures: every PNG from mike_plot_*.run: pass the whole mike_run return; the QA fields (status, completed, errors, warnings, issues, log_file, elapsed_s, active_simulation, result_files) are kept, the log tail is dropped.edits: the mike_set_values returns (they carry table / muids / before / after).tool_calls: ordered {tool, args, ok} records of what you called.notes: the user's plain-language goal + anything not done and why.schema: agentic-mike/manifest/1)created_utc, engine=mikeplus, case,
model{path, exists, bytes, sha256}, inputs[], results[], figures[] (same shape),
run{status, errors, warnings, ...}, edits[], tool_calls[], notes,
environment{python, platform, mikeplus_mcp}
run.status, errors and notes are the evidence of why.missing_files must be [] before you say the case is complete; a listed path means you referenced something that is not on disk.runs/<case>/manifest.json; one case = one manifest (overwrite on re-run of the same case, or use a new case name).... mike_run -> mike_results_* -> mike_plot_* ...
mike_manifest_write {out_json: runs/<case>/manifest.json, model, inputs, results, figures, run, edits, notes}
まだレビューはありません。使ってみた感想をお寄せください。
概要と使いどころ
Compare two MIKE+ runs via the mike-plus MCP server — baseline vs scenario, before vs after a parameter edit — to say WHERE a change mattered and BY HOW MUCH (delta peak, delta volume, timing shift, RMSE/NSE) and to draw the overlay hydrograph. Use whenever two .res1d files of the same model exist or a re-run just finished. Comparing needs no license; producing the second run does.
日本語の概要は準備中です。原文の説明を表示しています。
Standard operating procedure for an auditable, headless MIKE+ workflow via the mike-plus MCP server — how to plan, copy, run, QA-gate, compare, plot, audit, and safely stop. Use FIRST whenever an agent is handed a MIKE+ model (.sqlite) or result (.res1d) plus a plain-language goal; it picks the operating mode and tells the agent which mike-* skill to open next and in what order.
日本語の概要は準備中です。原文の説明を表示しています。
Bring an EPA SWMM (.inp) or EPANET (.inp) model into MIKE+ via the mike-plus MCP server — the bridge from an INP builder such as SWMMCanada or agentic-swmm into a MIKE+ database that mike_model_info / mike_run can use. Use when the user has an .inp and wants it in MIKE+. Requires MIKE+ installed + a checked-out license (demo mode caps the model size).
日本語の概要は準備中です。原文の説明を表示しています。
Inspect a MIKE+ .sqlite model via the mike-plus MCP server — active simulation/scenario/model type, unit system, the list of simulation setups and scenarios, element counts — and switch the active scenario. Use when an agent needs to understand a MIKE+ model before running or editing it, or must run a scenario prepared in the GUI. Requires MIKE+ installed + a license.
日本語の概要は準備中です。原文の説明を表示しています。
Read and modify MIKE+ model parameters via the mike-plus MCP server — inspect or change values in any model table (pipe Diameter/Manning in msm_Link, node levels in msm_Node, catchment properties in msm_Catchment, etc.). Use when an agent must identify a current parameter and change it before re-running. Requires MIKE+ + license; mutating tools must run on a copy.
日本語の概要は準備中です。原文の説明を表示しています。
Produce publication-style figures from MIKE+ .res1d results via the mike-plus MCP server — a stacked rainfall-runoff hydrograph, a single time series, a network map, a baseline-vs-scenario overlay, or a longitudinal profile (bed / crown / ground / max water level along a pipe string), in the house style (Arial 12, ticks inward, SI units, no title). Use when an agent must render MIKE+ results as a PNG without opening the GUI. No license required.
日本語の概要は準備中です。原文の説明を表示しています。