| { | |
| "detime": { | |
| "packages": [ | |
| "detime" | |
| ], | |
| "snippet": "from detime import DecompositionConfig, decompose\nfrom detime.io import read_series\n\nseries = read_series(\"examples/data/example_series.csv\", col=\"value\")\nresult = decompose(\n series,\n DecompositionConfig(method=\"SSA\", params={\"window\": 24, \"rank\": 6, \"primary_period\": 12}),\n)\n\nsummary = {\n \"trend_std\": float(result.trend.std()),\n \"season_std\": float(result.season.std()),\n \"component_count\": len(result.components),\n}\n", | |
| "step_count": 12 | |
| }, | |
| "narrative": [ | |
| "The specialist workflow mixes multiple imports, result objects, and package-specific APIs.", | |
| "The DeTime workflow keeps the same result contract and method switch path under one import surface." | |
| ], | |
| "specialist_glue": { | |
| "packages": [ | |
| "pandas", | |
| "statsmodels", | |
| "ssalib" | |
| ], | |
| "snippet": "import pandas as pd\nfrom statsmodels.tsa.seasonal import STL\nfrom ssalib import SingularSpectrumAnalysis\n\nframe = pd.read_csv(\"examples/data/example_series.csv\")\nseries = frame[\"value\"].to_numpy()\n\nstl = STL(series, period=12).fit()\nssa = SingularSpectrumAnalysis(window_size=24, rank=6)\nssa_components = ssa.fit_transform(series)\n\nsummary = {\n \"trend_std\": float(stl.trend.std()),\n \"season_std\": float(stl.seasonal.std()),\n \"ssa_components\": int(ssa_components.shape[1]),\n}\n", | |
| "step_count": 13 | |
| } | |
| } | |