Return the JSON Schema a climate processing config must satisfy.

Call this before process_climate_data to see the required and optional fields: input_csv, date_column, metrics, rolling_window_days, plot, summary.

List CSV files available under data/, with their column names.

Use one of the returned "filename" values as input_csv in a config passed to process_climate_data.

Validate a climate processing config against the schema, without running it.

Returns {"valid": true} or {"valid": false, "errors": [...]}.

Run the climate processing pipeline on an inline config and return the results.

config must satisfy the schema returned by get_config_schema. input_csv is resolved against the server's data/ directory (see list_sample_data for available files); the output_path fields under plot and summary are treated as filenames only -- every run writes to its own directory, so a config cannot choose where on disk anything is written.

missing_policy is optional: omit it and each metric uses the pipeline's default (interpolate for temperature_c, zero_fill for precipitation_mm). Set it per metric to interpolate, zero_fill, drop or fail. The right choice depends on the question being asked, not on the data alone: temperature is averaged over the month, so an estimated day washes out, while precipitation is summed, so a zero-filled day lowers the total permanently.

Returns a text report (row count, per-metric data quality, monthly summary table) followed by the rendered plot image. The data-quality section gives the missing count, coverage and longest consecutive gap for each metric, so the effect of the chosen policy is visible in the result.