this is so i will remember to continue here next week, I will add code example via PR later on next week
{
"type": "object",
"description": "Forecasted metric values at this budget level. Keys are either forecastable-metric values for delivery/engagement (impressions, reach, spend, etc.) or event-type values for outcomes (purchase, lead, app_install, etc.). Values are ForecastRange objects (low/mid/high). Use { \"mid\": value } for point estimates. Include spend when the platform predicts it will differ from budget.",
"additionalProperties": {
"title": "Forecast Range",
"description": "A forecast value with optional low/high bounds. The mid value represents the most likely outcome. When low and high are provided, they represent conservative and optimistic estimates respectively.",
"type": "object",
"properties": {
"low": {
"type": "number",
"description": "Conservative (low-end) forecast value",
"minimum": 0
},
"mid": {
"type": "number",
"description": "Expected (most likely) forecast value",
"minimum": 0
},
"high": {
"type": "number",
"description": "Optimistic (high-end) forecast value",
"minimum": 0
}
},
"required": [
"mid"
],
"additionalProperties": true
}
}
this is so i will remember to continue here next week, I will add code example via PR later on next week
but something like this:
will cause dtos that cannot represent the data. Maybe only if it is used with $ref i will double check and create an atomic example next week. Is used here:
https://github.com/adcontextprotocol/adcp/blob/main/dist/schemas/3.0.0-beta.3/bundled/media-buy/get-products-response.json#L1926
is removed in rc2 of the schema, but other places, maybe, still use this functionality or similar one e.g. here somewhat similar:
https://github.com/adcontextprotocol/adcp/blob/main/dist/schemas/3.0.0-rc.2/bundled/media-buy/package-request.json#L7533