linkbreakers 1.195.11

Official Rust SDK for the Linkbreakers API
Documentation
/*
 * Linkbreakers API
 *
 * This is a documentation of all the APIs of Linkbreakers  ## Entry points  `GET https://api.linkbreakers.com/` and `GET https://api.linkbreakers.com/v1` answer without credentials with the API's name, version and links. This document is served at `https://api.linkbreakers.com/openapi.json`.  ## Rate limits  Every response describes where the caller stands in the IETF `RateLimit-Policy` and `RateLimit` header fields (draft-ietf-httpapi-ratelimit-headers), and in `RateLimit-Limit`, `RateLimit-Remaining`, `RateLimit-Reset` and `X-RateLimit-Limit`, `X-RateLimit-Remaining`, `X-RateLimit-Reset` (a Unix timestamp) for clients that read those. A request past the limit gets `429 Too Many Requests` with `Retry-After` in seconds.  - Workspace API token: 720 requests per 60 seconds, per token. - Dashboard session or OAuth token: 600 requests per 60 seconds, per token. - No credentials: 300 requests per 60 seconds, per client IP.  Each plan also includes a number of API calls per billing cycle. Past it, calls answer `429` with the code `MAX_PUBLIC_API_QUERIED_REACHED` until the cycle renews or the plan changes.  ## Pagination  Every list pages with opaque cursors (AIP-158). Pass `pageSize`, then send the `nextPageToken` of each response back as `pageToken` (or its alias `cursor`) to get the next page. An empty or absent `nextPageToken` means the list is complete. Each list operation describes this in `x-pagination`.  ## Batches and background jobs  `POST /v1/links/bulk` creates up to 250 links in one transactional request; `POST /v1/links/batch` takes the same links as a bare JSON array. `POST /v1/links/bulk/jobs` accepts the same batch and answers `202 Accepted` at once with a job and a `Location` header; poll `GET /v1/jobs/{id}` until `status` is `JOB_STATUS_SUCCEEDED` (the created link ids are in `linksBulkCreate.linkIds`) or `JOB_STATUS_FAILED` (`error` says why, and nothing was created). Jobs can be read for 24 hours after they finish.  ## Versioning and deprecation  The version is in the path: every operation lives under `/v1`. Within a version, changes are additive only: new operations, new optional request fields, new response fields and new enum values. Clients should ignore fields and enum values they do not know. Anything that would break an existing client ships under a new path version instead.  An operation is retired in two steps. First it is deprecated: it keeps working, is flagged `deprecated: true` in this document, and every response carries a `Deprecation` header field (RFC 9745) with the time it was deprecated and a `Link` with `rel=\"deprecation\"` to the policy. Then, at least 180 days later, it may be removed; that date is announced in advance in a `Sunset` header field (RFC 8594) and as `x-sunset` on the operation. The full policy: https://linkbreakers.com/help/article/api-versioning-and-deprecation-policy  ## Idempotent retries  Every POST, PUT and PATCH accepts an optional `Idempotency-Key` header (draft-ietf-httpapi-idempotency-key-header). Send a new unique value, such as a UUID, with each logical request and reuse it only to retry that request. A retry with the same key and the same request returns the first response, marked `Idempotent-Replayed: true`, instead of running again. Keys are scoped to the token and kept for 24 hours. The same key with a different request is refused with `422`, and a retry while the first request is still running gets `409` with `Retry-After`. Server errors (5xx) and refusals such as 401, 403 and 429 are not stored, so retrying them runs the request again.  ## Identifying your client  Send a `Linkbreakers-Client` request header saying which client made the call: `<client>/<version> (<details>)`, for example `my-integration/1.2.0 (python 3.12; linux)`. The official SDKs, the CLI and the MCP bridge set it for you, as `sdk-python/2.3.1 (python 3.12.4; linux)` or `cli/1.4.0 (darwin arm64)`. It is optional and changes nothing about how a request is handled; it tells us which clients and versions are in use, so we know what can be retired and what to fix first. Keep the details short and free of personal data. Browsers may send it too: it is allowed by CORS.
 *
 * The version of the OpenAPI document: 1.195.11
 * 
 * Generated by: https://openapi-generator.tech
 */

use crate::models;
use serde::{Deserialize, Serialize};

/// TrafficSplitPayload : Traffic split payload for EVENT_ACTION_TRAFFIC_SPLIT, Linkbreakers' A/B test. Invisible to the visitor: no page, the visit goes straight to the variant's next step. 2 to 20 variants, weights sum to exactly 100, there is no else branch because the weights cover every visitor.  Sticky is the default: assignment is deterministic and per device. The variant is computed from the step id and the device id, nothing is stored, so the same device gets the same variant on every visit. A different split on the same device is an independent assignment. The device is the workspace's device cookie, so a new device, private browsing or some in-app browsers count as a new visitor.  With sticky set to false, the variant is computed from the step id and the visit's event id instead: every visit is assigned afresh, so one person can see several variants, while one visit stays on its variant across a form or a retry. That suits sharing traffic. For an A/B test, leave sticky on.  Percentages are approximate at low traffic: on a 50/50, expect about plus or minus 10 points at 100 visitors, 3 at 1,000 and 1 at 10,000.  Changing weights only moves visitors near the boundaries you changed: going from 90/10 to 80/20 only moves visitors from the first variant to the second. With three or more variants, a shift can move a few more than the difference. Keep variants in the same order, since the order decides who lands where.  To end a test, set the weights to 100 and 0: everyone goes to the winner, including those who had the other variant. To target a test, place a condition before the split, not inside it: a split has no rules of its own.
#[derive(Clone, Default, Debug, PartialEq, Serialize, Deserialize)]
pub struct TrafficSplitPayload {
    /// Keep returning visitors on the same variant. Left out means true, the default, and for an A/B test it should stay on. Set false only to share traffic: each visit is then assigned on its own by weight, so one person can see several variants. It is a draw per visit, not a strict round-robin. On an update, leaving this out keeps the current setting.
    #[serde(rename = "sticky", skip_serializing_if = "Option::is_none")]
    pub sticky: Option<bool>,
    /// The variants, from 2 to 20. Their order decides who lands where, so keep it stable across updates.
    #[serde(rename = "variants", skip_serializing_if = "Option::is_none")]
    pub variants: Option<Vec<models::TrafficSplitVariant>>,
}

impl TrafficSplitPayload {
    /// Traffic split payload for EVENT_ACTION_TRAFFIC_SPLIT, Linkbreakers' A/B test. Invisible to the visitor: no page, the visit goes straight to the variant's next step. 2 to 20 variants, weights sum to exactly 100, there is no else branch because the weights cover every visitor.  Sticky is the default: assignment is deterministic and per device. The variant is computed from the step id and the device id, nothing is stored, so the same device gets the same variant on every visit. A different split on the same device is an independent assignment. The device is the workspace's device cookie, so a new device, private browsing or some in-app browsers count as a new visitor.  With sticky set to false, the variant is computed from the step id and the visit's event id instead: every visit is assigned afresh, so one person can see several variants, while one visit stays on its variant across a form or a retry. That suits sharing traffic. For an A/B test, leave sticky on.  Percentages are approximate at low traffic: on a 50/50, expect about plus or minus 10 points at 100 visitors, 3 at 1,000 and 1 at 10,000.  Changing weights only moves visitors near the boundaries you changed: going from 90/10 to 80/20 only moves visitors from the first variant to the second. With three or more variants, a shift can move a few more than the difference. Keep variants in the same order, since the order decides who lands where.  To end a test, set the weights to 100 and 0: everyone goes to the winner, including those who had the other variant. To target a test, place a condition before the split, not inside it: a split has no rules of its own.
    pub fn new() -> TrafficSplitPayload {
        TrafficSplitPayload {
            sticky: None,
            variants: None,
        }
    }
}