ats

package
v0.17.0 Latest Latest
Warning

This package is not in the latest version of its module.

Go to latest
Published: Oct 25, 2026 License: MIT Imports: 43 Imported by: 0

Documentation

Overview

Package ats unifies the ATS-backed providers behind one company-parameterized search interface, so MCP clients name a company and never learn which ATS serves it.

Index

Constants

View Source
const (
	// DefaultDumpCacheTTL is the default write-expiry for board dumps.
	DefaultDumpCacheTTL = 10 * time.Minute
	// DefaultDumpCacheMaxEntries is the default max number of cached boards.
	DefaultDumpCacheMaxEntries = 100
)

Variables

This section is empty.

Functions

This section is empty.

Types

type ADPMyJobsAdapter added in v0.14.0

type ADPMyJobsAdapter struct {
	// contains filtered or unexported fields
}

ADPMyJobsAdapter serves public ADP MyJobs career boards (myjobs.adp.com). Keyword search uses the upstream OData $search parameter with server-side pagination, and filters use the tenant's own dimensions, both applied in the same request. Nothing is matched locally and no board dump is needed.

The dimensions are whatever the tenant configured — State and City on one board, Store or Brand or Compensation Range on another — so Adapter.Filters reports them per company and SearchParams.Location is not supported: a board that files its jobs by store cannot answer a free-text place. Workforce Now (workforcenow.adp.com) is out of scope (future adp_wfn).

func NewADPMyJobsAdapter added in v0.14.0

func NewADPMyJobsAdapter(hc *http.Client, dumpCache *DumpCache) *ADPMyJobsAdapter

NewADPMyJobsAdapter builds an adapter using the shared HTTP client. dumpCache is reused to hold each board's filter catalog rather than any board dump (see facets); nil disables that caching without changing behaviour.

func (*ADPMyJobsAdapter) CareersURL added in v0.14.0

func (a *ADPMyJobsAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster company's public MyJobs page.

func (*ADPMyJobsAdapter) Detail added in v0.14.0

func (a *ADPMyJobsAdapter) Detail(ctx context.Context, slug, jobID string) (*JobDetail, error)

func (*ADPMyJobsAdapter) Filters added in v0.14.0

func (a *ADPMyJobsAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*ADPMyJobsAdapter) Name added in v0.14.0

func (a *ADPMyJobsAdapter) Name() string

func (*ADPMyJobsAdapter) ParseCareersURL added in v0.14.0

func (a *ADPMyJobsAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes myjobs.adp.com/{slug}/... boards only.

func (*ADPMyJobsAdapter) Roster added in v0.14.0

func (a *ADPMyJobsAdapter) Roster() []CompanyInfo

func (*ADPMyJobsAdapter) Search added in v0.14.0

type Adapter

type Adapter interface {
	// Name reports the provider key used in registry diagnostics and URL guidance.
	Name() string
	// Roster lists every curated company on this ATS.
	Roster() []CompanyInfo
	// ParseCareersURL recognizes a careers URL for this ATS and returns a slug
	// accepted by [Adapter.Search], [Adapter.Filters], and [Adapter.Detail].
	// For curated companies, it returns [CompanyInfo.Slug]. For example,
	// Workday returns a canonical careers URL for unlisted tenants, since a
	// tenant name alone cannot identify its site. It returns ("", false) when
	// u is not a careers URL for this ATS.
	ParseCareersURL(u *url.URL) (slug string, ok bool)
	// CareersURL renders the public careers page for a curated company.
	// slug must come from [Adapter.Roster]; it returns ok=false when this ATS
	// has no stable public URL for that company.
	CareersURL(slug string) (url string, ok bool)
	// Search returns one page of jobs for the company identified by slug.
	// Pass a result's [JobSummary.JobID] and the same slug to [Adapter.Detail].
	Search(ctx context.Context, slug string, p SearchParams) (*SearchResult, error)
	// Filters returns the filter dimensions and values available for the company
	// identified by slug. Pass its keys and values through
	// [SearchParams.Filters] on a later call to [Adapter.Search].
	Filters(ctx context.Context, slug string) (FilterSet, error)
	// Detail returns the full posting identified by [JobSummary.JobID].
	// jobID must come from an [Adapter.Search] performed with the same slug.
	Detail(ctx context.Context, slug, jobID string) (*JobDetail, error)
}

Adapter is one ATS's implementation of the unified search interface. Methods address a company by slug: either one declared by Adapter.Roster and indexed by Registry, or one the adapter itself minted via Adapter.ParseCareersURL for careers-URL input.

type AshbyAdapter

type AshbyAdapter struct {
	// contains filtered or unexported fields
}

AshbyAdapter serves Ashby-hosted companies. Ashby's public API is a single full-board endpoint — no server-side search and no per-job endpoint (returns 401) — so Search filters the dump via searchDump and Detail scans the same dump. When the process-local dump cache is enabled, Search/Filters/Detail for one board share a single upstream fetch.

func NewAshbyAdapter

func NewAshbyAdapter(baseURL string, hc *http.Client, dumpCache *DumpCache) (*AshbyAdapter, error)

func (*AshbyAdapter) CareersURL added in v0.12.0

func (a *AshbyAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster company's public job board page.

func (*AshbyAdapter) Detail

func (a *AshbyAdapter) Detail(ctx context.Context, slug, jobID string) (*JobDetail, error)

func (*AshbyAdapter) Filters

func (a *AshbyAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*AshbyAdapter) Name

func (a *AshbyAdapter) Name() string

func (*AshbyAdapter) ParseCareersURL added in v0.5.0

func (a *AshbyAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes Ashby-hosted board URLs; the first path segment is the organization name, which is already this adapter's slug form (URL-decoded — Ashby org names may contain spaces).

func (*AshbyAdapter) Roster

func (a *AshbyAdapter) Roster() []CompanyInfo

func (*AshbyAdapter) Search

func (a *AshbyAdapter) Search(ctx context.Context, slug string, p SearchParams) (*SearchResult, error)

type AvatureAdapter added in v0.10.0

type AvatureAdapter struct {
	// contains filtered or unexported fields
}

AvatureAdapter serves public Avature career portals. Search and detail are server-rendered HTML (see internal/provider/avature/doc.go). Roster slugs are portal bases without the scheme (e.g. "koch.avature.net/careers" — one tenant can host several portals, so the portal name is part of the key). ParseCareersURL also accepts any *.avature.net portal URL so uncurated tenants work when passed as a careers URL; custom-domain portals (e.g. careers.unifiservice.com) resolve through the roster only.

func NewAvatureAdapter added in v0.10.0

func NewAvatureAdapter(hc *http.Client) *AvatureAdapter

func (*AvatureAdapter) CareersURL added in v0.12.0

func (a *AvatureAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster company's public job search page.

func (*AvatureAdapter) Detail added in v0.10.0

func (a *AvatureAdapter) Detail(ctx context.Context, slug, jobID string) (*JobDetail, error)

func (*AvatureAdapter) Filters added in v0.10.0

func (a *AvatureAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

Filters reports no dimensions: portal facets use per-tenant numeric field ids whose options load only via JS autocomplete, so there is nothing portable to offer.

func (*AvatureAdapter) Name added in v0.10.0

func (a *AvatureAdapter) Name() string

func (*AvatureAdapter) ParseCareersURL added in v0.10.0

func (a *AvatureAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes any *.avature.net portal URL, dropping an optional locale segment (e.g. /en_US/careers/SearchJobs and /careers/JobDetail/... both yield "<host>/careers"). A bare host is rejected: the portal name cannot be inferred.

func (*AvatureAdapter) Roster added in v0.10.0

func (a *AvatureAdapter) Roster() []CompanyInfo

func (*AvatureAdapter) Search added in v0.10.0

func (a *AvatureAdapter) Search(ctx context.Context, slug string, p SearchParams) (*SearchResult, error)

Search maps the unified page onto jobOffset requests. The portal page size is tenant-configured and cannot be raised, so one unified page may stitch several consecutive upstream fetches; jobOffset accepts arbitrary offsets, which keeps the stitch a plain cursor walk.

type BambooHRAdapter added in v0.10.0

type BambooHRAdapter struct {
	// contains filtered or unexported fields
}

BambooHRAdapter serves BambooHR hosted careers sites. The public /careers/list endpoint returns the complete board in one response but only as sparse summaries - descriptions, compensation, and posting dates live on the per-job /careers/{id}/detail endpoint. Filters and unqueried Search run searchDump over the list dump alone (results carry no posted date). When Search receives a non-empty query, the adapter fans out to /careers/{id}/detail so tier-3 skill/technology matching can see full JDs. Detail hits the per-job endpoint directly.

func NewBambooHRAdapter added in v0.10.0

func NewBambooHRAdapter(hc *http.Client, dumpCache *DumpCache) *BambooHRAdapter

NewBambooHRAdapter derives a redirect-blocking copy of hc: BambooHR 302-redirects unknown tenants to its marketing site, and following that redirect would turn a diagnosable "no such tenant" into an HTML decode error.

func (*BambooHRAdapter) CareersURL added in v0.12.0

func (a *BambooHRAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster company's public careers page.

func (*BambooHRAdapter) Detail added in v0.10.0

func (a *BambooHRAdapter) Detail(
	ctx context.Context,
	slug string,
	jobID string,
) (*JobDetail, error)

func (*BambooHRAdapter) Filters added in v0.10.0

func (a *BambooHRAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*BambooHRAdapter) Name added in v0.10.0

func (a *BambooHRAdapter) Name() string

func (*BambooHRAdapter) ParseCareersURL added in v0.10.0

func (a *BambooHRAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes BambooHR subdomain careers pages.

func (*BambooHRAdapter) Roster added in v0.10.0

func (a *BambooHRAdapter) Roster() []CompanyInfo

func (*BambooHRAdapter) Search added in v0.10.0

func (a *BambooHRAdapter) Search(
	ctx context.Context,
	slug string,
	p SearchParams,
) (*SearchResult, error)

type CompanyCandidate added in v0.12.0

type CompanyCandidate struct {
	Name       string
	CareersURL string
	Provider   string
}

CompanyCandidate is one human-readable choice in a CompanyResolution. CareersURL is empty when the owning adapter has no public URL for this entry; Provider is carried for logs and debug tooling and is never rendered in MCP-facing text.

type CompanyInfo

type CompanyInfo struct {
	Slug string // unique key; the provider roster's tenant/site/board slug
	Name string // display name; the resolver also matches on it
}

CompanyInfo is one company as the registry sees it: enough to resolve a user-supplied name to (adapter, slug). Connection config (Workday tenant/instance/site etc.) stays inside each adapter, looked up by slug.

type CompanyResolution added in v0.13.0

type CompanyResolution struct {
	// contains filtered or unexported fields
}

CompanyResolution is the successful result of resolving a company input. It may contain one exact match or multiple choices that the caller must disambiguate with the user.

func (CompanyResolution) Candidates added in v0.13.0

func (r CompanyResolution) Candidates() []CompanyCandidate

Candidates returns the human-readable choices in registration order.

func (CompanyResolution) IsAmbiguous added in v0.13.0

func (r CompanyResolution) IsAmbiguous() bool

IsAmbiguous reports whether the input matched more than one company.

func (CompanyResolution) Select added in v0.13.0

func (r CompanyResolution) Select(index int) (Adapter, string, bool)

Select returns the adapter and provider-local slug for a candidate index.

func (CompanyResolution) Single added in v0.13.0

func (r CompanyResolution) Single() (Adapter, string, bool)

Single returns the adapter and provider-local slug when exactly one company matched.

type DayforceAdapter added in v0.15.0

type DayforceAdapter struct {
	// contains filtered or unexported fields
}

DayforceAdapter serves Dayforce (Ceridian) candidate portal boards. Search, its filter catalogs, and job detail are all server-side JSON behind a next-auth CSRF pre-flight (see dayforce.BoardClient); only the board's public identity page is HTML. Roster slugs are tenant namespaces, suffixed with the job board code for a non-default board (see dayforce.Company.Slug); ParseCareersURL mints a canonical slug for non-roster boards. jobBoardId is resolved lazily via SiteInfo on the paths that need it (Filters, Detail, filtered Search).

func NewDayforceAdapter added in v0.15.0

func NewDayforceAdapter(baseURL string, hc *http.Client) (*DayforceAdapter, error)

NewDayforceAdapter creates a Dayforce ATS adapter. baseURL is the candidate portal API origin (production: dayforceJobsHost); tests point it at a mock server.

func (*DayforceAdapter) CareersURL added in v0.15.0

func (a *DayforceAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster company's public candidate portal page.

func (*DayforceAdapter) Detail added in v0.15.0

func (a *DayforceAdapter) Detail(ctx context.Context, slug, jobID string) (*JobDetail, error)

func (*DayforceAdapter) Filters added in v0.15.0

func (a *DayforceAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*DayforceAdapter) Name added in v0.15.0

func (a *DayforceAdapter) Name() string

func (*DayforceAdapter) ParseCareersURL added in v0.15.0

func (a *DayforceAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL accepts current jobs.dayforcehcm.com board or job URLs and legacy <region>.dayforcehcm.com CandidatePortal URLs.

Search has two entry points that cannot share one slug shape. A curated row keeps namespace, board, and culture in YAML, so its Search key is the short dayforce.Company.Slug (usually just the namespace). A careers URL for a board that is not in the roster has nowhere else to put those three fields, so this method mints "ns/board" and appends "/culture" only when culture is not en-US — [DayforceAdapter.resolveSlug] splits that form back apart. Roster matching is the namespace+board fields, not slug strings: a hit returns Company.Slug() (e.g. "pca"); "pca" and "pca/candidateportal" are never compared to each other.

Examples:

Non-Dayforce, API, asset, and incomplete board URLs return ("", false).

func (*DayforceAdapter) Roster added in v0.15.0

func (a *DayforceAdapter) Roster() []CompanyInfo

func (*DayforceAdapter) Search added in v0.15.0

func (a *DayforceAdapter) Search(ctx context.Context, slug string, p SearchParams) (*SearchResult, error)

type DumpCache added in v0.13.0

type DumpCache struct {
	// contains filtered or unexported fields
}

DumpCache is a short-TTL, size-bounded cache of full-board dumps for a single-user stdio MCP session. A nil *DumpCache means no caching (getOrLoadDump passes through to load). Cached dumps are shared read-only; callers must not mutate returned jobs or side.

func NewDumpCache added in v0.13.0

func NewDumpCache(cfg DumpCacheConfig) *DumpCache

NewDumpCache builds an enabled cache. Callers that want no caching pass nil to adapters instead of constructing one.

type DumpCacheConfig added in v0.13.0

type DumpCacheConfig struct {
	// TTL is how long each board dump is retained after write.
	// Values <= 0 become DefaultDumpCacheTTL in NewDumpCache.
	TTL time.Duration
	// MaxEntries is otter's MaximumSize (number of provider+slug boards).
	// Values <= 0 become DefaultDumpCacheMaxEntries.
	MaxEntries int
}

DumpCacheConfig configures a process-local full-dump cache. Callers that want no caching pass nil to adapters instead of constructing one.

type EightfoldAdapter added in v0.7.0

type EightfoldAdapter struct {
	// contains filtered or unexported fields
}

EightfoldAdapter serves Eightfold-hosted career sites (<tenant>.eightfold.ai). Search runs server-side: query and location are plain fuzzy-text parameters, but named facet filters (businessarea, employmenttype, city, ...) are tenant-specific, their query value often differs from the display label only in case, and the set of valid filters is unknowable ahead of time — so a filtered Search first probes an unfiltered search to resolve requested labels to values (mirrors Workday's GUID-facet probe, minus the GUIDs).

Some tenants have the pcsx API this all otherwise runs on disabled (403 "PCSX is not enabled for this user.") and serve postings exclusively through the newer, unfiltered-only v2 API instead; Search and Detail fall back to it on that specific error (see openapi.yaml's v2 tag). Filters and filtered Search still require pcsx and simply fail for v2-only tenants — narrower than not supporting them at all, since plain search still works.

Roster membership is required, unlike Workday/Greenhouse/Lever: every PCSX request needs the tenant's registered `domain` value alongside its subdomain, and that value can't be derived from a careers URL alone or safely guessed — a wrong domain gets the site's HTML shell back, not a clean error (see companies.yaml and provider/eightfold/openapi.yaml).

func NewEightfoldAdapter added in v0.7.0

func NewEightfoldAdapter(hc *http.Client) *EightfoldAdapter

NewEightfoldAdapter takes an *http.Client already wrapped in eightfold.BrowserTransport — Eightfold's edge 403s Go's default User-Agent instead of returning JSON.

func (*EightfoldAdapter) CareersURL added in v0.12.0

func (a *EightfoldAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster company's public career site.

func (*EightfoldAdapter) Detail added in v0.7.0

func (a *EightfoldAdapter) Detail(ctx context.Context, slug, jobID string) (*JobDetail, error)

func (*EightfoldAdapter) Filters added in v0.7.0

func (a *EightfoldAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*EightfoldAdapter) Name added in v0.7.0

func (a *EightfoldAdapter) Name() string

func (*EightfoldAdapter) ParseCareersURL added in v0.7.0

func (a *EightfoldAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes <tenant>.eightfold.ai career pages, but only resolves tenants already on the roster — see the domain-derivation note on EightfoldAdapter.

func (*EightfoldAdapter) Roster added in v0.7.0

func (a *EightfoldAdapter) Roster() []CompanyInfo

func (*EightfoldAdapter) Search added in v0.7.0

type EngageAdapter added in v0.12.0

type EngageAdapter struct {
	// contains filtered or unexported fields
}

EngageAdapter serves companies hosted on engage (エン・ジャパン), the free ATS whose tenants each get a careers site at en-gage.net/<slug>/. One request returns a tenant's board, so search and filtering are implemented over that dump; the board carries no posting text, so Detail fetches the posting page.

The board is a capped dump, not a complete one: engage returns at most engage.CategoryCap jobs per employment category and offers no pagination. For a tenant over that ceiling, Search results and TotalCount are a lower bound on the true board.

func NewEngageAdapter added in v0.12.0

func NewEngageAdapter(hc *http.Client, dumpCache *DumpCache) *EngageAdapter

func (*EngageAdapter) CareersURL added in v0.12.0

func (a *EngageAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster company's public engage tenant board.

func (*EngageAdapter) Detail added in v0.12.0

func (a *EngageAdapter) Detail(ctx context.Context, slug, jobID string) (*JobDetail, error)

Detail fetches the posting page, because the board lists jobs without their descriptions.

func (*EngageAdapter) Filters added in v0.12.0

func (a *EngageAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*EngageAdapter) Name added in v0.12.0

func (a *EngageAdapter) Name() string

func (*EngageAdapter) ParseCareersURL added in v0.12.0

func (a *EngageAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes en-gage.net tenant pages, both the board and a single posting. Like herp, it accepts slugs outside the curated roster: one request settles whether a slug resolves. Site-owned paths (/user/, /search/, sitemaps) are rejected, since they are not tenants.

func (*EngageAdapter) Roster added in v0.12.0

func (a *EngageAdapter) Roster() []CompanyInfo

func (*EngageAdapter) Search added in v0.12.0

func (a *EngageAdapter) Search(ctx context.Context, slug string, p SearchParams) (*SearchResult, error)

type FilterSet

type FilterSet map[string][]string

FilterSet maps a filter dimension to display labels accepted by SearchParams.Filters. Values are tenant-specific and discovered through Adapter.Filters.

type GreenhouseAdapter

type GreenhouseAdapter struct {
	// contains filtered or unexported fields
}

GreenhouseAdapter serves Greenhouse-hosted companies. The Job Board API has no server-side search, so Search fetches the whole board — with content=true, since JD text and department/office dimensions only exist on the content variant — and filters it via searchDump. Detail uses the per-job endpoint, one light request.

func NewGreenhouseAdapter

func NewGreenhouseAdapter(baseURL string, hc *http.Client, dumpCache *DumpCache) (*GreenhouseAdapter, error)

func (*GreenhouseAdapter) CareersURL added in v0.12.0

func (a *GreenhouseAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster company's public job board page.

func (*GreenhouseAdapter) Detail

func (a *GreenhouseAdapter) Detail(ctx context.Context, slug, jobID string) (*JobDetail, error)

func (*GreenhouseAdapter) Filters

func (a *GreenhouseAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*GreenhouseAdapter) Name

func (a *GreenhouseAdapter) Name() string

func (*GreenhouseAdapter) ParseCareersURL added in v0.5.0

func (a *GreenhouseAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes Greenhouse-hosted board URLs; the first path segment is the board token, which is already this adapter's slug form.

func (*GreenhouseAdapter) Roster

func (a *GreenhouseAdapter) Roster() []CompanyInfo

func (*GreenhouseAdapter) Search

type HerpAdapter added in v0.11.0

type HerpAdapter struct {
	// contains filtered or unexported fields
}

HerpAdapter serves companies listed on HERP Career, the job media that republishes the boards of HERP Hire ATS tenants. One request returns a company's whole board with full posting text, so all search, filter, and detail behavior is implemented over that dump.

func NewHerpAdapter added in v0.11.0

func NewHerpAdapter(hc *http.Client, dumpCache *DumpCache) *HerpAdapter

func (*HerpAdapter) CareersURL added in v0.12.0

func (a *HerpAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster company's public HERP Career board.

func (*HerpAdapter) Detail added in v0.11.0

func (a *HerpAdapter) Detail(ctx context.Context, slug, jobID string) (*JobDetail, error)

func (*HerpAdapter) Filters added in v0.11.0

func (a *HerpAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*HerpAdapter) Name added in v0.11.0

func (a *HerpAdapter) Name() string

func (*HerpAdapter) ParseCareersURL added in v0.11.0

func (a *HerpAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes both herp.careers URL shapes. Unlike join, it accepts slugs outside the curated roster: one request settles whether a slug resolves. The HERP Hire form is accepted for convenience even though its tenant set is wider than HERP Career's — a tenant missing from the media surfaces as the not-listed error from Search or Detail.

func (*HerpAdapter) Roster added in v0.11.0

func (a *HerpAdapter) Roster() []CompanyInfo

func (*HerpAdapter) Search added in v0.11.0

func (a *HerpAdapter) Search(ctx context.Context, slug string, p SearchParams) (*SearchResult, error)

type HrmosAdapter added in v0.12.0

type HrmosAdapter struct {
	// contains filtered or unexported fields
}

HrmosAdapter serves HRMOS 採用 (hrmos.co)-hosted companies. HRMOS has no server-side keyword search (?word=/?keyword=/?q= are byte-identical), so Search fetches a tenant's whole job dump and filters it via searchDump, the same shape as Lever and join.com.

func NewHrmosAdapter added in v0.12.0

func NewHrmosAdapter(baseURL string, hc *http.Client, dumpCache *DumpCache) *HrmosAdapter

func (*HrmosAdapter) CareersURL added in v0.12.0

func (a *HrmosAdapter) CareersURL(slug string) (string, bool)

CareersURL renders any tenant's jobs page, roster or not, for the same reason HrmosAdapter.ParseCareersURL accepts any tenant: the slug alone identifies the page, so there is nothing to look up.

func (*HrmosAdapter) Detail added in v0.12.0

func (a *HrmosAdapter) Detail(ctx context.Context, slug, jobID string) (*JobDetail, error)

func (*HrmosAdapter) Filters added in v0.12.0

func (a *HrmosAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*HrmosAdapter) Name added in v0.12.0

func (a *HrmosAdapter) Name() string

func (*HrmosAdapter) ParseCareersURL added in v0.12.0

func (a *HrmosAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes any hrmos.co tenant URL, not just curated roster ones: unlike join.com, a HRMOS slug is itself the identifier Adapter.Search and Adapter.Detail need, so no companyId lookup gates it (the Workday shape, not the join.com one).

func (*HrmosAdapter) Roster added in v0.12.0

func (a *HrmosAdapter) Roster() []CompanyInfo

func (*HrmosAdapter) Search added in v0.12.0

func (a *HrmosAdapter) Search(ctx context.Context, slug string, p SearchParams) (*SearchResult, error)

type ICIMSAdapter added in v0.8.0

type ICIMSAdapter struct {
	// contains filtered or unexported fields
}

ICIMSAdapter serves public iCIMS career portals. Search and detail are server-rendered HTML (see internal/provider/icims/openapi.yaml). Roster slugs are lowercase portal hostnames (e.g. careers-peraton.icims.com). ParseCareersURL also accepts any *.icims.com host so uncurated tenants work when passed as a careers URL.

func NewICIMSAdapter added in v0.8.0

func NewICIMSAdapter(hc *http.Client) *ICIMSAdapter

func (*ICIMSAdapter) CareersURL added in v0.12.0

func (a *ICIMSAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster company's public job search page.

func (*ICIMSAdapter) Detail added in v0.8.0

func (a *ICIMSAdapter) Detail(ctx context.Context, slug, jobID string) (*JobDetail, error)

func (*ICIMSAdapter) Filters added in v0.8.0

func (a *ICIMSAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

Filters reports the portal's category and position-type selects, resolved per tenant at call time. Tenants that do not render a select simply omit that dimension.

func (*ICIMSAdapter) Name added in v0.8.0

func (a *ICIMSAdapter) Name() string

func (*ICIMSAdapter) ParseCareersURL added in v0.8.0

func (a *ICIMSAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes any *.icims.com host. Curated roster names are preferred via the registry's name index; uncurated hosts still resolve so callers can pass a careers URL directly.

func (*ICIMSAdapter) Roster added in v0.8.0

func (a *ICIMSAdapter) Roster() []CompanyInfo

func (*ICIMSAdapter) Search added in v0.8.0

func (a *ICIMSAdapter) Search(ctx context.Context, slug string, p SearchParams) (*SearchResult, error)

type JobDetail

type JobDetail struct {
	JobID       string
	Title       string
	Company     string
	Location    string
	PostedAt    string
	URL         string
	Description string
}

JobDetail is one full posting returned by Adapter.Detail, with its description normalized to plain text.

type JobSummary

type JobSummary struct {
	// JobID is the provider-native value accepted by [Adapter.Detail], such as
	// a Workday externalPath, Lever UUID, or Ashby ID.
	JobID    string
	Title    string
	Location string
	// PostedAt is an ISO 8601 date when the upstream provides one; otherwise,
	// it is the upstream's raw text.
	PostedAt string
	// URL is the human-clickable posting page.
	URL string
}

JobSummary carries summary fields only; Adapter.Detail returns the full description, keeping search responses small for the LLM.

type JoinAdapter added in v0.10.0

type JoinAdapter struct {
	// contains filtered or unexported fields
}

JoinAdapter serves join.com-hosted companies. The public GraphQL search endpoint has no server-side keyword search (see the provider's API.md), so Search fetches a company's whole job dump and filters it via searchDump, the same shape as Greenhouse. Unlike Greenhouse, the dump carries no description — join.com's search endpoint never populates one — so Detail is a separate SSR page scrape, and searchDump's full-text tier only ever matches title and category.

func NewJoinAdapter added in v0.10.0

func NewJoinAdapter(baseURL string, hc *http.Client, dumpCache *DumpCache) *JoinAdapter

func (*JoinAdapter) CareersURL added in v0.12.0

func (a *JoinAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster company's public career page.

func (*JoinAdapter) Detail added in v0.10.0

func (a *JoinAdapter) Detail(ctx context.Context, slug, jobID string) (*JobDetail, error)

func (*JoinAdapter) Filters added in v0.10.0

func (a *JoinAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*JoinAdapter) Name added in v0.10.0

func (a *JoinAdapter) Name() string

func (*JoinAdapter) ParseCareersURL added in v0.10.0

func (a *JoinAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes join.com company page URLs. There is no public way to resolve an arbitrary slug's numeric companyId without a network call (see the provider's API.md), and this method has no context to make one with, so it only matches slugs already in the curated roster.

func (*JoinAdapter) Roster added in v0.10.0

func (a *JoinAdapter) Roster() []CompanyInfo

func (*JoinAdapter) Search added in v0.10.0

func (a *JoinAdapter) Search(ctx context.Context, slug string, p SearchParams) (*SearchResult, error)

type LeverAdapter

type LeverAdapter struct {
	// contains filtered or unexported fields
}

LeverAdapter serves Lever-hosted companies. Lever's list endpoint dumps the whole board (its native filter params are exact-match only, verified useless for fuzzy search), so searching happens in searchDump.

func NewLeverAdapter

func NewLeverAdapter(baseURL string, hc *http.Client, dumpCache *DumpCache) (*LeverAdapter, error)

func (*LeverAdapter) CareersURL added in v0.12.0

func (a *LeverAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster company's public job board page.

func (*LeverAdapter) Detail

func (a *LeverAdapter) Detail(ctx context.Context, slug, jobID string) (*JobDetail, error)

func (*LeverAdapter) Filters

func (a *LeverAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*LeverAdapter) Name

func (a *LeverAdapter) Name() string

func (*LeverAdapter) ParseCareersURL added in v0.5.0

func (a *LeverAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes Lever-hosted board URLs; the first path segment is the organization, which is already this adapter's slug form.

func (*LeverAdapter) Roster

func (a *LeverAdapter) Roster() []CompanyInfo

func (*LeverAdapter) Search

func (a *LeverAdapter) Search(ctx context.Context, slug string, p SearchParams) (*SearchResult, error)

type MokaHRAdapter added in v0.12.0

type MokaHRAdapter struct {
	// contains filtered or unexported fields
}

MokaHRAdapter serves MokaHR-hosted careers sites. Structured filters (city, category) run server-side. Text search does not: MokaHR's keyword parameter matches job titles only, so honoring the unified Query semantics means pulling the filtered board and matching locally. The list endpoint already carries every posting's full description, so that dump needs no per-job requests.

func NewMokaHRAdapter added in v0.12.0

func NewMokaHRAdapter(baseURL string, hc *http.Client) (*MokaHRAdapter, error)

func (*MokaHRAdapter) CareersURL added in v0.12.0

func (a *MokaHRAdapter) CareersURL(slug string) (string, bool)

CareersURL renders a curated site's public listing page. Slugs minted from a careers URL are not curated, so they resolve to nothing here even though the same "<org>/<siteId>" would address them upstream.

func (*MokaHRAdapter) Detail added in v0.12.0

func (a *MokaHRAdapter) Detail(ctx context.Context, slug, jobID string) (*JobDetail, error)

func (*MokaHRAdapter) Filters added in v0.12.0

func (a *MokaHRAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*MokaHRAdapter) Name added in v0.12.0

func (a *MokaHRAdapter) Name() string

func (*MokaHRAdapter) ParseCareersURL added in v0.12.0

func (a *MokaHRAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes app.mokahr.com careers sites. The captured "<org>/<siteId>" is already the slug form the roster uses, so a curated company's own URL resolves to its roster entry.

func (*MokaHRAdapter) Roster added in v0.12.0

func (a *MokaHRAdapter) Roster() []CompanyInfo

func (*MokaHRAdapter) Search added in v0.12.0

func (a *MokaHRAdapter) Search(ctx context.Context, slug string, p SearchParams) (*SearchResult, error)

type OracleAdapter added in v0.8.0

type OracleAdapter struct {
	// contains filtered or unexported fields
}

OracleAdapter serves Oracle Recruiting Cloud Candidate Experience sites. Curated companies use host/site_number roster slugs. Careers URLs outside the roster remain usable as self-describing slugs; the adapter discovers their internal site number from the public careers page before each call.

func NewOracleAdapter added in v0.8.0

func NewOracleAdapter(hc *http.Client) *OracleAdapter

func (*OracleAdapter) CareersURL added in v0.12.0

func (a *OracleAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster company's public job search page.

func (*OracleAdapter) Detail added in v0.8.0

func (a *OracleAdapter) Detail(
	ctx context.Context,
	slug string,
	jobID string,
) (*JobDetail, error)

func (*OracleAdapter) Filters added in v0.8.0

func (a *OracleAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*OracleAdapter) Name added in v0.8.0

func (a *OracleAdapter) Name() string

func (*OracleAdapter) ParseCareersURL added in v0.8.0

func (a *OracleAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes Oracle Candidate Experience job-search and posting URLs. Curated sites fold back to their roster slug; uncurated sites get a canonical careers URL carrying the host, language, and public site.

func (*OracleAdapter) Roster added in v0.8.0

func (a *OracleAdapter) Roster() []CompanyInfo

func (*OracleAdapter) Search added in v0.8.0

func (a *OracleAdapter) Search(
	ctx context.Context,
	slug string,
	p SearchParams,
) (*SearchResult, error)

type RecruiteeAdapter added in v0.6.0

type RecruiteeAdapter struct {
	// contains filtered or unexported fields
}

RecruiteeAdapter serves Recruitee career sites. The public /api/offers endpoint returns the complete board with full descriptions, so all search, filter, and detail behavior is implemented over that dump.

func NewRecruiteeAdapter added in v0.6.0

func NewRecruiteeAdapter(hc *http.Client, dumpCache *DumpCache) *RecruiteeAdapter

func (*RecruiteeAdapter) CareersURL added in v0.12.0

func (a *RecruiteeAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster company's public jobs page.

func (*RecruiteeAdapter) Detail added in v0.6.0

func (a *RecruiteeAdapter) Detail(
	ctx context.Context,
	slug string,
	jobID string,
) (*JobDetail, error)

func (*RecruiteeAdapter) Filters added in v0.6.0

func (a *RecruiteeAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*RecruiteeAdapter) Name added in v0.6.0

func (a *RecruiteeAdapter) Name() string

func (*RecruiteeAdapter) ParseCareersURL added in v0.6.0

func (a *RecruiteeAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes Recruitee subdomain career pages.

func (*RecruiteeAdapter) Roster added in v0.6.0

func (a *RecruiteeAdapter) Roster() []CompanyInfo

func (*RecruiteeAdapter) Search added in v0.6.0

func (a *RecruiteeAdapter) Search(
	ctx context.Context,
	slug string,
	p SearchParams,
) (*SearchResult, error)

type Registry

type Registry struct {
	// contains filtered or unexported fields
}

Registry is the read-only union of every adapter's roster, built once at startup. It owns name resolution; adapters never see unresolved input.

The same normalized slug or display name can belong to entries from different adapters — ATS-local identifiers were never a shared namespace. bySlug and byName therefore map to every entry sharing a key, and Registry.Resolve disambiguates when more than one survives.

func NewRegistry

func NewRegistry(adapters ...Adapter) (*Registry, error)

NewRegistry unions the adapters' rosters. Cross-adapter slug and name collisions are expected — ATS-local identifiers are only unique within their own ATS — so they are indexed for Registry.Resolve to disambiguate at lookup time, not rejected here.

A collision within one adapter's own roster stays a fatal curation bug: two rows sharing a slug would render the same careers URL, so a retry could never disambiguate them, and two rows sharing a name in a roster this repo owns outright has no upstream cause.

An entry whose adapter cannot render a careers URL for it (CareersURL returns ok=false) is required to have a normalized name that collides with no other entry's normalized name or slug: that name is the only retry a caller can offer for it, so the name must resolve uniquely.

func (*Registry) Resolve

func (r *Registry) Resolve(company string) (CompanyResolution, error)

Resolve maps a user-supplied company string to a CompanyResolution. The input can be a roster slug, a display name, or a careers URL. A selected slug is not always a roster key: a careers URL for a company outside the roster resolves to whatever slug the owning adapter minted via Adapter.ParseCareersURL (Workday mints the canonical careers URL).

A careers URL is authoritative and answers immediately. Anything else — including a URL shape no adapter recognizes — falls through to the roster union of slug and name hits. One or more hits are a successful resolution; callers decide how to handle multiple choices. Zero returns a teaching error with either the closest roster slugs or the supported careers-page hosts when the input was URL-shaped.

type RipplingAdapter added in v0.10.0

type RipplingAdapter struct {
	// contains filtered or unexported fields
}

RipplingAdapter serves Rippling-hosted companies. The Job Board API has no server-side search, so Search fetches the whole board and filters it via searchDump. The list carries no JD text and no timestamps, so query matching covers titles and departments only, and ordering falls back to job id. Detail uses the per-job endpoint, one light request.

func NewRipplingAdapter added in v0.10.0

func NewRipplingAdapter(baseURL string, hc *http.Client, dumpCache *DumpCache) (*RipplingAdapter, error)

func (*RipplingAdapter) CareersURL added in v0.12.0

func (a *RipplingAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster company's public job board page.

func (*RipplingAdapter) Detail added in v0.10.0

func (a *RipplingAdapter) Detail(ctx context.Context, slug, jobID string) (*JobDetail, error)

func (*RipplingAdapter) Filters added in v0.10.0

func (a *RipplingAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*RipplingAdapter) Name added in v0.10.0

func (a *RipplingAdapter) Name() string

func (*RipplingAdapter) ParseCareersURL added in v0.10.0

func (a *RipplingAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes Rippling-hosted board URLs; the first path segment is the board slug. Rippling mints lowercase slugs and its API is case-sensitive, so the captured segment is lowercased before use.

func (*RipplingAdapter) Roster added in v0.10.0

func (a *RipplingAdapter) Roster() []CompanyInfo

func (*RipplingAdapter) Search added in v0.10.0

func (a *RipplingAdapter) Search(ctx context.Context, slug string, p SearchParams) (*SearchResult, error)

type SearchParams

type SearchParams struct {
	// Query searches titles, skills, and technology. It excludes locations.
	Query string
	// Location performs a fuzzy text match. Full-dump adapters use their remote
	// fields for "remote".
	Location string
	// Filters uses keys and values returned by [Adapter.Filters]. Values within
	// a key use OR semantics; different keys use AND semantics.
	Filters FilterSet
	// Page is 1-based. Values below 1 request page 1.
	Page int
}

SearchParams are the unified search inputs. Semantics are identical across adapters; how each maps them upstream is the adapter's business.

type SearchResult

type SearchResult struct {
	Jobs       []JobSummary
	TotalCount int
	Page       int
	TotalPages int
}

type SmartRecruitersAdapter added in v0.8.0

type SmartRecruitersAdapter struct {
	// contains filtered or unexported fields
}

SmartRecruitersAdapter serves SmartRecruiters-hosted companies via the public Posting API. Structured filters run server-side. Text search uses q only to select a bounded candidate set, then applies the unified Query and Location semantics locally because SmartRecruiters ORs q terms and searches both titles and locations.

func NewSmartRecruitersAdapter added in v0.8.0

func NewSmartRecruitersAdapter(baseURL string, hc *http.Client) (*SmartRecruitersAdapter, error)

func (*SmartRecruitersAdapter) CareersURL added in v0.12.0

func (a *SmartRecruitersAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster company's public career site.

func (*SmartRecruitersAdapter) Detail added in v0.8.0

func (a *SmartRecruitersAdapter) Detail(ctx context.Context, slug, jobID string) (*JobDetail, error)

func (*SmartRecruitersAdapter) Filters added in v0.8.0

func (a *SmartRecruitersAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*SmartRecruitersAdapter) Name added in v0.8.0

func (a *SmartRecruitersAdapter) Name() string

func (*SmartRecruitersAdapter) ParseCareersURL added in v0.8.0

func (a *SmartRecruitersAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes jobs.smartrecruiters.com career-site URLs; the first path segment is the companyIdentifier, which alone addresses a company (the API accepts it case-insensitively), so non-roster companies need no special slug form. An unknown identifier cannot be validated — the list endpoint answers HTTP 200 with zero results — so a typo'd URL degrades to an empty search, mirroring the raw API.

func (*SmartRecruitersAdapter) Roster added in v0.8.0

func (a *SmartRecruitersAdapter) Roster() []CompanyInfo

func (*SmartRecruitersAdapter) Search added in v0.8.0

type SuccessFactorsAdapter added in v0.7.0

type SuccessFactorsAdapter struct {
	// contains filtered or unexported fields
}

SuccessFactorsAdapter serves SAP SuccessFactors Career Site Builder tenants. Search and detail are server-rendered HTML pages (see internal/provider/successfactors/openapi.yaml); slugs are lowercase career-site hostnames rather than a shared-host tenant token, since every tenant has its own custom domain.

func NewSuccessFactorsAdapter added in v0.7.0

func NewSuccessFactorsAdapter(hc *http.Client) *SuccessFactorsAdapter

func (*SuccessFactorsAdapter) CareersURL added in v0.12.0

func (a *SuccessFactorsAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster company's public job search page.

func (*SuccessFactorsAdapter) Detail added in v0.7.0

func (a *SuccessFactorsAdapter) Detail(ctx context.Context, slug, jobID string) (*JobDetail, error)

func (*SuccessFactorsAdapter) Filters added in v0.7.0

func (a *SuccessFactorsAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*SuccessFactorsAdapter) Name added in v0.7.0

func (a *SuccessFactorsAdapter) Name() string

func (*SuccessFactorsAdapter) ParseCareersURL added in v0.7.0

func (a *SuccessFactorsAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL only recognizes curated hosts: unlike Teamtailor's shared *.teamtailor.com suffix, every SuccessFactors CSB tenant uses its own custom domain with no common pattern to match uncurated tenants against, so recognition is a roster lookup rather than a URL regex.

Examples:

func (*SuccessFactorsAdapter) Roster added in v0.7.0

func (a *SuccessFactorsAdapter) Roster() []CompanyInfo

func (*SuccessFactorsAdapter) Search added in v0.7.0

type TeamtailorAdapter added in v0.6.0

type TeamtailorAdapter struct {
	// contains filtered or unexported fields
}

TeamtailorAdapter serves Teamtailor career sites. The public /jobs.json endpoint returns the complete board with full descriptions, so all search, filter, and detail behavior is implemented over that dump.

func NewTeamtailorAdapter added in v0.6.0

func NewTeamtailorAdapter(hc *http.Client, dumpCache *DumpCache) *TeamtailorAdapter

func (*TeamtailorAdapter) CareersURL added in v0.12.0

func (a *TeamtailorAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster company's public jobs page.

func (*TeamtailorAdapter) Detail added in v0.6.0

func (a *TeamtailorAdapter) Detail(
	ctx context.Context,
	slug string,
	jobID string,
) (*JobDetail, error)

func (*TeamtailorAdapter) Filters added in v0.6.0

func (a *TeamtailorAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*TeamtailorAdapter) Name added in v0.6.0

func (a *TeamtailorAdapter) Name() string

func (*TeamtailorAdapter) ParseCareersURL added in v0.6.0

func (a *TeamtailorAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes every non-reserved Teamtailor-hosted career site and curated custom-domain sites. The lowercase hostname is already the slug form accepted by the other adapter methods.

func (*TeamtailorAdapter) Roster added in v0.6.0

func (a *TeamtailorAdapter) Roster() []CompanyInfo

func (*TeamtailorAdapter) Search added in v0.6.0

func (a *TeamtailorAdapter) Search(
	ctx context.Context,
	slug string,
	p SearchParams,
) (*SearchResult, error)

type UltiProAdapter added in v0.10.0

type UltiProAdapter struct {
	// contains filtered or unexported fields
}

UltiProAdapter serves UltiPro (UKG Pro Recruiting) career boards. Search and its filter catalogs are server-side JSON; job detail is HTML with the posting embedded as a JSON object literal (see internal/provider/ultipro/openapi.yaml). Roster slugs are lowercase company codes (e.g. "tec1006teser"); ParseCareersURL mints a canonical board URL as the slug for non-roster boards, since a company code alone cannot carry the host and board id a non-roster board also needs.

func NewUltiProAdapter added in v0.10.0

func NewUltiProAdapter(hc *http.Client) *UltiProAdapter

func (*UltiProAdapter) CareersURL added in v0.12.0

func (a *UltiProAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster company's public job board page.

func (*UltiProAdapter) Detail added in v0.10.0

func (a *UltiProAdapter) Detail(ctx context.Context, slug, jobID string) (*JobDetail, error)

func (*UltiProAdapter) Filters added in v0.10.0

func (a *UltiProAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*UltiProAdapter) Name added in v0.10.0

func (a *UltiProAdapter) Name() string

func (*UltiProAdapter) ParseCareersURL added in v0.10.0

func (a *UltiProAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes UltiPro career-board URLs. A roster company folds back to its roster slug only when host, company code, AND board id all match the curated entry — UltiPro addresses a board with all three, so a same-coded tenant's *other* board must not collapse onto the curated one. Anything else (including a same-code URL with a different host or board id) gets the canonical URL as a self-describing slug.

func (*UltiProAdapter) Roster added in v0.10.0

func (a *UltiProAdapter) Roster() []CompanyInfo

func (*UltiProAdapter) Search added in v0.10.0

func (a *UltiProAdapter) Search(ctx context.Context, slug string, p SearchParams) (*SearchResult, error)

type WorkableAdapter added in v0.10.0

type WorkableAdapter struct {
	// contains filtered or unexported fields
}

WorkableAdapter serves Workable-hosted companies via the public job board API behind apply.workable.com. Structured filters and locations run server-side. Text search uses query only to select a bounded candidate set, then applies the unified Query semantics locally, because Workable ORs query terms and matches location text too. List rows carry no JD text, so residual filtering fans out to the detail endpoint to keep body-only keyword matches.

func NewWorkableAdapter added in v0.10.0

func NewWorkableAdapter(baseURL string, hc *http.Client) (*WorkableAdapter, error)

func (*WorkableAdapter) CareersURL added in v0.12.0

func (a *WorkableAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster company's public careers page.

func (*WorkableAdapter) Detail added in v0.10.0

func (a *WorkableAdapter) Detail(ctx context.Context, slug, jobID string) (*JobDetail, error)

func (*WorkableAdapter) Filters added in v0.10.0

func (a *WorkableAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*WorkableAdapter) Name added in v0.10.0

func (a *WorkableAdapter) Name() string

func (*WorkableAdapter) ParseCareersURL added in v0.10.0

func (a *WorkableAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes apply.workable.com careers URLs; the first path segment is the account subdomain, which alone addresses a company, so non-roster companies need no special slug form. "api" is the API prefix on the same host, not an account.

func (*WorkableAdapter) Roster added in v0.10.0

func (a *WorkableAdapter) Roster() []CompanyInfo

func (*WorkableAdapter) Search added in v0.10.0

func (a *WorkableAdapter) Search(ctx context.Context, slug string, p SearchParams) (*SearchResult, error)

type WorkdayAdapter

type WorkdayAdapter struct {
	// contains filtered or unexported fields
}

WorkdayAdapter serves Workday CXS tenants. Search runs server-side; location and filters name facet labels, which the adapter resolves to tenant-specific GUIDs via a probe request (appliedFacets wants GUIDs but get_filters reports labels — the stateless price is one extra upstream call whenever location or filters are set).

func NewWorkdayAdapter

func NewWorkdayAdapter(hc *http.Client) *WorkdayAdapter

func (*WorkdayAdapter) CareersURL added in v0.12.0

func (a *WorkdayAdapter) CareersURL(slug string) (string, bool)

CareersURL renders the roster tenant's public career site.

func (*WorkdayAdapter) Detail

func (a *WorkdayAdapter) Detail(ctx context.Context, slug, jobID string) (*JobDetail, error)

func (*WorkdayAdapter) Filters

func (a *WorkdayAdapter) Filters(ctx context.Context, slug string) (FilterSet, error)

func (*WorkdayAdapter) Name

func (a *WorkdayAdapter) Name() string

func (*WorkdayAdapter) ParseCareersURL added in v0.5.0

func (a *WorkdayAdapter) ParseCareersURL(u *url.URL) (string, bool)

ParseCareersURL recognizes myworkdayjobs.com careers URLs. Roster tenants fold back to their roster slug so display names stay identical to name-based resolution; unknown tenants get the canonical URL as a self-describing slug (workday config is three values, which a bare tenant slug can't carry).

func (*WorkdayAdapter) Roster

func (a *WorkdayAdapter) Roster() []CompanyInfo

Roster dedupes by tenant slug: fox and dowjones each hold two share-class rows in companies.yaml sharing one tenant, and the registry treats duplicate slugs as curation bugs.

func (*WorkdayAdapter) Search

func (a *WorkdayAdapter) Search(ctx context.Context, slug string, p SearchParams) (*SearchResult, error)

Search accepts two slug forms: a roster tenant slug, or the canonical careers URL that ParseCareersURL mints for tenants outside the roster. The URL form exists because addressing a Workday site takes three values (tenant, instance host, site name); the roster stores them for curated tenants, and for everyone else the URL is the only slug that can carry all three across stateless calls. Filters and Detail accept the same two forms.

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL
⌂ 换网址