Workplace evidence pipeline

What “practice enrichment queued” actually means.

Every exact registration can receive a durable D1 work item. Queue state, source plans, dated attempts, blockers and review outcomes are reported separately so a stored job is never confused with an active crawler.

Bounded public-web discovery worker status.

CURRENT RUNTIME TRUTH

The recurring service is running; public-web discovery is waiting for its free-search credentials.

Every cycle records a durable run, plans eligible sources and processes at most three exact-registration discovery jobs when the licensed search adapter is configured. Results remain unverified candidates for review. It never automatically creates a facility claim or doctor affiliation.

Runner mode
RECURRING BOUNDED SERVICE
Recurring schedule
CLOUDFLARE CRON EVERY 6 HOURS
Last completed cycle
21 Jul 2026, 1:04 pm UTC
Next planned cycle
21 Jul 2026, 7:04 pm UTC
Eligible external adapters
0
Snapshot generated
21 Jul 2026, 1:41 pm UTC
Low-cost deployment pathThe existing Sites worker and D1 database perform the work, so no always-on server is required. One Google Cloud Scheduler job can call the protected endpoint; Google currently includes three Scheduler jobs per billing account, while the Programmable Search JSON API documents 100 queries per day at no charge. Social and editable-directory pages are retained only as candidate URLs returned by the licensed API and are never scraped automatically.
All D1 work items15,956One durable job per medical system, authority and registration number.
Awaiting planning7,443Oldest created 19 Jul 2026, 12:00 am UTC.
Waiting for source access8,510Oldest blocker recorded 19 Jul 2026, 6:37 pm UTC.
Researching or in review10 actively researching; 1 in review.
Latest planner run21 Jul 2026, 1:04 pm UTCCOMPLETED WITH SOURCE BLOCKER
Jobs classified in latest run100100 jobs considered.
External queries in latest run0Only a real source request can increase this number.
Jobs with any dated attempt20.0% of all D1 work items.

STATE DEFINITIONS

Each label describes a different operational fact.

QUEUEDAwaiting source planning

The exact registration has a D1 work item, but no external query has started.

WAITING_FOR_SOURCE_ACCESSBlocked by source access

Planning found no approved person-keyed adapter. The job stays open and is not a no-match result.

RESEARCHINGExternal check in progress

A dated query using an approved adapter has actually started.

IN_REVIEWEvidence under review

Candidate, facility or affiliation evidence exists but is not yet publishable.

NEEDS_RETRYRetry required

A dated attempt failed or became unavailable and has a future action.

COMPLETEDCycle completed

The reviewed cycle finished. This never means the doctor has no workplace.

Claim AFacility identity

A facility exists at a professional address.

Stored independently even when no doctor relationship has been found.

Identity
Name, facility type and source record ID
Location
Professional address, city, State and PIN
Identifiers
HFR ID, establishment ID, accreditation or other source identifier
Web evidence
Official website, evidence URL and observation date
Claim BDoctor affiliation

A particular doctor is linked to that facility.

Requires an exact registration-scoped identity and its own dated evidence.

Doctor
Canonical registration key and source-reported name
Relationship
Department and role only when explicitly published
Evidence
Affiliation source URL and observation date
Decision
Confidence, identity-match basis and review status
1Queue exact identityCreated during register ingestion or identity-specific search.
2Plan permissible sourcesRecord eligible adapters and explicit access or policy blockers.
3Run a dated source attemptOccurs only when an approved person-keyed adapter is actually available.
4Review separate claimsFacility and doctor-affiliation evidence are resolved independently before display.
6 dated source attempts across 5 source types

A zero-result or unavailable response is retained as an attempt, never converted into “the doctor has no workplace.” Candidate-only observations remain unpublished until corroborated.

Review source rules

SOURCE COVERAGE PLAN

Sources have deliberately different jobs.

Tier 1ENABLED REVIEWED ONLY

ABDM Health Facility Registry public search

An HFR profile can establish a facility name, identifier, professional address, type and website. It does not by itself establish that a named doctor works there.

Facility claim
Supported
Affiliation claim
Never inferred
Collection
Public facility search or approved HFR interface
Open source →
Tier 1ONBOARDING REQUIRED

ABDM HPR public professional directory

Enrollment and public visibility are voluntary. Absence is never an adverse finding, and a facility relationship is used only when the public professional record states it.

Facility claim
Supported
Affiliation claim
Evidence may be retained
Collection
Approved public interface or onboarded API only
Open source →
Tier 1ENABLED REVIEWED ONLY

National Hospital Directory with geocode

The dataset supplies facility identity and location fields. It must not be used to infer a doctor affiliation, and source freshness is retained per snapshot.

Facility claim
Supported
Affiliation claim
Never inferred
Collection
Versioned official CSV snapshot under the Government Open Data License - India
Open source →
Tier 1ENABLED REVIEWED ONLY

Official hospital or clinic clinician directory

The page must identify the doctor and facility unambiguously. A directory entry is time-bounded and does not prove hospital privileges.

Facility claim
Supported
Affiliation claim
Evidence may be retained
Collection
Per-domain policy review, respectful rate limits, and dated observations
Open source →
Tier 1ENABLED REVIEWED ONLY

Direct healthcare-facility confirmation

Confirmation is valid only for the stated facility, relationship, and review period.

Facility claim
Supported
Affiliation claim
Evidence may be retained
Collection
Authenticated, consent-aware submission and review
Open source →
Tier 2ENABLED REVIEWED ONLY

National and State Clinical Establishments registers

Coverage varies by State. Facility registration confirms the establishment, not that a particular doctor currently practises there.

Facility claim
Supported
Affiliation claim
Never inferred
Collection
Published register or export when available
Open source →
Tier 2ENABLED REVIEWED ONLY

CGHS empanelled hospitals and diagnostic centres

Empanelment can corroborate facility identity and address for the publication period. It does not identify every facility or establish an individual doctor relationship.

Facility claim
Supported
Affiliation claim
Never inferred
Collection
Versioned official publication or export
Open source →
Tier 2ENABLED REVIEWED ONLY

NABH accredited healthcare-organisation directory

The directory identifies participating facilities and accreditation context; it does not establish an individual doctor affiliation.

Facility claim
Supported
Affiliation claim
Never inferred
Collection
Public directory within published access rules
Open source →
Tier 2ENABLED REVIEWED ONLY

Doctor-confirmed practice location

A doctor statement is displayed as doctor-confirmed until independently confirmed by the facility.

Facility claim
Supported
Affiliation claim
Evidence may be retained
Collection
Authenticated claim with private identity review
Open source →
Tier 4ONBOARDING REQUIRED

Google Programmable Search professional-web discovery

Search results may identify possible public professional, facility, directory or social pages. They create discovery candidates only; they do not create facility or doctor-affiliation claims automatically. Private profiles, personal contact data, residential addresses and inferred social graphs are excluded.

Facility claim
Supported
Affiliation claim
Evidence may be retained
Collection
Official JSON API, bounded batches and review-only candidate retention
Open source →
Tier 4PERMISSION REQUIRED

Justdial public professional and facility directory

Editable directory information may be stale or user-supplied. It can generate a candidate but cannot by itself produce a verified affiliation.

Facility claim
Supported
Affiliation claim
Evidence may be retained
Collection
Manual observation or licensed/permissioned feed only
Open source →
Tier 4PERMISSION REQUIRED

Public professional social profile

Only deliberately public professional claims may be considered. Personal posts, private profiles, contact details, inferred locations, and scraped social graphs are excluded.

Facility claim
Supported
Affiliation claim
Evidence may be retained
Collection
Official platform API, permissioned feed, or manual observation only
Open source →