linuteresa/hackerrank-orchestrate-october26

How linuteresa/hackerrank-orchestrate-october26 works

This is a deterministic personal-data deletion auditor. A request is classified for deletion intent, then the audit resolves the subject and linked resources, ingests evidence, evaluates per-copy state and policy, and produces ordered findings, remediation, and verification steps. The main audit owns these stages; Claude is limited to intent classification and screenshot transcription, while code determines statuses. The graph follows README and source-index relationships; several parser and state internals were not excerpted, so their detailed runtime links are omitted.

Main entry points: audit.py Audit orchestrator · loader.py Evidence loader · model.py Evidence records · main.py Audit entry

Components 21

Audit workflow

Audit entry code/main.py
Audit orchestrator code/audit.py

Evidence ingestion

Requests and context code/ingest/csv_inputs.py
Evidence loader code/ingest/loader.py
Structured parsers code/ingest/json_records.py
Log evidence code/ingest/logs.py
SQL snapshots code/ingest/sql_snapshots.py
Screenshot facts code/ingest/png_facts.py

Identity and state

Evidence store code/core/store.py
Evidence records code/core/model.py
Identity resolution code/core/identity.py
Domain vocabulary code/core/vocab.py
Category mapping code/core/categories.py
Per-copy state code/core/resource_state.py

Decision and output

Deletion policy code/policy.py
Step ordering code/planner.py
Validated output code/main.py

Model assistance

Intent classifier code/agent/intent.py
Screenshot transcription code/agent/vision.py
Claude API and cache code/agent/llm.py

Other components

Requester

Connections 30

Requester → Audit entry: submits request README.md
Audit entry → Audit orchestrator: runs audit code/main.py
Audit entry → Requests and context: loads requests code/main.py
Audit orchestrator → Intent classifier: classifies intent code/audit.py
Audit orchestrator → Evidence loader: loads evidence code/audit.py
Audit orchestrator → Requests and context: loads context code/audit.py
Audit orchestrator → Identity resolution: resolves subject code/audit.py
Audit orchestrator → Category mapping: maps category code/audit.py
Audit orchestrator → Per-copy state: evaluates copies code/audit.py
Audit orchestrator → Deletion policy: applies rules code/audit.py
Audit orchestrator → Step ordering: orders steps README.md
Audit orchestrator → Validated output: returns audit result code/main.py
Evidence loader → Structured parsers: dispatches parsing code/ingest/loader.py
Evidence loader → Log evidence: dispatches parsing code/ingest/loader.py
Evidence loader → SQL snapshots: dispatches parsing code/ingest/loader.py
Evidence loader → Screenshot facts: dispatches parsing code/ingest/loader.py
Evidence loader → Evidence store: populates store code/ingest/loader.py
Structured parsers → Evidence records: creates facts code/ingest/json_records.py
Log evidence → Evidence records: creates facts code/ingest/logs.py
SQL snapshots → Evidence records: creates facts code/ingest/sql_snapshots.py
Screenshot facts → Evidence records: creates facts code/ingest/png_facts.py
Identity resolution → Domain vocabulary: uses identifiers code/core/identity.py
Category mapping → Domain vocabulary: uses mappings code/core/categories.py
Evidence store → Evidence records: stores records code/core/store.py
Per-copy state → Evidence records: reads observations code/core/resource_state.py
Deletion policy → Per-copy state: uses copy state code/policy.py
Intent classifier → Claude API and cache: requests classification code/agent/intent.py
Screenshot facts → Screenshot transcription: requests transcription code/ingest/loader.py
Screenshot transcription → Claude API and cache: uses model code/agent/vision.py
Audit entry → Validated output: writes output README.md

Drawn by GitDiagram on November 11, 2026 from the repository's file tree, README and source files. Read as Markdown ·