How scpz24/requirementseeker works
RequirementSeeker is a set of Python packages for collecting and preparing comment datasets and for developing a requirements-analysis agent. The README explicitly says there is not yet a complete application or integrated analysis pipeline, so this map shows package capabilities without implying a connected end-to-end runtime. The sampled agent orchestrator supports its internal analysis-stage links; collector and dataset-tool wiring is less fully evidenced, so uncertain connections are omitted.
Main entry points: m2.py M2 orchestrator · runner.py Collection runner · sanitize.py Sanitization · policy.py Quality and sampling
Components 21
Collection
- Collector CLI
packages/collector/src/requirementseeker_collector/cli.py - Collection planning
packages/collector/src/requirementseeker_collector/planning.py - Candidate discovery
packages/collector/src/requirementseeker_collector/candidates/discovery.py - Coverage validation
packages/collector/src/requirementseeker_collector/candidates/coverage.py - Collection runner
packages/collector/src/requirementseeker_collector/runner.py - Platform adapters
packages/collector/src/requirementseeker_collector/adapters - Raw data contracts
packages/collector/src/requirementseeker_collector/contracts.py
Dataset preparation
- Sanitization
packages/dataset-tools/src/requirementseeker_dataset/sanitize.py
Analysis agent
- Quality and sampling
packages/agent/src/requirementseeker_agent/sampling/policy.py - Analysis contracts
packages/agent/src/requirementseeker_agent/contracts/requests.py - M2 orchestrator
packages/agent/src/requirementseeker_agent/pipeline/m2.py - Signal extraction
packages/agent/src/requirementseeker_agent/pipeline/signals.py - Signal clustering
packages/agent/src/requirementseeker_agent/pipeline/clusters.py - Consensus rules
packages/agent/src/requirementseeker_agent/rules/consensus.py - Budget and cache
packages/agent/src/requirementseeker_agent/runtime - Analysis results
packages/agent/src/requirementseeker_agent/contracts/results.py
Other components
- Operator
- Video platforms
- Raw collection data
- Host application
- Model gateway
packages/agent/src/requirementseeker_agent/model/types.py
Connections 19
- Operator → Collector CLI: invokes
- Video platforms → Collection runner: serves pages
- Collection runner → Platform adapters: parses responses
- Collection runner → Raw data contracts: builds records
- Collection runner → Raw collection data: writes collection
- Raw collection data → Sanitization: provides raw input
- Operator → Sanitization: approves plan
- Sanitization → Quality and sampling: provides sanitized manifest
- Host application → Quality and sampling: requests sampling
- Quality and sampling → Host application: returns plan
- Host application → M2 orchestrator: submits request
- M2 orchestrator → Analysis contracts: validates input
- M2 orchestrator → Signal extraction: extracts signals
- M2 orchestrator → Signal clustering: clusters signals
- M2 orchestrator → Consensus rules: checks clusters
- Signal extraction → Model gateway: invokes model
- Signal clustering → Model gateway: invokes model
- M2 orchestrator → Budget and cache: uses budget and cache
- M2 orchestrator → Analysis results: produces result