5G Vector

Ingest Carrier Statement

Full pipeline: ingest a carrier PDF or Excel statement, normalize, validate against BigQuery, and stage for review or posting. Args: file_path: Absolute path to the PDF or Excel file carrier: Carrier slug (e.g. 'nationwide', 'travelers') mode: 'trial' (default, safe) or 'live' (writes to Epic) Returns: run_summary with counts by status and run_id for follow-up

Ingest PDF Statement

Ingest a single carrier PDF statement. Runs the full pipeline. Alias for ingest_carrier_statement for explicit PDF files.

Ingest Excel Bordereaux

Ingest a single carrier Excel bordereaux. Runs the full pipeline.

Normalize Transactions

Parse and normalize a carrier document WITHOUT validation or staging. Returns the raw normalized transactions for inspection. Useful for debugging normalization quality before running full pipeline. Args: file_path: Path to PDF or Excel file carrier: Carrier slug mode: 'trial' or 'live'

Validate Against Data Lake

Parse, normalize, AND validate a carrier document against BigQuery — but do NOT stage or post. Returns the full validation report including confidence scores, policy matches, and queue classifications. Args: file_path: Path to PDF or Excel file carrier: Carrier slug mode: 'trial' or 'live'

Score Transaction Confidence

Score a single transaction's confidence without running the full pipeline. Looks up the policy in BigQuery, checks for duplicates, and returns the confidence score breakdown. Useful for testing and debugging. Args: carrier: Carrier slug policy_number: Carrier policy number client_name: Insured/client name amount: Transaction amount as string (e.g. '1500.00') effective_date: Effective date YYYY-MM-DD (optional)

Browse Carrier Portal

Log into a carrier portal via Playwright and download latest statements. Returns paths to downloaded files which can then be passed to ingest_carrier_statement. Args: carrier: Carrier slug (must have portal_url in its config) username: Portal login username password: Portal login password mode: 'trial' or 'live' download_dir: Optional override for download directory

Get Exception Queue

Return all transactions in the human review queue for today. These are transactions with confidence score 0.80-0.94 that need accounting team review before posting to Epic.

Approve Transaction

Approve a single transaction from the review queue. In live mode, this will trigger an Epic posting. Args: transaction_id: UUID of the transaction reviewer: Name/email of the reviewer notes: Optional review notes

Reject Transaction

Reject a transaction from the review queue. Args: transaction_id: UUID of the transaction reviewer: Name/email of the reviewer reason: Reason for rejection (required)

Bulk Approve Run

Bulk approve all review-queue transactions from a specific run. Use after verifying the trial run report is accurate. Args: run_id: The run ID from ingest_carrier_statement output reviewer: Name/email of the approving accountant

Post to Applied Epic

Post all approved transactions from a run to Applied Epic via REST API. Only works in live mode. In trial mode, returns a simulation report. Args: run_id: The run ID to post

Post to Epic via Browser

Fallback: post approved transactions to Applied Epic by automating the Epic web UI via Playwright. Use when the REST API / SDK is unavailable. Args: run_id: The run ID to post epic_username: Epic web UI login username epic_password: Epic web UI login password

Generate Epic CSV Import

Generate an Applied Epic-compatible CSV import file for a run's approved transactions. Use for batch imports when neither the SDK nor UI automation is available. Args: run_id: The run ID to export output_path: Optional output file path (auto-generated if omitted)

Rollback Epic Entries

Rollback all Epic entries from a run. Use if a bad batch was posted. Calls Epic void for each entry. Args: run_id: The run ID to roll back reason: Reason for rollback (required, logged to audit trail)

Daily Accuracy Metrics

Get the daily accuracy scorecard for the accounting team. Shows total processed, auto-approved, review queue, exceptions, accuracy %. Args: target_date: Date string YYYY-MM-DD (defaults to today)

Carrier Accuracy Report

Get accuracy metrics for a specific carrier over the past N days. Useful for deciding when to promote a carrier from trial to live mode. Args: carrier: Carrier slug (e.g. 'nationwide') days: Number of days to look back (default 30)

Run History

Get recent ingestion run history. Shows run_id, carrier, mode, counts, and status for each run over the past N days. Args: days: Number of days to look back (default 7)

List Supported Carriers

List all carriers currently configured in the system.

Reconciliation Report

Compare posted transactions in our data lake against Applied Epic entries. For each posted transaction, reads the Epic entry and checks for discrepancies in amount, policy ID, and status. Returns match rate, mismatched records, and entries missing from Epic. Args: run_id: Specific run to reconcile (optional) carrier: Filter by carrier slug (optional) target_date: Date string YYYY-MM-DD (optional, defaults to today)

Trial Diff Report

Side-by-side comparison: what the system WOULD have posted to Epic (from the shadow table) vs what currently exists in Epic for those policies. This is the key report for the accounting team during trial/shadow mode. Shows every parsed transaction alongside existing Epic data, highlighting amount deltas, name mismatches, and an overall accuracy recommendation. Args: run_id: Specific trial run to diff (optional) carrier: Filter by carrier slug (optional) target_date: Date string YYY