bbchops/transcription
1
0
Fork 0
generated from john/python-template
Code Issues 4 Pull Requests Actions Packages Projects Releases Wiki Activity

Delete V3 duplicates

Browse Source
This commit is contained in:
Jim Lancaster
2026-08-09 13:34:36 -05:00
parent 5753eb0135
commit 9b4d6f0340
6 changed files with 0 additions and 541 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/architecture_v3.md
-142
View File
@@ -1,142 +0,0 @@
# System Architecture (Version 3)
This document describes the V3 production architecture of the personal historical-document transcription system.
## Architecture Objectives
* Preserve source material as immutable transcribed text alongside page-level spatial AI metadata and complete provider API envelopes.
* Support batching multi-image and folder uploads cleanly into sequential pages (`page_number`).
* Capture complete input prompt provenance (`system_prompt`, `user_prompt`, `prompt_hash`) and execution parameters (`temperature`, `top_p`) at submission time on `Job`.
* Leverage asynchronous worker pools (`asyncio`) for parallel single-image API execution bounded by rate limiters (`asyncio.Semaphore`).
* Maintain relational data-model portability across the supported backends by using SQLModel/SQLAlchemy and compatibility types so the same domain schema works in SQLite for local development/testing and PostgreSQL in production.
* Keep operator tooling and local maintenance workflows OS-independent by using Python or other cross-platform interfaces for canonical project automation.
* Verify image asset integrity via SHA-256 file hashing (`file_hash`) while storing binary assets on the local filesystem.
* Standardize all data validation, API parsing, and database models on **Pydantic V2** and **SQLModel**.
* Support rich historical attribution (multi-author and multi-recipient relationships via `DocumentPerson`).
## Runtime Topology
The V3 runtime operates as an asynchronous Python application:
* FastAPI + NiceGUI web application process.
* In-process `asyncio` background task orchestrator for parallel API execution.
* Relational persistence via SQLModel / SQLAlchemy, using SQLite for local development/testing and PostgreSQL as the production persistence target.
* Pydantic V2 validation layer wrapping API payloads, prompt configurations, and JSON metadata schemas.
* Cross-platform operator workflows implemented in Python so core local operations run consistently on Windows, Linux, and macOS.
^^^mermaid
flowchart LR
U[Browser User] --> A[FastAPI + NiceGUI App]
A --> W[Asyncio Worker Engine]
A --> DB[(Relational DB\nSQLite / PostgreSQL)]
W --> P[Vision Provider APIs\nOpenAI / Claude / OpenRouter]
W --> DB
^^^
## Lifecycle Ownership
Application lifespan owns runtime setup/teardown:
* Initialize environment logging, directory paths, and Pydantic configuration.
* Manage asynchronous database engine connection pools (`aiosqlite` or `asyncpg`).
* Execute database bootstrap (`SQLModel.metadata.create_all()`) or migrations.
* Recover stale or interrupted processing jobs on startup.
* Manage graceful shutdown of active `asyncio` worker pools.
## Layered Module Structure
### Interface Layer
* `src/transcription/ui/**` (NiceGUI pages, multi-page renderers, person cards)
* `src/transcription/api/**` (FastAPI routes and JSON error handlers)
### Application & Async Worker Layer
* `src/transcription/services/workflows.py`
* `src/transcription/worker.py`
Responsibilities:
* Batch orchestration and status transitions (`queued` -> `processing` -> `transcribed` | `partial_success` | `failed`).
* Parallel single-image API execution using `asyncio.gather` bounded by `asyncio.Semaphore`.
* Resolve prompt configuration at submission time and persist frozen snapshot fields on `Job`.
* Pydantic schema parsing and validation prior to database storage.
### Domain & Service Layer
* `src/transcription/db/models.py` (SQLModel schema definitions for Document, Source, Job, JobSource, Person, DocumentPerson)
* `src/transcription/services/*.py` (Transactional operations for `Document`, `Person`, `Source`, `Job`, and `JobSource`)
### Infrastructure Layer
* `src/transcription/db/**` (Async database session factory, engine creation, and JSON dialect abstractions)
* `src/transcription/providers/**` (OpenAI, Anthropic, and OpenRouter Vision SDK adapters)
## Processing Workflow
1. User uploads a folder or batch of images for a `Document`.
2. System hashes each image file (SHA-256), writes image files to filesystem storage, and creates `Document`, `Job(status='queued')`, and ordered `Source` pages (`page_number = 1..N`).
3. Worker claims job, sets `Job.status = 'processing'`, and spawns parallel `asyncio` tasks bounded by semaphore.
4. Each task reads the frozen prompt snapshot from `Job` and calls Vision API for a **single** `Source` image.
5. On task completion:
* Writes a `JobSource` record containing `status='transcribed'`, `raw_transcription`, operational `ai_metadata`, and complete unedited `raw_api_response`.
* Caches active output text to `Source.raw_transcription`.
6. On page failure:
* Writes `JobSource` record with `status='failed'` and `error_detail`.
7. Once all page tasks resolve:
* Marks `Job.status` as `transcribed` (100% success), `partial_success` (at least 1 success, 1 failure), or `failed` (all failed).
## Domain Ownership & Invariants
* **Immutable AI Outputs:** `source.raw_transcription` and `job_source.raw_transcription` store original, point-in-time machine output and are immutable.
* **Complete Input & Output Provenance:** Every `job` stores the exact frozen input configuration sent to the model, and every `job_source` stores per-page output evidence including the complete REST response envelope returned.
* **Inlined Revisions:** Human corrections occur on `source.revised_text`. UI renders `COALESCE(revised_text, raw_transcription)`.
* **Sequential Integrity:** Multi-page documents are strictly ordered by `source.page_number ASC`.
* **Page Execution Isolation:** A failure on one page image does not invalidate successful transcriptions on sister pages in the same batch job.
## Data Model Summary
* `Document` has many `Source` pages, many `Job` runs, and many `Person` records via `DocumentPerson` junction (`author` or `recipient`).
* `Source` belongs to one `Document` and can be processed across many `JobSource` executions.
* `Job` has many `JobSource` execution records.
* `JobSource` holds page-level execution status, output text, and raw response JSON.
## Test Strategy
* Unit tests for SQLModel/Pydantic V2 models, JSON cross-dialect serialization, and file hashing functions.
* Integration tests for async database connection handling, session management, and queries.
* Async workflow tests using mock AI providers to verify `partial_success`, page-level failure isolation, and retry logic.
* UI integration tests for multi-page rendering and person attribution management.
---
## Technology References
* [FastAPI documentation](https://fastapi.tiangolo.com/)
* [NiceGUI documentation](https://nicegui.io/documentation)
* [SQLModel documentation](https://sqlmodel.tiangolo.com/)
* [SQLAlchemy Async I/O documentation](https://docs.sqlalchemy.org/en/20/orm/extensions/asyncio.html)
* [Python asyncio](https://www.google.com/search?q=https://docs.python.org/3/library/asyncio.html%23module-asyncio)
* [Pydantic Validation](https://pydantic.dev/docs/validation/latest/get-started/)
## Related Local References
- [System Overview](index_v3.md)
- [System Design Intent](invariant/intent.md)
- [Transcription Methodology](invariant/transcription_methodology.md)
- System Architecture (this document)
- [System Requirements](requirements_v3.md)
- [Data model](schema_v3.md)
- [Error Handling Policy](error_handling_v3.md)
- [Implementation Plan](implementation_plan_v3.md)
docs/error_handling_v3.md
-87
View File
@@ -1,87 +0,0 @@
# Error Handling Policy (Version 3)
This document defines the canonical error-handling policy for the v3 document transcription system.
## Error Handling Objectives
* Make failures visible in clear, actionable language at both the document and individual page levels.
* Support **isolated failure handling** in multi-image batches so single page errors do not crash an entire batch job.
* Preserve diagnostic detail (Pydantic validation errors, raw provider REST envelopes, exact input prompts) in generic database JSON structures for fast troubleshooting.
* Ensure consistent error envelope structure across API, UI, and async worker boundaries.
## Scope And Authority
Governs error behavior across NiceGUI pages, FastAPI routes, service orchestration, `asyncio` background tasks, database interactions, and AI provider adapters.
## Error Taxonomy
| Category | Definition | Retriable |
| --- | --- | --- |
| `validation_error` | Pydantic payload or parameter schema validation failure | no |
| `user_input_error` | Unacceptable user file (unsupported image type, corrupt file) | no |
| `not_found_error` | Requested resource (`Document`, `Source`, `Person`, `Job`) missing | no |
| `conflict_error` | Operation violates state constraints (e.g., duplicate `document_person` role) | no |
| `external_provider_error` | AI Provider API failure (rate limit, vision execution error) | yes |
| `infrastructure_transient_error` | Temporary DB connection reset or HTTP timeout | yes |
| `infrastructure_persistent_error` | Database down, missing API credentials, misconfiguration | no |
| `internal_unexpected_error` | Uncaught Python exception or logic defect | no |
## Async Batch & Page-Level Error Behavior
In multi-image `asyncio` batch processing:
1. **Page Isolation:** Exceptions caught during individual page calls are trapped within the `asyncio` task wrapper.
2. **Page Record Logging:** Page failure details, along with the prompt inputs and hyperparameters attempted, are written directly to `job_source.error_detail` and `job_source.status = 'failed'`.
3. **Batch Aggregate State:**
* If **all** page tasks succeed -> `job.status = 'completed'`.
* If **some** page tasks fail -> `job.status = 'partial_success'`.
* If **all** page tasks fail -> `job.status = 'failed'`.
4. **Retry Strategy:** The UI exposes a "Retry Failed Pages" option for `partial_success` jobs, which spawns a new targeted `Job` containing *only* the `Source` IDs marked as `failed`.
## API Error Response Contract
API error responses return a structured JSON envelope:
^^^json
{
"error_id": "err_uuid_12345",
"category": "validation_error",
"message": "The uploaded payload failed schema validation.",
"suggestion": "Check file format and metadata fields, then try again.",
"details": {
"pydantic_errors": [...]
},
"timestamp": "2026-08-08T15:00:00Z"
}
^^^
HTTP Status Mappings:
* `validation_error`, `user_input_error` -> `400`
* `not_found_error` -> `404`
* `conflict_error` -> `409`
* `external_provider_error` -> `502` / `503`
* `infrastructure_transient_error` -> `503`
* `infrastructure_persistent_error`, `internal_unexpected_error` -> `500`
---
## Technology References
* [FastAPI documentation](https://fastapi.tiangolo.com/)
* [NiceGUI documentation](https://nicegui.io/documentation)
* [SQLModel documentation](https://sqlmodel.tiangolo.com/)
* [Python asyncio](https://www.google.com/search?q=https://docs.python.org/3/library/asyncio.html%23module-asyncio)
* [Pydantic Validation](https://pydantic.dev/docs/validation/latest/get-started/)
## Related Local References
- [System Overview](index_v3.md)
- [System Design Intent](invariant/intent.md)
- [Transcription Methodology](invariant/transcription_methodology.md)
- [System Architecture](architecture_v3.md)
- [System Requirements](requirements_v3.md)
- [Data model](schema_v3.md)
- Error Handling Policy (this document)
- [Implementation Plan](implementation_plan_v3.md)
docs/implementation_plan_v3.md
-81
View File
@@ -1,81 +0,0 @@
# Implementation Plan (Version 3)
## Goal
Replace the current v2 SQLModel schema with the approved v3 schema and make sure every database operation works through the existing async SQLAlchemy/SQLModel session layer.
Use a fresh database. There will be no migrations, data conversion, legacy compatibility shims, or parallel v2/v3 code paths.
## Current Project Impact
* `src/transcription/db/models.py` defines the SQLModel tables. It must be updated to match the approved v3 schema (`Document`, `Person`, `DocumentPerson`, `Source`, `Job`, `JobSource`).
* The v3 target adds frozen submission-time prompt snapshot fields (`prompt_name`, `prompt_hash`, `system_prompt`, `user_prompt`, `temperature`, `top_p`) to `Job` and full output payloads (`raw_api_response`, `ai_metadata`) to `JobSource`.
* The v3 target adds image asset verification fields (`file_hash`, `file_size_bytes`) to `Source`.
* Database operations must utilize `JSONBCompat` and the existing SQLModel/SQLAlchemy abstractions to preserve the same logical schema and JSON behavior across the supported backends, while keeping PostgreSQL as the intended production database.
* Async CRUD lives in `DocumentService`, `JobService`, `TranscriptionService`, and upload helpers. Their queries and relationship loading must be updated for v3 fields.
* Canonical operator tooling must remain OS-independent; safety workflows such as destructive-test backup and restore should run through Python or other cross-platform entry points rather than platform-specific shells.
## Implementation
### 1. Update the Schema and Domain Models
* Replace the models in `src/transcription/db/models.py` with the approved v3 tables, enums, relationships, foreign keys, constraints, and indexes.
* Ensure all JSON fields use `JSONBCompat` for dialect portability across SQLite and PostgreSQL.
* Keep `SQLModel.metadata.create_all()` as the schema bootstrap for fresh databases.
* Delete `_ensure_sqlite_compat_columns()` and all legacy schema patching from `src/transcription/db/operations.py`.
* Keep the Python models and `docs/schema_v3.md` perfectly synchronized.
### 2. Update Data Services and Async Worker Layer
* Update job creation and worker orchestration so prompt configuration is resolved at submission and frozen onto `Job` (`prompt_name`, `prompt_hash`, `system_prompt`, `user_prompt`, `temperature`, `top_p`) before execution starts.
* Update `TranscriptionService` and provider adapters to store the complete unedited API REST response dictionary into `job_source.raw_api_response` alongside operational metrics in `job_source.ai_metadata`.
* Update upload handlers to calculate and store file metadata (`file_hash` via SHA-256, `file_size_bytes`) on `Source` records during file ingestion.
* Remove legacy single-source compatibility flows so worker paths persist per-page outcomes only through `JobSource` updates.
### 3. Update Integration Tests and Mock AI Providers
* Update mock provider fixtures in test suites to return realistic complete API response envelopes.
* Verify test coverage for `JSONBCompat` field writes and reads under SQLite in-memory test databases.
* Add assertions in async workflow tests to verify frozen prompt snapshot fields on `Job`, plus per-page failure isolation and output evidence on `JobSource`.
### 4. Update the UI for the v3 Schema
* Review the UI components and views displaying document, job, person, and source data so they reference v3 schema properties instead of v2 relationships.
* Ensure the UI correctly renders `COALESCE(revised_text, raw_transcription)` for page viewing and inline editing.
* Ensure resubmit actions only queue failed pages and preserve frozen prompt snapshot behavior on the existing `Job`.
* Consider the guidance in `docs/ui_style_guide.md` when making UI changes so updated views remain consistent with the project’s visual conventions.
### 5. Keep Operational Tooling Portable
* Implement destructive-test backup and restore workflows in Python so the canonical path runs on Windows, Linux, and macOS.
* Avoid making core developer or recovery procedures depend on PowerShell-only or shell-specific semantics.
* Keep operational documentation aligned with the cross-platform command path used by the repository.
## Done When
* A fresh database is created directly from the v3 SQLModel metadata.
* Frozen prompt input provenance is captured on `Job` for each submission, and full per-page output evidence is captured on `JobSource` for every AI execution task.
* The focused tests and full test suite pass on both SQLite and PostgreSQL backends.
* Canonical operator workflows required for development and destructive-test recovery run without a Windows-only shell dependency.
## Out of Scope
* Database migrations or preservation of v2 data
* Legacy compatibility code
* UI redesign, batch orchestration, worker concurrency, deployment, and operational runbooks
---
## Related Local References
- [System Overview](index_v3.md)
- [System Design Intent](invariant/intent.md)
- [Transcription Methodology](invariant/transcription_methodology.md)
- [System Architecture](architecture_v3.md)
- [System Requirements](requirements_v3.md)
- [Data model](schema_v3.md)
- [Error Handling Policy](error_handling_v3.md)
- Implementation Plan (this document)
docs/index_v3.md
-49
View File
@@ -1,49 +0,0 @@
# Document Transcription System Overview (Version 3)
This project is a personal-scale application for transcribing, indexing, and preserving historical family documents, letters, postcards, and journals.
## Start Here
Read [architecture_v3.md](https://www.google.com/search?q=architecture_v3.md) first for technical overview and system design.
## Core V3 Capabilities
* **Folder & Multi-Image Ingestion:** Upload one or more images that map sequentially (`page_number`) under a single `Document`.
* **Parallel Async AI Vision Engine:** Concurrently process single-page image transcriptions using Python `asyncio` bounded by rate limiters.
* **Portable Relational Storage:** SQLModel and SQLAlchemy preserve a portable relational model across the supported backends, with SQLite for local development/testing and PostgreSQL as the production database target.
* **Cross-Platform Operations:** Canonical developer and recovery workflows run through Python-based, OS-independent tooling rather than platform-specific shell scripts.
* **Complete Auditability & Provenance:** Capture frozen submission-time input prompts (`system_prompt`, `user_prompt`) and hyperparameters (`temperature`, `top_p`) on `Job`, plus per-page operational metrics (`ai_metadata`) and full provider response envelopes (`raw_api_response`) on `JobSource`.
* **Asset Integrity Tracking:** Calculate and store cryptographic hashes (SHA-256) and file sizes on `Source` image records while preserving clean filesystem storage.
* **Pydantic V2 Validation:** End-to-end type safety, DB row mapping, and JSON payload validation.
* **Historical Person Management:** Track authors and recipients across documents with rich biographical entities (`Person`).
* **Page-Level Execution Auditing & Revisions:** Store immutable point-in-time machine output per run while enabling inline human corrections (`revised_text`).
* **Partial Failure Recovery:** Bounded batch execution that isolates single-page API errors (`partial_success`) for simple retries.
## Technical Stack
* **Application Web Framework:** FastAPI + NiceGUI
* **Persistence Engine:** SQLModel / SQLAlchemy (SQLite for development/testing, PostgreSQL for production)
* **Data Validation & Schemas:** Pydantic V2
* **Concurrency & Workers:** Python `asyncio` worker pool with `asyncio.Semaphore`
* **Vision Providers:** OpenAI, Anthropic, and OpenRouter Vision models via native SDK adapters
---
## Technology References
* [FastAPI documentation](https://fastapi.tiangolo.com/)
* [NiceGUI documentation](https://nicegui.io/documentation)
* [SQLModel documentation](https://sqlmodel.tiangolo.com/)
* [Python asyncio](https://www.google.com/search?q=https://docs.python.org/3/library/asyncio.html%23module-asyncio)
* [Pydantic Validation](https://pydantic.dev/docs/validation/latest/get-started/)
## Documentation Index
- System Overview (this document)
- [System Design Intent](invariant/intent.md)
- [Transcription Methodology](invariant/transcription_methodology.md)
- [System Architecture](architecture_v3.md)
- [System Requirements](requirements_v3.md)
- [Data model](schema_v3.md)
- [Error Handling Policy](error_handling_v3.md)
- [Implementation Plan](implementation_plan_v3.md)
docs/requirements_v3.md
-45
View File
@@ -1,45 +0,0 @@
# Document Transcription System Requirements (Version 3)
This document captures the **Version 3 baseline requirements** for the production implementation.
## Requirements Model
| ID | Category | Requirement | Verify Method |
| --- | --- | --- | --- |
| REQ-0 | System | Provide end-to-end multi-page document transcription with persistent, inspectable async job states. | demonstration |
| REQ-1 | Functional | Allow users to upload multi-image batches as sequential `Source` pages under a `Document`. | test |
| REQ-2 | Functional | Process multi-page jobs asynchronously using an `asyncio` worker pool bounded by rate limits. | test |
| REQ-3 | Functional | Persist frozen submission-time execution parameters and full input prompts (`system_prompt`, `user_prompt`, `prompt_name`, `prompt_hash`, `temperature`, `top_p`) on `Job`, and persist page-level output responses (`raw_transcription`, `ai_metadata`, `raw_api_response`) on `JobSource`. | test |
| REQ-4 | Functional | Support job states (`queued`, `processing`, `completed`, `partial_success`, `failed`) and page states (`pending`, `transcribed`, `failed`). | inspection |
| REQ-5 | Functional | Allow users to manage historical `Person` records and link multiple authors/recipients to a `Document` via `DocumentPerson`. | test |
| REQ-6 | Functional | Maintain immutable original machine output on `Source.raw_transcription` while permitting inline human edits on `Source.revised_text`. | test |
| REQ-7 | Data Constraint | Use SQLModel/SQLAlchemy to preserve a portable relational domain model and compatible data shape across the supported backends, with SQLite for local development/testing and PostgreSQL as the production system of record. | inspection |
| REQ-8 | Data Constraint | Validate all API requests, database rows, and JSON structures using Pydantic V2 schemas and SQLModel. | test |
| REQ-9 | Interface | Render multi-page transcriptions sequentially by `page_number` in the web UI with author/recipient metadata. | demonstration |
| REQ-10 | Operations | Allow operators to resubmit only failed pages for queued reprocessing while preserving the frozen prompt snapshot on the existing `Job`. | test |
| REQ-11 | Data Constraint | Calculate and store cryptographic file hashes (SHA-256) and file sizes for uploaded source images to track asset integrity. | test |
| REQ-12 | Operations Constraint | Keep core development, testing, restore, and recovery workflows OS-independent across Windows, Linux, and macOS; do not require a platform-specific shell for canonical project processes. | inspection |
## Element Satisfaction Mapping
* **UI (NiceGUI):** Satisfies REQ-1, REQ-5, REQ-6, REQ-9, REQ-10.
* **API (FastAPI):** Satisfies REQ-1, REQ-4, REQ-5, REQ-8.
* **WORKER (asyncio):** Satisfies REQ-2, REQ-3, REQ-4, REQ-10.
* **PERSISTENCE (SQLModel/SQLAlchemy):** Satisfies REQ-3, REQ-6, REQ-7, REQ-11.
* **MODELS (Pydantic V2 / SQLModel):** Satisfies REQ-8.
* **OPERATIONS TOOLING (Python / OS-neutral automation):** Satisfies REQ-12.
---
## Related Local References
- [System Overview](index_v3.md)
- [System Design Intent](invariant/intent.md)
- [Transcription Methodology](invariant/transcription_methodology.md)
- [System Architecture](architecture_v3.md)
- System Requirements (this document)
- [Data model](schema_v3.md)
- [Error Handling Policy](error_handling_v3.md)
- [Implementation Plan](implementation_plan_v3.md)
docs/schema_v3.md
-137
View File
@@ -1,137 +0,0 @@
# Database Schema (Version 3)
This document describes the relational schema for the transcription platform. It incorporates multi-image batch orchestration, page-level execution tracking, many-to-many author/recipient attribution, submission-time prompt snapshot capture, and raw API payload evidence for archival auditing.
The schema uses generic JSON columns compatible with SQLite in local development and PostgreSQL native JSONB/UUID types in production.
## Entity Relationship Diagram
```mermaid
erDiagram
PERSON {
UUID id PK
TEXT full_name
TEXT display_name
TEXT maiden_name
DATE birth_date
TEXT birth_date_raw
TEXT birth_place
DATE death_date
TEXT death_date_raw
TEXT death_place
TEXT biography
TEXT portrait_path
JSONB metadata
TIMESTAMPTZ created_at
TIMESTAMPTZ updated_at
}
DOCUMENT {
UUID id PK
TEXT name
TEXT document_type
DATE document_date
TEXT document_date_raw
TEXT location_created
TEXT notes
TEXT archive_identifier
TIMESTAMPTZ created_at
TIMESTAMPTZ updated_at
}
DOCUMENT_PERSON {
UUID id PK
UUID document_id FK
UUID person_id FK
VARCHAR role "author | recipient"
TIMESTAMPTZ created_at
}
JOB {
UUID id PK
UUID document_id FK
VARCHAR status "queued | processing | transcribed | completed | partial_success | failed"
INTEGER retry_count
TEXT provider
TEXT model
TEXT prompt_name
TEXT prompt_hash
TEXT system_prompt
TEXT user_prompt
FLOAT temperature
FLOAT top_p
TIMESTAMPTZ date_created
TIMESTAMPTZ date_updated
}
SOURCE {
UUID id PK
UUID document_id FK
INTEGER page_number
TEXT upload_name
TEXT filename
TEXT file_path
TEXT file_hash
BIGINT file_size_bytes
TEXT raw_transcription
TEXT revised_text
TIMESTAMPTZ date_uploaded
TIMESTAMPTZ date_revised
}
JOB_SOURCE {
UUID id PK
UUID job_id FK
UUID source_id FK
VARCHAR status "pending | transcribed | failed"
TEXT raw_transcription
JSONB ai_metadata
JSONB raw_api_response
TEXT error_detail
TIMESTAMPTZ executed_at
}
DOCUMENT ||--o{ DOCUMENT_PERSON : "has_people"
PERSON ||--o{ DOCUMENT_PERSON : "participates_in"
DOCUMENT ||--o{ JOB : "has_jobs"
DOCUMENT ||--o{ SOURCE : "contains_pages"
JOB ||--o{ JOB_SOURCE : "executes"
SOURCE ||--o{ JOB_SOURCE : "processed_in"
```
## Domain Invariants & Provenance Rules
### Page-Level Execution & AI Outputs
* **Execution Granularity:** Every single image execution attempt by an AI model produces a dedicated record in `job_source`.
* **Submission Snapshot Provenance:** Every `job` captures the frozen prompt identifier details (`prompt_name`, `prompt_hash`), full prompt text strings (`system_prompt`, `user_prompt`), and hyperparameters (`temperature`, `top_p`) at submission time.
* **Point-in-Time Output Auditability:** `job_source.raw_api_response` stores the complete, unedited provider REST response envelope for that specific image page call. `job_source.ai_metadata` stores spatial bounding boxes, normalized token usage, latency, and cost details for fast querying.
* **Active Output Caching:** Upon successful completion of an image call, `source.raw_transcription` is updated with the latest output string from `job_source.raw_transcription` for fast UI rendering.
### Image Storage & Integrity
* **Filesystem Storage:** Binary images are stored on disk in the local file system. The `source` table holds the relative `file_path`.
* **File Integrity Tracking:** `source` captures `file_hash` (SHA-256) and `file_size_bytes` at upload time to guarantee document file integrity and duplicate checking over long-term preservation.
### Page Ordering & Revisions
* **Sequential Integrity:** `source.page_number` dictates page ordering within a document. Reads assembling full documents must query `ORDER BY source.document_id, source.page_number ASC`.
* **Inlined Human Corrections:** User edits occur at the page level inside `source.revised_text`. `source.raw_transcription` remains immutable. If `source.revised_text` is non-null, application frontends must render `source.revised_text`.
### Async Job Lifecycle & Failure Isolation
* **Batch Orchestrator:** A job represents an overarching execution run across one or more source images belonging to a document.
* **Isolated Failures:** API requests run concurrently (e.g., using `asyncio`). A failure on page 3 does not invalidate successful transcriptions on page 1 or 2.
* **Job States:**
* `queued`: Created, awaiting worker execution.
* `processing`: Concurrent HTTP tasks actively running.
* `completed`: 100% of linked `job_source` tasks succeeded (`transcribed`).
* `partial_success`: At least one `job_source` succeeded and at least one failed.
* `failed`: All linked `job_source` tasks failed or a job-level runtime error occurred.
### Attribution & Person Roles
* **Multi-Person Roles:** Documents support zero, one, or many authors and recipients linked via `document_person`.
* **Role Uniqueness:** `(document_id, person_id, role)` must be unique to prevent duplicate role tagging.