- Python 100%
John A. Hoeven
16c592c971
Adds link_discovery.py: sitemap-first, same-domain-crawl-fallback discovery of sub-pages from a single index URL, so a source with many sub-pages no longer needs each page entered by hand. Never pulls page titles or body text into a ticket field, only URLs, to avoid a prompt-injection path into tickets Claude Code later reads. Adds a review-and-approve screen so nothing is queued without a human checking the actual discovered list first. Replaces the free-text licence field with a required typed permission question, since licence itself is better researched from the source during actual processing rather than typed in at ticket time. Replaces the priority dropdown with a single urgent checkbox. Fixes a few real bugs found along the way: colliding ticket filenames when writing many tickets in one batch, a couple of layout issues where long labels or URLs were getting clipped instead of wrapping or scrolling, and dead links slipping through discovery from malformed relative link resolution |
||
|---|---|---|
| .gitignore | ||
| ambrosiana_intake.py | ||
| LICENSE | ||
| link_discovery.py | ||
| README.md | ||
| requirements.txt | ||
| TODO.md | ||
Ambrosiana Document Intake
Status: Beta (v0.1.0)
Document intake UI for the Ambrosiana RAG library.
A minimal Textual TUI that collects the metadata needed to start a document's journey into an Ambrosiana instance's RAG corpus. It does not touch the pipeline, any agent, or any model — its only job is to turn what a human knows about an incoming document into a structured ticket that Claude Code (or a human) then acts on.
Deployed per-device: each Ambrosiana instance gets its own copy, and an intake ticket always targets that device's own RAG — there is no cross-device routing field.
Requirements
- Python 3.11+ (uses stdlib
tomllib) textual(seerequirements.txt)
pip install -r requirements.txt
Running
python3 ambrosiana_intake.py
Configuration
The ticket output directory is read from
~/.config/ambrosiana-intake/config.toml (tickets_root key). If the
config file doesn't exist yet, it's created automatically on first run
with a default of ~/documents/rag-administration/ticket-queue. Edit
the file directly to point at a different location.
Fields
| Field | Required | Type | Notes |
|---|---|---|---|
| Source location | Yes | filepath or URL | |
| Source type | Yes | upstream-doc | ce-authored | ce-experience |
|
| Source tier | Yes | primary | secondary | forum/unmoderated |
|
| Licence | Yes | text | |
| Priority | Yes | immediate need | add to queue |
|
| Title / description | No | text | used for the ticket filename slug if given |
| Notes / context | No | text | free-form, passed through to the ticket body |
Output
Each submission writes a markdown file with YAML front matter directly
into the configured tickets_root — there are no immediate//queue/
subdirectories; priority is encoded in the filename instead:
- Priority "immediate need" →
immediate_YYYYMMDD-HHMMSS_slug.md - Priority "add to queue" →
YYYYMMDD-HHMMSS_slug.md(no tag — the absence of a tag means normal, order-of-arrival handling)
slug comes from the title if provided, otherwise the source
location's basename.
On a successful save the form clears itself and a toast notification confirms the ticket path — no need to scroll to see confirmation.
The ticket is the handoff artefact: a human (or Claude Code, on
request) picks it up from tickets_root and runs the actual
scrape → vet → transform → commit → index pipeline against it.
Keybindings
| Key | Action |
|---|---|
Ctrl+S |
Save ticket |
Ctrl+N |
Clear form |
Ctrl+Q |
Quit |
Built standing on the shoulders of billions of dwarves Created by John A. Hoeven with the ethical assistance of Claude AI