Academic coursework · IT Project / Project 11
Computer Science · Ton Duc Thang University
Reconia
Ongoing Document AI coursework with a document-browsing interface
Project focus
The current team project includes a searchable, filterable document table with formatted metadata and explicit empty states. It supports ongoing invoice-extraction research; the interface scope is document browsing at this stage.
- Information hierarchy
- Search & filters
- Document workflows
- Context
- IT Project · Team coursework
- Timeline
- 2026 — present
- Project status
- In progress
- Tools & technologies
- Python · FastAPI · Jinja2 · HTMX · Document AI
01 / The project
The project
in context.
Reconia is a team IT project comparing approaches to invoice information extraction and line-item recognition. The current repository combines model experiments and evaluation reports with a small document-management interface. It is included as ongoing academic work, with the interface scope limited to what is present in the current implementation.
The user flow
- 01Open the document list
- 02Search by document number
- 03Filter by document type
- 04Inspect document metadata and status
02 / The making
What went
into it.
Readable document records
The document table separates type, number, date, direction, source, amount, and status. Missing metadata has an explicit placeholder, while amounts and dates use consistent formatting.
A lightweight filtering flow
Search and document-type filters update the list through HTMX and preserve the filter URL. An empty state explains that no documents are available and points toward adding one.
Work in progress, clearly scoped
The wider coursework compares extraction models and records evaluation outputs. The portfolio does not imply a finished invoice-review product or an implemented annotation workspace.