Technical Architecture

1. Overview

Single docker container Go application built on PocketBase as an embedded framework (https://pocketbase.io/). The binary contains:

PocketBase's HTTP/API server and admin dashboard;
the custom review dashboard;
the S3 poller and the PDF processing pipeline, executed through PocketBase's standard job scheduler (app.Cron());
the declarative rule engine (JSONC config + expr-lang/expr expressions);
access to the self-hosted S3 storage via the MinIO Go SDK.

Everything runs offline on the server infrastructure: no cloud APIs, no outbound calls other than to the S3 cluster.

2. Stack

Component	Choice
Framework / runtime	PocketBase (Go framework, single binary)
Language	Go
Scheduling	`app.Cron()` (S3 poller, periodic tasks)
S3	MinIO Go SDK (`github.com/minio/minio-go/v7`)
PDF text extraction	`github.com/ledongthuc/pdf` (native text, no OCR)
Rule expressions	`github.com/expr-lang/expr`
Regex	stdlib `regexp` (RE2); `regexp2` only if a lookaround proves indispensable
Config validation	`santhosh-tekuri/jsonschema` (schema generated from structs via `invopop/jsonschema`)
Database	SQLite (embedded in PocketBase)
Full-text search	SQLite FTS5 index synchronized via hooks

3. S3 Layout

A single versioned bucket:

inbox/       # drop zone for new PDFs (via Nextcloud / S3 client)
pending/     # files awaiting human review
files/       # renamed and filed files (configured prefix structure)

Nextcloud mounts the bucket as external storage: it is the user-facing interface for uploading, downloading, and sharing. The application only manipulates S3 keys, never a local filesystem for documents.

4. Data Model (PocketBase Collections)

`documents`

One row per processed PDF file:

sha256: the sha256 is a primary key as the system will reject as duplicate any file with the same hash
original_filename: original file name in inbox/
extracted_text: full text extracted from all pages (base of the FTS5 index)
first_page_text: first-page text (extraction base)
size: file size
processed_at: timestamp of the file initial process
status: pending | processed | validated
updated_at: timestamp of the last update to the file metadata

all the fields declared by the user in the configuration file

The append-only document_events collection is the single source of truth for a document's history; no history is stored on the documents record itself.

`document_events`

Append-only history: document, event_type (classified, moved, corrected, validated), payload as JSON, timestamp.

5. Configuration File (JSONC)

A config.jsonc file kept alongside the binary, reloaded on restart. Three sections:

fields: declaration of fields to extract from the files:
- name: name of the field
- type: string | number | boolean
- mandatory: true | false
rules: ordered list of rules that will be evaluated one after the other:
- when: an expr-lang/expr expression evaluated over the document text and the current field values; rules run in order and each when may reference fields set by earlier rules (e.g. firstPageText contains "Relevé de compte", or emittingCompany in lists.banks && documentType == "Relevé de compte")
- actions:
  - set: set the value of one field; fields accumulate as rules are applied in order
  - stop: stop rule evaluation; the fields set so far are kept and the usual validations still apply
templates: naming templates per document type (e.g. {sender}/{date}_{type}_{reference}.pdf), with required-field validation — a template with a missing field renders the result invalid and routes the file to pending/.
lists: optional named constant lists (e.g. banks, account numbers) usable in when expressions as lists.<name>.

Validation on load: JSON schema, regex compilation, expr expression compilation against a typed context (Document{FirstPageText, ExtractedText, Fields, Lists}). Full rejection if invalid, no partial processing.

6. Processing Pipeline

Executed by the PocketBase cron job (configurable interval, e.g. every minute):

Poll: list objects in inbox/ not yet processed (detection by key recorded in the database). Write-completion wait: an object is a candidate only if its size and ETag are stable across two consecutive polls.
Download: fetch the object into memory (streamed), no disk writes. The sha256 is computed during download; if a documents record with the same hash already exists, the object is a duplicate: the inbox/ key is deleted (delete marker, data recoverable), a duplicate event is logged, and processing stops there.
Extraction: first-page text (for rule evaluation) + full text (for FTS) via ledongthuc/pdf, including headers and footers
Rule evaluation: rules evaluated in order against the extracted text and the fields set so far; full log stored in rule_trace (built-in dry-run mode for debugging from the dashboard).
Validation: a result is valid if and only if: all rules have been processed with no error, all mandatory fields are identified, and the naming template evaluated with no error. Otherwise: route directly to pending/.
Filing:
- valid → copy to the new key (prefix + templated name), verify the target object exists, then delete the inbox/ key only after the copy is confirmed; status processed. If a file with the same name already exists, append a number to the filename to avoid overwriting.
- invalid → move to pending/ (incremental suffix on name conflict); status pending.
Indexing: write the documents record, add an event to the file history including the original filename.

The bucket is versioned: any inbox/ object deletion creates a delete marker, the data remains recoverable.

7. Review Dashboard

Custom pages on the PocketBase instance:

File list: all files with status, filters, combined search over full text (FTS5) + metadata/fields.
Detail: PDF preview streamed from S3 through a dedicated route, along with extracted fields and rule_trace visible in debug mode.
Correction: forms to edit fields, type, and sender; correcting a processed file re-files it (copy to the new key, delete the old one, history preserved). A corrected pending file follows the same path: once validated, it is copied to its templated key in files/ (incremental suffix on conflict) and removed from pending/.
Validation: transition to validated; processed files can also be validated. Validation is what takes a corrected pending file out of the review queue.
Rule management: config display and dry-run on an existing document. The configuration is reloaded only on restart; there is no runtime reload.

Authentication: single account via PocketBase's standard auth.

8. Error Handling

Transient S3 error: retry with backoff, the object stays in inbox/ and will be polled again.
Unreadable PDF: pending status, reason recorded.
Target key conflict (two files renamed identically, whether filing to files/ or moving to pending/): automatic incremental suffix, event logged.
Invalid config: error message displayed and the application does not start.

7.3 KiB Raw Blame History