aviconv-aip 1.0.0

Template-driven extraction of airports, runways, frequencies, and navaids from PDF AIPs into schema-valid AIXM 5.1.1.

aviconv-aip

Template-driven extraction of airports, runways, radio frequencies, and navaids from a published PDF AIP into schema-valid AIXM 5.1.1.

Part of the aviconv converter family. This is the v1.0 headline feature: most States publish their AIP only as PDF, and nothing free turns those PDFs into structured AIXM.

What it is — and is not

aviconv-aip is a template-driven, human-in-the-loop extractor, not a magic button. PDF AIP extraction is heuristic — layouts vary between publishers and OCR mis-reads. So the honest deliverable is:

• one declarative template per AIP publisher (adding a publisher means adding a YAML template, not changing extractor code);
• a provenance record on every extracted value — AIP edition, AIRAC cycle, template version, and source page — so a reviewer can audit any value back to the PDF;
• a coarse per-field confidence signal (high = matched table cell, low = OCR/heuristic fallback);
• a human-review report listing every value beside its source page;
• output routed through the released aviconv-aixm writer and the aviconv aixm validate --semantic gate, so an implausible extraction surfaces as a finding before it reaches a user.

Scope is locked to airports/runways/frequencies/navaids — not procedures, not airspace.

Offline / Pyodide boundary

The PDF and OCR dependencies (pdfplumber, optional pytesseract) are not Pyodide-safe. aviconv-aip is therefore a CLI-and-service feature only, never a browser-playground feature. This boundary is locked.

Install

pip install aviconv-aip           # base install (pdfplumber)
pip install "aviconv-aip[ocr]"    # + optional OCR fallback (needs tesseract)

Without the ocr extra, image-only pages are flagged in the extraction result rather than causing a crash.

Usage

aviconv aip list-templates                          # bundled publishers
aviconv aip extract --template pakistan AD-2.pdf     # PDF → records JSON
aviconv aip extract --template pakistan --review AD-2.pdf   # human-review report
aviconv aip to-aixm  --template pakistan AD-2.pdf    # PDF → AIXM 5.1.1

See the package docs and the template-authoring guide.

Source AIP licensing

No source AIP PDF is committed to this repository — only extraction templates and small synthetic, licence-clean fixtures ship. AIP content licensing varies by State; see LICENSES/README.md.