bluegreencity-ki

Citree Species Extraction

Overview

This workflow extracts all species profiles and all available criteria from the public Citree website via its internal AJAX endpoints.

The extraction currently uses:

• AJAX/ArtSearch.php with an empty search term to enumerate all available species profiles and Citree IDs
• AJAX/getSteckbrief.php to fetch the detailed criteria tables for each profile

The resulting files are written to data/derived/species:

• citree_species_criteria.csv
• citree_species_criteria.json

Script

Script entry point:

• scripts/extract_citree_species_catalog.py

Reusable parser and HTTP logic:

• bluegreencity_ki/io/citree_species_catalog.py

Extracted Fields

The CSV contains one row per Citree profile and one column per extracted criterion.

Metadata columns:

• art_id
• citree_name
• deutscher_name
• botanischer_name
• sorte
• source_url

Criteria columns include, among others:

• Lichtanspruch
• Trockenheitstoleranz
• Hitzeverträglichkeit
• Bodenverdichtungstoleranz
• Staunässetoleranz
• Salzverträglichkeit
• Bodenfeuchtetoleranz
• Bodeneigenschaft
• Gründigkeit
• Krankheiten / Schädlinge
• Wuchsgeschwindigkeit
• Kronendurchlässigkeit
• Kronenform
• oberflächennahe Wurzeln

The extractor also keeps all additional criteria that appear in the Citree detail tables.

Label Normalization

Citree uses inconsistent label spellings in some places, for example around umlauts or separators.

The extractor normalizes known variants to stable output columns, for example:

• Krankheiten/Schädlinge → Krankheiten / Schädlinge
• Kronendurchlaessigkeit → Kronendurchlässigkeit
• Gruendigkeit → Gründigkeit

Unknown labels are preserved as they appear after whitespace normalization.

Usage

Run from the project root:

python scripts/extract_citree_species_catalog.py

Optional test run:

python scripts/extract_citree_species_catalog.py --limit 10 --delay-seconds 0

Output Format

CSV

Flat table for downstream joins, manual inspection, and import into other tools.

JSON

The JSON file stores:

• extraction metadata
• the flat criteria dictionary per profile
• the original parsed criteria entries including panel, section, raw label, normalized label, value, and row description

Notes

• The Citree endpoints are not an officially documented public API.
• Responses are HTML fragments and require parsing.
• The extractor uses a short configurable delay between detail requests to avoid unnecessary load on the source website.
• The website footer references a Creative Commons BY-NC-SA 4.0 license; downstream reuse should respect that license.