Projective Drawing Database

Rajzvizsgálati adatbázis / Projective Drawing Database ! Vass, Z. (2025). Rajzvizsgálati adatbázis / Projective Drawing Database (Version 1.0) [Database]. Károli Gáspár Reformed University. https://krepozit.kre.hu/ Database Documentation Last updated: 2025-10-10 What this database is ! Rajzvizsgálati adatbázis (Projective Drawing Database) – a comprehensive methodology for building thematic databases of projective drawing analysis from validated case studies. Timeframe: January 6, 2025 - May 15, 2025 Delivery format: Multi-component digital system including: Searchable JSON database ( patterns.json ) Interactive HTML search interface ( search_interface.html ) RAG (Retrieval-Augmented Generation) search engine LLM-based pattern extraction pipeline Comprehensive sign indexing system Source materials: Case studies and professional documentation collected over five years Author and affiliation: see Metadata below What the database contains (methodology and schema) ! Core Components ! The database methodology encompasses the following structured units: 1. Validated, anonymized case studies Full compliance with GDPR requirements Complete anonymization protocols Professional-ethical guidelines for data usage 2. Diagnostically instructive case collections Educationally valuable case presentations Rare or exceptional drawing assessment phenomena Typical case demonstrations for teaching purposes Sample analyses supporting diagnostic practice 3. Methodological framework Structured template for drawing assessment documentation Systematic pattern extraction procedures Sign co-occurrence analysis protocols Confidentiality requirement adherence guidelines Database Schema (patterns.json) ! Top-level JSON structure: metadata: object title: string ("Rajzvizsgálati adatbázis / Projective Drawing Database") author: string ("Vass, Zoltan") year: string ("2025") version: string ("1.0") publisher: string ("Károli Gáspár Reformed University") location: string ("Budapest, Hungary") url: string ("https://krepozit.kre.hu/") citation: string (APA 7 formatted) total_source documents: number (currently partially processed) total_patterns: number (currently 240) total_reference_signs: number (1,116 indexed signs) categories: array (11 sign categories) patterns: object<string, object> Key: pattern ID (normalized combination of sign IDs, e.g., "20_146_996") Value: object containing: sign_ids: array (IDs of co-occurring signs) sign_names: array (descriptive names of signs) occurrences: array (source document IDs where this pattern appears) interpretations: array (contextual interpretations from source documents) frequency: number (count of occurrences) comprehensive_sign_index: object<string, object> Key: sign ID as string (e.g., "990", "1087", "1112") Value: object containing: id: number (unique sign identifier from reference file) name: string (sign description in Hungarian) category: string (one of 11 categories: ANAMNÉZIS, EMBERRAJZ, ÁLLATRAJZ, FARAJZ, HÁZRAJZ, SZABADRAJZ, AKTUÁLGENEZIS, INTUITÍV ELEMZÉS, EGÉSZLEGES ELEMZÉS, FORMAI-SZERKEZETI JELLEMZŐK, TARTALMI-SZIMBOLIKUS JELLEMZŐK) patterns: array (pattern IDs containing this sign) source documents: array (source document IDs mentioning this sign) Reference Sign Categories (11 categories) ! 1. ANAMNÉZIS (Anamnestic Data) Demographic information Psychological diagnosis Social behavior Emotional characteristics Cognitive functioning Somatic diseases 2. AKTUÁLGENEZIS (Process Analysis) Verbal behavior during testing Gestures and facial expressions Test-taking behavior Psychomotor characteristics Drawing time and reaction time Drawing sequence Erasing patterns 3. INTUITÍV ELEMZÉS (Intuitive Analysis) Receptive observation methods Spontaneous attention focus Motor empathy Kinesthetic empathy Empathic questioning 4. EGÉSZLEGES ELEMZÉS (Holistic Analysis) Emotional-mood tone Integration and harmony Hárdi drawing personality levels Movement and form imagery Color usage Elkisch analysis framework 5. FORMAI-SZERKEZETI JELLEMZŐK (Formal-Structural Characteristics) Size and proportions Symmetry Position on page Line quality Shading Detail level Transparency 6. TARTALMI-SZIMBOLIKUS JELLEMZŐK (Content-Symbolic Characteristics) Eleven areas of content-symbolic analysis Iconic identification of motifs Semantic connections and interactions Cultural and social influences Personal relationships Psychoanalytic symbol analysis 7. HÁZRAJZ (House Drawing) Thematic stereotypes Door characteristics Windows Roof and chimney Walls Stairs and pathways Environment and accessories 8. FARAJZ (Tree Drawing) Holistic tree characteristics Thematic stereotypes Tree species Foliage and branches Tree trunk Roots Hollow Special markings 9. ÁLLATRAJZ (Animal Drawing) Animal species categories Movement and posture Head and features Body trunk Limbs Latent animal imagery 10. EMBERRAJZ (Human Figure Drawing) Gender representation Age characteristics Thematic stereotypes Posture and movement Head and neck Facial features Body trunk Arms, hands, fingers Legs, feet, toes Hair and body hair Shoulders Chest Hip and genital area Clothing and accessories 11. SZABADRAJZ (Free Drawing) Thematic stereotypes Typical representation forms Accessory elements (ground line, background, objects, natural elements) Current statistics (from patterns.json) ! Total reference signs: 1,116 (all indexed and searchable) Processed source documents: (654 available, currently partially processed) Extracted patterns: 240 (co-occurrence patterns of 3+ signs) Signs appearing in patterns: 20 signs Signs searchable but not yet in patterns: 1,096 signs Sign categories: 11 major categories Database file size: approximately 540 KB (JSON format) Methodology and processing pipeline ! 1. Source Data Collection ! Input: CSV file containing source document data with the following fields: Title (Hungarian) Summary/Abstract (Hungarian) Analysis/Discussion (Hungarian) Metadata (year, author, etc.) Current dataset: 654 source documents available for processing 2. LLM-Based Sign Extraction ! Technology stack: Extraction method: Structured prompt-based extraction Reference library: 1,305 unique sign IDs from reference file Language: Hungarian (all processing in native language) Extraction process ( DrawingPatternExtractor ): 1. Load source document text (summary + analysis) 2. Submit to LLM with reference sign library 3. Extract: Identified sign IDs mentioned in text Contextual interpretations Demographics (age, gender, occupation, anamnestic features) 4. Validate extracted sign IDs against reference library 5. Cache results to avoid re-processing Batch processing ( PatternDatabaseBuilder ): Processes source documents in configurable batches (default: 10) Manual checkpoints between batches Progress tracking with visual indicators Extraction cache for efficiency 3. Pattern Discovery ! Co-occurrence analysis: Minimum pattern size: 3 signs appearing together Pattern identification: Normalized combination of sign IDs Frequency tracking: Count of occurrences across source documents Interpretation aggregation: Collect contextual meanings Example pattern: 4. Comprehensive Sign Indexing ! Reference library parsing ( build_comprehensive_sign_index.py ): 1. Parse drawing-and-anamnestic-signs-features.txt 2. Extract 1,116 unique signs with categories 3. Assign sequential IDs (1-1305, with gaps for hierarchical structure) 4. Index all signs regardless of source document occurrence 5. Link signs to patterns (when applicable) 6. Link signs to source documents (when mentioned) Result: Every reference sign is searchable, whether or not it has been found in processed source documents yet. 5. Search Engine ! Dual search functionality ( RAGSearchEngine ): Pattern search ( search_patterns() ): Searches across 240 co-occurrence patterns Returns matches with full source document context Includes interpretations and frequency data Sign search ( search_signs() ): { "20_146_996": { "sign_ids": [20, 146, 996], "sign_names": [ "alacsony szocioökonómiai státus", "kényszeres személyiségvonások", "fejábrázolás egyszerű körsémával" ], "occurrences": ["source document_1", "source document_3"], "interpretations": [ "Simple drawing style correlating with socioeconomic factors", "Obsessive-compulsive traits reflected in controlled circular forms" ], "frequency": 2 } } Searches all 1,116 reference signs Keyword-based matching (sign name, category) Relevance scoring (exact match > partial match > keyword match) Returns sign metadata + pattern/source document links Example searches: ujj (finger) → finds 10+ finger-related signs határozott (determined) → finds 3 patterns + 5 individual signs FARAJZ (tree drawing) → finds all tree-related signs anamnézis → finds anamnestic data signs 6. Interactive HTML Interface ! Features ( search_interface.html ): Real-time search with dual result display Pattern matches section (with source document examples) Individual sign matches section (all reference signs) Example query buttons for quick exploration Statistics dashboard Responsive design for readability Technology: Single-file HTML with embedded JavaScript Fetches data from JSON via local server Client-side search and filtering No external dependencies 7. Deployment ! Local server ( serve_search.py ): Server features: Simple HTTP server for local development Auto-opens browser on start Serves JSON database and HTML interface CORS-enabled for local testing Design principles and quality assurance ! Professional-Ethical Framework ! 1. Anonymization protocols: Complete removal of personal identifiers python3 serve_search.py # Opens browser at: http://localhost:8000/search_interface.html GDPR-compliant data handling Secure storage and access controls 2. Confidentiality requirements: Professional use only (research, education) No distribution of raw case data Aggregated patterns for public presentation 3. Data validation: Expert review of extracted patterns Cross-validation with reference literature Manual verification of diagnostic correlations Technical Quality Standards ! 1. LLM extraction accuracy: Reference library validation (all sign IDs checked) Contextual interpretation preservation Hungarian language processing fidelity 2. Database integrity: Unique pattern identification Consistent source document ID mapping Traceability to source data (CSV row numbers) 3. Search performance: Keyword-based relevance scoring Fast client-side filtering Comprehensive coverage (all 1,116 signs) 4. Extensibility: Modular pipeline design Incremental processing capability Merge-friendly database structure Methodological documentation template ! The database provides a structured template for drawing assessment documentation, based on the eleven-category framework from drawing-and-anamnestic-signs-features.txt : I. Anamnestic Data (ANAMNÉZIS) ! Demographics: age, gender, occupation Psychological diagnosis (DSM-IV, BNO-10) Personality characteristics Social behavior patterns Cognitive functioning Somatic conditions II. Process Analysis (AKTUÁLGENEZIS) ! Verbal behavior during testing Gestures and facial expressions Test-taking behavior Psychomotor characteristics Temporal features (drawing time, reaction time) Drawing sequence and modifications Erasing patterns III. Intuitive Analysis (INTUITÍV ELEMZÉS) ! Initial impressions Spontaneous attention focus Motor and kinesthetic empathy Visualization techniques Empathic questioning IV. Holistic Analysis (EGÉSZLEGES ELEMZÉS) ! Emotional-mood tone Integration and harmony levels Drawing personality levels (Hárdi framework) Movement and form emphasis Color usage patterns Elkisch analysis dimensions V. Formal-Structural Characteristics (FORMAI-SZERKEZETI JELLEMZŐK) ! Size and proportions Symmetry features Position on page Line quality Shading techniques Detail level Transparency effects VI. Content-Symbolic Characteristics (TARTALMI-SZIMBOLIKUS JELLEMZŐK) ! Iconic identification Semantic connections Cultural influences Personal relationships Symbol analysis (archetypal, cultural, individual)