scripts/generate_pdf_metadata.py

#!/usr/bin/env python3
"""
generate_pdf_metadata.py
========================
Generates a human-readable PDF metadata report for MatchGeo-DEM.

Requires: fpdf2
Install:  pip install fpdf2

Usage:
  python generate_pdf_metadata.py --output MatchGeo-DEM_Metadata_Report.pdf
"""

import argparse
from pathlib import Path
from datetime import datetime

try:
    from fpdf import FPDF
    HAS_FPDF = True
except ImportError:
    HAS_FPDF = False
    raise ImportError("fpdf2 is required. Install: pip install fpdf2")


CITIES = [
    {"id": "ATA_MV", "name": "Mount Athos, Greece", "epsg": 3031, "resolution": 2.0, "method": "Satellite InSAR", "n_tiles": 5625, "labelled": False, "year": "2011–2015", "provider": "Copernicus DEM (ESA)"},
    {"id": "BRA_SP", "name": "São Paulo, Brazil", "epsg": 31983, "resolution": 0.5, "method": "Airborne LiDAR", "n_tiles": 558, "labelled": True, "year": "2020", "provider": "GeoSampa (PMSP)"},
    {"id": "CHN_WS", "name": "Wutai Shan, China", "epsg": 32649, "resolution": 1.0, "method": "UAV SfM", "n_tiles": 1076, "labelled": False, "year": "2021", "provider": "OpenTopography (Zhou, C.)"},
    {"id": "ESP_EH", "name": "El Hierro, Spain", "epsg": 3040, "resolution": 1.0, "method": "Airborne LiDAR", "n_tiles": 2460, "labelled": False, "year": "2022–2025", "provider": "PNOA-LiDAR (CNIG)"},
    {"id": "FIN_LM", "name": "Lahti, Finland", "epsg": 3067, "resolution": 2.0, "method": "LiDAR + Photogrammetry", "n_tiles": 248, "labelled": False, "year": "2020–2026", "provider": "National Land Survey of Finland"},
    {"id": "GER_BN", "name": "Bonn, Germany", "epsg": 25832, "resolution": 1.0, "method": "Airborne LiDAR", "n_tiles": 1759, "labelled": True, "year": "2016–2018", "provider": "Geobasis NRW"},
    {"id": "IDN_SV", "name": "Sinabung Volcano, Indonesia", "epsg": 32647, "resolution": 0.87, "method": "UAS SfM", "n_tiles": 181, "labelled": False, "year": "2018", "provider": "OpenTopography (Carr, B.)"},
    {"id": "KAZ_AC", "name": "Almaty City, Kazakhstan", "epsg": 32643, "resolution": 1.0, "method": "Pleiades Tristereo", "n_tiles": 887, "labelled": False, "year": "2017", "provider": "OpenTopography (Amey et al.)"},
    {"id": "KSA_WA", "name": "Wadi Al-Akhdar, Saudi Arabia", "epsg": 32637, "resolution": 1.6, "method": "SPOT 6 Stereo", "n_tiles": 3880, "labelled": False, "year": "2016", "provider": "OpenTopography (Matthieu et al.)"},
    {"id": "NAM_HF", "name": "Hebron Fault, Namibia", "epsg": 32733, "resolution": 0.53, "method": "WorldView-3 Stereo", "n_tiles": 1457, "labelled": False, "year": "2017", "provider": "OpenTopography (Salomon et al.)"},
    {"id": "NZL_KP", "name": "Kapiti Coast, New Zealand", "epsg": 2193, "resolution": 1.0, "method": "Airborne LiDAR", "n_tiles": 1776, "labelled": False, "year": "2010–2025", "provider": "LINZ"},
    {"id": "PHL_TA", "name": "Tarlac, Philippines", "epsg": 32651, "resolution": 1.0, "method": "Airborne LiDAR", "n_tiles": 286, "labelled": False, "year": "2014–2017", "provider": "LiPAD (UP Diliman)"},
    {"id": "USA_GC", "name": "Grand Canyon, United States", "epsg": 6341, "resolution": 10.0, "method": "LiDAR + IfSAR", "n_tiles": 600, "labelled": False, "year": "2020–2026", "provider": "USGS 3DEP"},
]


class PDF(FPDF):
    def header(self):
        if self.page_no() == 1:
            return
        self.set_font("Helvetica", "B", 10)
        self.set_text_color(40, 40, 40)
        self.cell(0, 8, "MatchGeo-DEM Dataset Metadata Report", border=0, align="L")
        self.cell(0, 8, f"Page {self.page_no()}", border=0, align="R")
        self.ln(10)
        self.set_draw_color(180, 180, 180)
        self.line(10, self.get_y(), 200, self.get_y())
        self.ln(5)

    def footer(self):
        self.set_y(-15)
        self.set_font("Helvetica", "I", 8)
        self.set_text_color(128, 128, 128)
        self.cell(0, 10, f"Generated: {datetime.now().strftime('%Y-%m-%d %H:%M')} | DOI: 10.5281/zenodo.19339008", align="C")

    def chapter_title(self, title, level=1):
        if level == 1:
            self.set_font("Helvetica", "B", 16)
            self.set_text_color(0, 51, 102)
            self.ln(8)
            self.cell(0, 10, title, ln=True)
            self.set_draw_color(0, 51, 102)
            self.line(10, self.get_y(), 200, self.get_y())
            self.ln(6)
        else:
            self.set_font("Helvetica", "B", 12)
            self.set_text_color(0, 51, 102)
            self.ln(6)
            self.cell(0, 8, title, ln=True)
            self.ln(2)

    def body_text(self, text, bold=False):
        self.set_font("Helvetica", "B" if bold else "", 10)
        self.set_text_color(40, 40, 40)
        self.multi_cell(0, 5, text)
        self.ln(2)

    def info_row(self, label, value):
        self.set_font("Helvetica", "B", 10)
        self.set_text_color(60, 60, 60)
        self.cell(50, 6, label + ":", align="L")
        self.set_font("Helvetica", "", 10)
        self.set_text_color(40, 40, 40)
        self.cell(0, 6, str(value), align="L")
        self.ln()


def generate_pdf(output_path):
    pdf = PDF()
    pdf.set_auto_page_break(auto=True, margin=15)
    pdf.add_page()

    # ===== COVER PAGE =====
    pdf.set_font("Helvetica", "B", 24)
    pdf.set_text_color(0, 51, 102)
    pdf.ln(40)
    pdf.cell(0, 15, "MatchGeo-DEM", ln=True, align="C")
    pdf.set_font("Helvetica", "", 14)
    pdf.cell(0, 10, "Multi-City Digital Elevation Model Dataset", ln=True, align="C")
    pdf.cell(0, 10, "for Local Feature Matching", ln=True, align="C")
    pdf.ln(20)

    pdf.set_font("Helvetica", "", 11)
    pdf.set_text_color(80, 80, 80)
    pdf.multi_cell(0, 6, 
        "This report provides a human-readable summary of the MatchGeo-DEM dataset "
        "(Version 1.1), including its structure, provenance, licensing, and per-city coverage. "
        "It is intended for data managers, reviewers, and users who need a quick reference "
        "without opening machine-readable metadata files.",
        align="C"
    )
    pdf.ln(30)

    pdf.set_font("Helvetica", "B", 11)
    pdf.set_text_color(40, 40, 40)
    pdf.cell(0, 8, "Dataset DOI: 10.5281/zenodo.19339008", ln=True, align="C")
    pdf.cell(0, 8, "License: CC BY 4.0", ln=True, align="C")
    pdf.cell(0, 8, f"Report Date: {datetime.now().strftime('%Y-%m-%d')}", ln=True, align="C")
    pdf.cell(0, 8, "Contact: sabrina.correa@ufv.br", ln=True, align="C")

    # ===== OVERVIEW =====
    pdf.add_page()
    pdf.chapter_title("1. Dataset Overview")
    pdf.body_text(
        "MatchGeo is a curated, multi-city Digital Elevation Model (DEM) dataset designed for "
        "training and benchmarking local feature matching algorithms in urban and natural terrain analysis. "
        "It aggregates high-resolution elevation data from 13 distinct environments across 6 continents, "
        "supporting cross-domain generalization studies under varying acquisition methods, climates, and terrain types."
    )

    pdf.info_row("Title", "MatchGeo: Multi-City DEM Dataset for Local Feature Matching")
    pdf.info_row("Version", "1.1")
    pdf.info_row("Release Date", "2026-05-11")
    pdf.info_row("Total Cities", "13")
    pdf.info_row("Total Tiles", "20,793")
    pdf.info_row("Labelled Tiles", "213 (Bonn, Germany)")
    pdf.info_row("Total Annotations", "27,000+ (handcrafted keypoints)")
    pdf.info_row("Tile Size", "333 × 333 pixels")
    pdf.info_row("Pixel Depth", "Float32")
    pdf.info_row("NoData Value", "-9999")
    pdf.info_row("Compression", "DEFLATE")
    pdf.info_row("Format", "GeoTIFF (BigTIFF, tiled, OGC 23-008r3 compliant)")
    pdf.ln(5)

    pdf.chapter_title("2. Authors & Contact", level=2)
    pdf.info_row("Authors", "Correa, S. P. L. P.; Santos, A. de Paula; Oliveira, H. N.; Beltons, D.")
    pdf.info_row("Institution", "Universidade Federal de Viçosa (UFV)")
    pdf.info_row("Contact", "sabrina.correa@ufv.br")
    pdf.info_row("Repository", "https://doi.org/10.5281/zenodo.19339008")
    pdf.ln(5)

    # ===== LICENSE =====
    pdf.chapter_title("3. License & Attribution", level=2)
    pdf.body_text(
        "This dataset is released under the Creative Commons Attribution 4.0 International License (CC BY 4.0). "
        "You are free to share and adapt the material for any purpose, even commercially, provided you give "
        "appropriate credit, provide a link to the license, and indicate if changes were made."
    )
    pdf.body_text(
        "When using this dataset, you must cite the dataset DOI and acknowledge the original data providers "
        "for each city used in your study. Full attribution statements are provided in the DATASET_DESCRIPTION.md file."
    )
    pdf.ln(5)

    # ===== TECHNICAL SPECIFICATIONS =====
    pdf.chapter_title("4. Technical Specifications", level=2)
    pdf.info_row("Raster Format", "GeoTIFF (BigTIFF variant)")
    pdf.info_row("Internal Tiling", "256 × 256 pixels")
    pdf.info_row("Patch Dimensions", "333 × 333 pixels")
    pdf.info_row("Data Type", "Float32")
    pdf.info_row("Coordinate Systems", "City-specific UTM / local CRS (EPSG)")
    pdf.info_row("Metadata Standard", "ISO 19115-2 + OGC 23-008r3")
    pdf.info_row("Machine Catalog", "JSON-LD manifest.json + STAC 1.0.0 collection")
    pdf.ln(5)

    # ===== PROCESSING PIPELINE =====
    pdf.chapter_title("5. Processing Pipeline", level=2)
    pdf.body_text(
        "All cities were processed through a standardized PDAL 2.6.0 pipeline with city-specific adaptations:"
    )
    steps = [
        "1. Acquisition — Raw data retrieved from source portals in native CRS and resolution.",
        "2. Preprocessing — City-specific filtering (ground classification, outlier removal, noise filtering).",
        "3. Rasterization — PDAL writers.gdal with output_type=max (DSM), float32, nodata=-9999.",
        "4. Standardization — BigTIFF, TILED=YES, COMPRESS=DEFLATE.",
        "5. Patch Extraction — Non-overlapping 333×333 pixel grid (no resampling).",
        "6. Annotation — Handcrafted keypoints in normalized coordinates (Bonn, São Paulo).",
        "7. Metadata — Per-city ISO 19115-2 JSON; central JSON-LD manifest.",
    ]
    for step in steps:
        pdf.body_text(step)
    pdf.ln(5)

    # ===== PER-CITY TABLE =====
    pdf.add_page()
    pdf.chapter_title("6. Per-City Coverage")
    pdf.body_text(
        "The following table summarizes each city's geographic coverage, acquisition method, resolution, "
        "and annotation status. All tiles are 333×333 pixel GeoTIFF patches."
    )
    pdf.ln(3)

    # Table header
    pdf.set_fill_color(0, 51, 102)
    pdf.set_text_color(255, 255, 255)
    pdf.set_font("Helvetica", "B", 9)
    pdf.cell(22, 7, "City", fill=True)
    pdf.cell(45, 7, "Location", fill=True)
    pdf.cell(30, 7, "Method", fill=True)
    pdf.cell(18, 7, "Res (m)", fill=True)
    pdf.cell(18, 7, "Tiles", fill=True)
    pdf.cell(20, 7, "Labelled", fill=True)
    pdf.cell(25, 7, "Year", fill=True)
    pdf.ln()

    # Table rows
    pdf.set_text_color(40, 40, 40)
    pdf.set_font("Helvetica", "", 8)
    fill = False
    for city in CITIES:
        if pdf.get_y() > 260:
            pdf.add_page()
            pdf.set_fill_color(0, 51, 102)
            pdf.set_text_color(255, 255, 255)
            pdf.set_font("Helvetica", "B", 9)
            pdf.cell(22, 7, "City", fill=True)
            pdf.cell(45, 7, "Location", fill=True)
            pdf.cell(30, 7, "Method", fill=True)
            pdf.cell(18, 7, "Res (m)", fill=True)
            pdf.cell(18, 7, "Tiles", fill=True)
            pdf.cell(20, 7, "Labelled", fill=True)
            pdf.cell(25, 7, "Year", fill=True)
            pdf.ln()
            pdf.set_text_color(40, 40, 40)
            pdf.set_font("Helvetica", "", 8)
            fill = False

        pdf.set_fill_color(240, 240, 240) if fill else pdf.set_fill_color(255, 255, 255)
        pdf.cell(22, 6, city["id"], fill=True)
        pdf.cell(45, 6, city["name"], fill=True)
        pdf.cell(30, 6, city["method"], fill=True)
        pdf.cell(18, 6, str(city["resolution"]), fill=True)
        pdf.cell(18, 6, str(city["n_tiles"]), fill=True)
        pdf.cell(20, 6, "Yes" if city["labelled"] else "No", fill=True)
        pdf.cell(25, 6, city["year"], fill=True)
        pdf.ln()
        fill = not fill

    pdf.ln(5)
    pdf.set_font("Helvetica", "I", 8)
    pdf.set_text_color(100, 100, 100)
    pdf.multi_cell(0, 4,
        "Note: BRA_SP is marked as labelled in the dataset schema but currently has zero annotation files "
        "(annotations pending). GER_BN contains 27,000+ handcrafted keypoint annotations.")

    # ===== DATA SPLITS =====
    pdf.add_page()
    pdf.chapter_title("7. Data Splits")
    pdf.body_text(
        "The dataset is partitioned into train / validation / test splits stratified by city and difficulty. "
        "Split manifests are provided as CSV files in the splits/ directory."
    )
    pdf.info_row("Train", "80%")
    pdf.info_row("Validation", "10%")
    pdf.info_row("Test", "10%")
    pdf.info_row("Stratification", "By city and difficulty (flat, medium, steep, urban_density)")
    pdf.ln(5)

    # ===== KNOWN LIMITATIONS =====
    pdf.chapter_title("8. Known Limitations", level=2)
    limitations = [
        "• Geographic bias: Dense annotations are currently available only for GER_BN. BRA_SP annotations are pending.",
        "• Temporal mismatch: Data spans 2011–2026 across cities; users should account for temporal drift.",
        "• Sensor heterogeneity: LiDAR, photogrammetry, SfM, and satellite stereo have different noise characteristics.",
        "• Resolution heterogeneity: Native resolutions range from 0.5 m to 10 m; all tiles are 333×333 pixels.",
        "• Missing data: Water bodies and ocean areas are excluded (NoData = -9999).",
    ]
    for lim in limitations:
        pdf.body_text(lim)
    pdf.ln(5)

    # ===== CITATION =====
    pdf.chapter_title("9. How to Cite", level=2)
    pdf.body_text("Dataset citation (BibTeX):", bold=True)
    pdf.set_font("Courier", "", 8)
    pdf.set_text_color(40, 40, 40)
    bibtex = """@dataset{correa_2026_matchgeo,
  author = {Correa, S. P. L. P. and Santos, A. de Paula and Oliveira, H. N. and Beltons, D.},
  title  = {MatchGeo: Multi-City Digital Elevation Model Dataset for Local Feature Matching},
  year   = 2026,
  publisher = {Zenodo},
  version = {1.1},
  doi    = {10.5281/zenodo.19339008},
  url    = {https://doi.org/10.5281/zenodo.19339008}
}"""
    pdf.multi_cell(0, 4, bibtex)
    pdf.ln(5)

    pdf.set_font("Helvetica", "", 10)
    pdf.body_text("Plain text citation:", bold=True)
    pdf.body_text(
        "Correa, S. P. L. P., Santos, A. de Paula, Oliveira, H. N., & Beltons, D. (2026). "
        "MatchGeo: Multi-City Digital Elevation Model Dataset for Local Feature Matching (Version 1.1) [Data set]. "
        "Zenodo. https://doi.org/10.5281/zenodo.19339008"
    )

    # ===== BACK PAGE =====
    pdf.add_page()
    pdf.set_font("Helvetica", "B", 14)
    pdf.set_text_color(0, 51, 102)
    pdf.ln(80)
    pdf.cell(0, 10, "End of Report", ln=True, align="C")
    pdf.set_font("Helvetica", "", 10)
    pdf.set_text_color(100, 100, 100)
    pdf.cell(0, 8, "For questions or bug reports, contact: sabrina.correa@ufv.br", ln=True, align="C")
    pdf.cell(0, 8, "Zenodo: https://doi.org/10.5281/zenodo.19339008", ln=True, align="C")
    pdf.cell(0, 8, "Hugging Face: https://huggingface.co/datasets/paeslemesa/matchgeo", ln=True, align="C")

    # Save
    pdf.output(output_path)
    print(f"✅ PDF report saved: {output_path}")


def main():
    parser = argparse.ArgumentParser(description="Generate MatchGeo-DEM PDF metadata report")
    parser.add_argument("--output", default="MatchGeo-DEM_Metadata_Report.pdf", help="Output PDF path")
    args = parser.parse_args()
    generate_pdf(args.output)


if __name__ == "__main__":
    main()