Changelog

Unreleased

  • Bug fixes

    • Fix geoextent --version (and the version recorded in exported output / extraction metadata) to report the installed package version rather than a version inferred from the current working directory.

    • Fix --placename returning no placename for remote sources, and the missing placename property in the resulting --geojsonio URL. Also fix an AttributeError crash when --geojsonio is combined with from_remote.

    • Fix --placename combined with --convex-hull on merged geometries (every from_remote call, multi-file directories, multi-layer vectors). convex_hull_merge emits the merged hull as a GeoJSON Polygon dict, but the placename block in from_file / from_directory / from_remote only knew the 4-tuple and list-of-pairs shapes, so the dict fell through to the bbox path and produced “Invalid bounding box format for placename extraction”.

  • Breaking change — canonical output shape

    • Public output of from_file / from_directory / from_remote / from_text switches from {"bbox": ..., "crs": "4326", ...} to {"envelope": [minx, miny, maxx, maxy], "geometry": <GeoJSON Polygon>, "convex_hull": <bool>, ...}. crs is no longer surfaced; geoextent always emits EPSG:4326 (which RFC 7946 mandates for GeoJSON anyway).

    • The merge shape (--convex-hull output) no longer overloads bbox with a GeoJSON Polygon dict — envelope is always the 4-tuple, geometry is always the Polygon dict.

    • Migration: downstream code reading result["bbox"] / result["crs"] should read result["envelope"] / result["geometry"] (or just rely on the GeoJSON output, which is unchanged for end users).

0.13.0

  • New Content Providers

    • Generalise Zenodo provider into InvenioRDM base provider supporting CaltechDATA, TU Wien, Frei-Data, GEO Knowledge Hub, TU Graz, Materials Cloud Archive, FDAT, DataPLANT ARChive, KTH, Prism, and NYU Ultraviolet (#81)

    • Add Mendeley Data as content provider (#58)

    • Add ioerDATA (Leibniz IOER) Dataverse instance support (#85)

    • Add heiDATA (Heidelberg University) Dataverse instance support (#94)

    • Add Edmond (Max Planck Society) Dataverse instance support (#93)

    • Add Wikidata as content provider for extracting geographic extents from Wikidata items via SPARQL (#83)

    • Add 4TU.ResearchData as content provider with support for both data download and metadata-only extraction (#92)

    • Add RADAR (FIZ Karlsruhe) as content provider for cross-disciplinary German research data (#87)

    • Add NSF Arctic Data Center as content provider with metadata-only and data download support (#90)

    • Add DEIMS-SDR (Dynamic Ecological Information Management System) as metadata-only content provider for long-term ecological research sites and datasets, with support for dataset and site URLs, WKT geometry parsing (POINT, POLYGON, MULTIPOLYGON), and temporal ranges (#101)

    • Add BAW (Bundesanstalt für Wasserbau) Datenrepository as content provider with CSW 2.0.2 metadata extraction via OWSLib, supporting DOIs (10.48437/*), landing page URLs, and bare UUIDs (#89)

    • Add B2SHARE (EUDAT) as InvenioRDM instance, supporting DOIs (10.23728/b2share.*), record URLs, and old-style hex DOIs (#16)

    • Add MDI-DE (Marine Daten-Infrastruktur Deutschland) as content provider with CSW 2.0.2 metadata extraction via OWSLib, WFS-based data download, and direct file download; supports NOKIS landing page URLs and bare UUIDs (#86)

    • Add HALO DB (DLR) as metadata-only content provider for the HALO research aircraft database, extracting flight track geometry from the GeoJSON search API and temporal extent from dataset HTML pages (#88)

    • Add GBIF (Global Biodiversity Information Facility) as content provider with metadata-only extraction from the Registry API and optional Darwin Core Archive (DwC-A) data download from institutional IPT servers; supports DOI prefixes 10.15468, 10.15470, 10.15472, 10.25607, 10.71819, 10.82144 and gbif.org dataset URLs (#62)

    • Add GeoScienceWorld as metadata-only content provider for geoscience journal articles, extracting geographic coordinates from GeoRef metadata (WKT POLYGON/POINT) embedded in article landing pages; supports article URLs, GeoRef record URLs, and DOIs from multiple publisher prefixes (10.1190, 10.1144, etc.) (#109)

    • Add SEANOE (SEA scieNtific Open data Edition) as content provider for marine science data from Ifremer/SISMER, with metadata-only extraction from the REST API (geographic bounding boxes, temporal extents) and data download of open-access files; supports DOI prefix 10.17882 and seanoe.org landing page URLs (#104)

    • Add UKCEH (UK Centre for Ecology & Hydrology) EIDC as content provider with metadata-only extraction from the catalogue JSON API (bounding boxes, temporal extents) and dual data download pattern (Apache datastore listing or data-package ZIP); supports DOI prefix 10.5285 and catalogue.ceh.ac.uk URLs (#103)

    • Add GDI-DE (Geodateninfrastruktur Deutschland / geoportal.de) as metadata-only content provider for the national German spatial data infrastructure catalogue (771,000+ records), with CSW 2.0.2 metadata extraction via OWSLib; supports geoportal.de landing page URLs and bare UUIDs (#84)

    • Add generic CKAN content provider supporting any CKAN instance via dataset URL matching, with known-host fast matching and dynamic API discovery; includes GeoJSON spatial metadata parsing with geometry preservation for convex hull, multi-format temporal field support, and UK bbox-* extras pattern; known hosts include GeoKur TU Dresden, data.gov.uk, GovData.de, data.gov.au, and catalog.data.gov (#98)

    • Add STAC (SpatioTemporal Asset Catalog) as metadata-only content provider, extracting spatial bounding boxes and temporal intervals directly from STAC Collection JSON; supports known STAC API hosts (Element84, DLR, Terradue, WorldPop, Lantmateriet, etc.), /stac/ URL path patterns, JSON content-inspection fallback, and content negotiation for HTML/JSON servers (#25)

    • Add NFDI4Earth Knowledge Hub as metadata-only content provider for the Cordra-based digital object repository (1.3M+ datasets), with SPARQL endpoint extraction (spatial WKT, temporal ranges) and Cordra REST API fallback; supports OneStop4All landing page URLs and direct Cordra object URLs; follows landingPage to other supported providers (#100)

    • Add GitHub as content provider for extracting geospatial extent from public GitHub repositories, with GitHostProvider abstract base class for git hosting platforms; uses Git Trees API (2 API calls per repo) and raw file downloads; supports repository root, branch, and subdirectory URLs; preserves directory structure for co-located files (e.g. shapefile components); optional GITHUB_TOKEN for higher rate limits (#96)

    • Add GitLab as content provider for extracting geospatial extent from public GitLab repositories on gitlab.com and self-hosted instances; uses paginated Repository Tree API and raw file API; supports repository root, branch, and subdirectory URLs with nested namespace paths (group/subgroup/project); three-stage instance detection (known hosts, hostname heuristic, API probe); optional GITLAB_TOKEN for higher rate limits (#97)

    • Add Forgejo/Gitea as content provider for extracting geospatial extent from public Forgejo and Gitea repositories (including Codeberg.org and Helmholtz DataHub); uses Gitea REST API v1 for paginated tree listing and raw file downloads; three-stage instance detection (known hosts, hostname heuristic containing “forgejo” or “gitea”, /api/v1/version probe); optional FORGEJO_TOKEN for higher rate limits

    • Add Software Heritage as content provider for extracting geospatial extent from the universal software archive (Inria + UNESCO); supports persistent SWHIDs (swh:1:dir:..., swh:1:ori:..., swh:1:rev:..., swh:1:cnt:...), browse URLs (archive.softwareheritage.org), and SWHID qualifiers (origin=, path=); subpath optimization for targeted subdirectory extraction; optional SWH_TOKEN for higher rate limits (1200 req/hr vs 120 anonymous)

    • Add DataONE (Data Observation Network for Earth) as metadata-only content provider, extracting pre-computed spatial bounding boxes and temporal ranges from the Coordinating Node Solr API; covers ~20 member node repositories (KNB, PISCO, EDI/LTER, NEON, BCO-DMO, ESS-DIVE, etc.) through a single API; supports DOI prefixes 10.5063/ (KNB), 10.6085/ (PISCO), and search.dataone.org URLs; defers Arctic Data Center, PANGAEA, and Dryad datasets to their dedicated providers (#15)

    • Add journals/ umbrella package covering OJS (Open Journal Systems) and Janeway as metadata-only content providers for journal article landing pages. The shared extractor follows a richer-geometry-first priority (JSON-LD spatialCoverage is preferred over bbox-only encodings, with ICBM / geo.position as a points-only last resort) and an equivalent fallback chain for temporal extent. The existing Pensoft provider is folded into the same hierarchy with no external API change. Article DOIs are lifted from the landing page so a user can pass a journal article URL (not just a DOI) and still get CrossRef / DataCite enrichment via --external-metadata (#76)

  • New Features — Text source (spaCy NER, issue #112 )

    • New text source. spaCy-based named entity recognition extracts geospatial and temporal information from plain-text inputs. New optional [nlp] extra (pip install geoextent[nlp]) installs spaCy; the default en_core_web_sm model is auto-downloaded on first use (suppress with --no-auto-download). New handle_text handler module routes text files (mime text/*: .txt, .md, .markdown, .rst, .text) automatically; new public from_text() API processes in-memory strings. CLI: --text-method ("ner" is the default; "none" disables the text handler entirely) plus --text STRING, - (stdin), single text files, directories of text files, and mixed inputs (text + geospatial in one call) — all inputs in a multi-input run are merged into a single envelope / convex hull by default, matching the existing files-only behaviour. When the [nlp] extra is not installed, the handler silently declines so that pre-existing workflows that happen to include README.md keep working.

    • Place names from text. Detected LOC/GPE entities are forward-geocoded via the configured gazetteer (Nominatim, GeoNames, or Photon) and contribute to the bbox. Each detected name is recorded in a place_names provenance list carrying gazetteer_id and gazetteer_url. Areal hits use their administrative boundary when the gazetteer provides one (Nominatim only), so "Field campaign in Saxony" now yields the state’s polygon envelope instead of a centroid point; point-only gazetteers (GeoNames, Photon) fall back to the centroid. CLI: --place-geometry {auto,boundary,point} (default auto); --ner-model, --ner-labels (default LOC,GPE), --ner-score-threshold, --ner-gazetteer (default: Nominatim, no API key required), --ner-ambiguity {drop,top} (drop is default — ambiguous names like "Paris" or "Springfield" are skipped rather than guessed; use top to keep the highest-ranked candidate). Dropped names are reported via a one-shot WARNING on stderr.

    • Convex hull on mixed geometries. --convex-hull now consumes polygon boundaries when the gazetteer supplies them: a single polygon hit yields the polygon’s convex hull (acting as a polygon simplification for irregular outlines); a polygon hit plus an outside point extends the hull to enclose both; two or more point-only hits keep the line / polygon-of-points behaviour from before. --place-geometry point forces a centroid-only hull. See Extracting extents from free text for examples.

    • List the bundled period gazetteer. New --list-periods CLI flag (with --list-periods-format {json,text} and --list-periods-filter SUBSTR) and geoextent.lib.period_gazetteer.list_periods() API expose the full ICS GTS2020 chart that ships with geoextent. The output carries a versioned metadata block (schema_version, source, source_url, source_revision, source_revision_date, license, license_url, attribution, built_at, built_by, period_count) so downstream UIs can attribute the data and detect refreshes. Useful for autocomplete widgets and reference panels.

    • Calendar dates from text. DATE/TIME entities (parsed by dateutil) are joined with new structured parsing for ranges and granularities: year-only ("1987" → 1987-01-01..1987-12-31), decades ("the 1990s" → 1990-01-01..1999-12-31), centuries ("the 19th century" → 1801-01-01..1900-12-31), and range expressions captured by spaCy as a single span ("between 2010 and 2015", "from 1820 to 1850", "1820–1850", "January to March 2024"). Splitter handles ASCII hyphens, en-dashes, em-dashes, and to/until/through/and connectors.

    • Named time periods from text. Mentions like "Holocene", "Pleistocene", "Mesozoic Era", or "Late Cretaceous" are detected against a bundled ICS International Chronostratigraphic Chart (GTS2020, CC0; 178 geological eons, eras, periods, epochs, and ages) and resolved to a [start, end] pair via the bundled period gazetteer (new geoextent.lib.period_gazetteer). CLI: --period-gazetteer {bundled,none}, --period-ambiguity {drop,top}, --no-period-resolution.

    • Signed ISO 8601 extended year output for deep-time temporal extents. Geological mentions are emitted as signed ISO year strings (-9750-01-01 for the start of the Holocene, -251900050-01-01 for the base of the Mesozoic). Pre-CE and CE values merge correctly into a single tbox; CE-only output is unchanged and continues to honour --time-format.

    • Standoff offsets contract for downstream highlighting. Every place / date / period mention now carries char_start / char_end indices into the source string, and result dicts echo the (normalised) source text. --no-source-text / include_source_text=False opts out of the echo while keeping the offsets. See Highlighting matches from the text/NER source for the full contract.

    • In-terminal and pipe-friendly match highlighting. New --annotate {auto,ansi,brackets,off} CLI flag (and geoextent.lib.annotate.render_annotated_text() library helper) prints the source text with matched place names, dates, and named time periods highlighted after the JSON output. ansi uses terminal SGR colours (default: places cyan, dates yellow, periods magenta); brackets uses textual markers ([[Berlin|place]], [[Holocene|period]], [[2024-05-12|date]]) for log capture, pipelines, and chat clients. auto (default) picks ansi when stdout is a TTY else brackets. Override colours with --annotate-classes "place=red,date=green,period=blue".

    • Documentation. New how-to guide at Highlighting matches from the text/NER source walks through the standoff offset contract, source-text echo, --annotate modes, a Python consumer snippet, and a JavaScript/Java re-encoding recipe for callers that count UTF-16 code units. Examples in the README cover place-name extraction, range/decade/century parsing, geological periods, signed-ISO output, and --annotate brackets invocation. --formats / --list-features now lists “Text (NER)” with its capability metadata.

  • New Features

    • Add Cloud Optimized GeoTIFF (COG) support: direct HTTP(S) URLs to GeoTIFF files are opened via GDAL /vsicurl/ for efficient header-only metadata extraction without downloading the full file (#11)

    • Add point cloud support for LAS/LAZ files via laspy, extracting bounding boxes from file headers and temporal extent from creation dates (#10)

    • Add --join CLI flag and join_files() Python API to merge multiple export files (from --output) into a single file, dropping summary rows and keeping only individual-file features; supports cross-format joins (GPKG, GeoJSON, CSV)

    • Enhanced --output export: support single-file input, auto-detect GeoJSON/CSV format from extension, export_to_file() Python API, proper date fields (tbox_start/tbox_end), convex hull geometry export, --format interaction for CSV (#21)

    • Add metadata-only extraction for InvenioRDM instances (metadata.locations GeoJSON, metadata.dates)

    • Implement metadata-only extraction for Figshare provider (download_data=False / --no-data-download), supporting --metadata-first strategy (#68)

    • Expand Figshare provider to support institutional portal URLs (*.figshare.com, e.g. springernature.figshare.com, monash.figshare.com)

    • Add --time-format CLI option and time_format API parameter for configurable temporal extent output format: date-only (default), ISO 8601, or custom strftime strings (#39)

    • Add --metadata-first CLI flag and metadata_first API parameter for smart metadata-then-download extraction strategy: tries metadata-only extraction first if the provider supports it, falls back to data download if metadata yields no results

    • Add automatic metadata fallback: when data download yields no files and the provider supports metadata extraction, automatically fall back to metadata-only extraction (enabled by default, disable with --no-metadata-fallback or metadata_fallback=False)

    • DEIMS-SDR provider now follows external DOIs/URLs to other supported providers (e.g., Zenodo, PANGAEA) for actual data extent extraction. Disable with --no-follow or follow=False.

    • Extract temporal extent from raster files: NetCDF CF time dimensions, GeoTIFF TIFFTAG_DATETIME, ACDD time_coverage_start/end global attributes, and band-level ACQUISITIONDATETIME (IMAGERY domain) (#22)

    • Add --assume-wgs84 CLI flag and assume_wgs84 API parameter to explicitly enable WGS84 fallback for ungeoreferenced rasters (disabled by default)

    • Add support for Esri File Geodatabase (.gdb) format via GDAL’s OpenFileGDB driver

    • Add support for Zarr format (.zarr) via GDAL’s Zarr driver (V2 and V3), including directory-based dataset handling in CLI and directory extraction (#9)

    • All content providers now support interactive download size confirmation via --max-download-size. When the total download exceeds the limit, the CLI prompts for confirmation instead of silently truncating. API: download_size_soft_limit=True in from_remote().

    • Add -p / --parallel CLI flag and workers API parameter for parallel file extraction within directories using thread-based parallelism. -p auto-detects CPU count, -p N uses N workers. API: from_directory(..., workers=N) and from_remote(..., workers=N). (#34)

    • Add progress_callback parameter to from_file(), from_directory(), and from_remote() for structured progress reporting. Callbacks receive ProgressEvent dataclass instances with phase, message, current/total counters, and byte-level download progress. Three built-in callbacks: TqdmProgressCallback (tqdm bars), LoggingProgressCallback (logger), CollectingProgressCallback (list collection for testing). See API Docs for usage. (#80)

    • Add --map, --preview, and --map-dim CLI flags for static map preview of extracted spatial extents on OpenStreetMap tiles (requires pip install geoextent[preview]). --map saves a PNG to a temporary file, --map FILE saves to a specific path, --preview displays it inline in the terminal, --map-dim WxH sets image dimensions. The saved path is printed to stderr unless --quiet is used. (#35)

    • Add AppImage packaging for portable Linux distribution: single-file executable bundling Python, GDAL, PROJ, and all dependencies; CI workflow builds on tag push and attaches to GitHub Releases (#40)

  • Breaking Changes

    • Drop support for bare numeric Zenodo record IDs (e.g., 820562); use DOI (10.5281/zenodo.820562) or URL (https://zenodo.org/records/820562) instead

  • Bug Fixes

    • Skip raster files with pixel-based coordinates (outside WGS84 bounds) instead of merging them into geographic extents

    • Validate bounding boxes against WGS84 coordinate ranges before including in results

    • Reject vector files with projected coordinates falsely reported as WGS84 (e.g., GeoJSON files without CRS declaration)

    • --geojsonio (CLI) now distinguishes no spatial extent in output from geojson.io service call failed and surfaces the underlying reason — most commonly anonymous GitHub Gist uploads (the fallback used by geojsonio for large payloads) returning 401 Requires authentication. When the payload is large, the warning hints at --convex-hull as the user-actionable fix. Library callers can opt into a new GeojsonioUrlError exception via helpfunctions.generate_geojsonio_url(raise_on_error=True).

    • --placename default service changed from geonames (which requires the GEONAMES_USERNAME env var) to nominatim (no API key needed). Use --placename-service geonames to opt back into GeoNames.

    • --convex-hull now strips boundary polygons from the place_names provenance after consuming them for the hull, since the hull already encodes the spatial extent. This dramatically reduces output size for NER-driven extractions.

    • All placename / gazetteer warnings now name the service and endpoint (e.g. Nominatim reverse-geocoding via nominatim.openstreetmap.org failed for (52.3, 13.4): timeout).

    • --placename runs only once on the final merged extent now, eliminating duplicate No geometry provided for placename extraction warnings on multi-file inputs.

    • --place-geometry point now suppresses the boundary polygon in the place_names provenance — only the centroid lat/lon is kept.

    • Ambiguous-place-name warning now uses ; between candidates (e.g. Paris, Île-de-France, France; Paris, Lamar County, Texas).

    • Fix a crash when --convex-hull is combined with text files inside a directory walk.

    • --text (and - stdin) inputs are now merged with positional file inputs by default, matching the behaviour of multi-file runs.

    • Fix --convex-hull silently dropping single-point or two-point hulls (e.g. from --text "A to B") when merging them with peer geometries. Degenerate hulls now survive the merge.

    • Text-file detection now recognises extensionless project files (README, LICENSE, CHANGELOG, …) and additional text extensions (.log, .adoc, .org) as NER candidates. Source-code files (.py, .js, .sh, .yaml, …) are explicitly excluded so directory walks do not accidentally run NER on code.

  • API Changes

    • Rename all public API functions from camelCase to snake_case per PEP 8: from_file(), from_directory(), from_remote() (#31)

    • Remove old camelCase aliases (fromFile, fromDirectory, fromRemote)

    • Rename internal handler modules: handleCSV → handle_csv, handleRaster → handle_raster, handleVector → handle_vector

    • Rename all internal handler functions to snake_case (e.g., checkFileSupported → check_file_supported, getBoundingBox → get_bounding_box)

  • Improvements

    • Document installation with uv, conda/mamba, Poetry, and pipx (#4, #5, #41)

    • Fix and enable skipped multi-input CLI tests, add convex hull geometry tests for 2–5 file inputs, and add documentation for multiple input processing

    • Use GDAL CSV driver open options for coordinate column detection, supporting GDAL column naming conventions (X/Y, Easting/Northing) and CSVT sidecar files (#53)

    • Add GeoCSV format support: recognise CoordX/CoordY column names (giswiki.ch GeoCSV spec), .prj sidecar files for CRS information, WKT polygon geometry columns, and EarthScope GeoCSV #-prefixed metadata header lines (#52)

    • Move content provider metadata into provider classes (provider_info() classmethod), eliminating duplication in features.py

    • Verify bare UUIDs against BGR and Opara catalogs before accepting them.

    • Correct 4TU.ResearchData platform description: uses Djehuty, not Figshare

    • Automatically skip restricted files in Dataverse downloads

0.12.0

  • Breaking Changes

    • Default coordinate order for plain bounding boxes is now EPSG:4326 native axis order: (latitude, longitude)

    • Bounding boxes are returned as [minlat, minlon, maxlat, maxlon] instead of [minlon, minlat, maxlon, maxlat]

    • GeoJSON output always uses [longitude, latitude] coordinate order per RFC 7946

    • Add --legacy CLI flag and legacy=True Python API parameter to restore the previous [lon, lat] order for plain bounding boxes (does not affect GeoJSON output)

  • New Content Providers

    • Add Senckenberg Biodiversity and Climate Research Centre data portal provider (#66)

    • Support for CKAN-based repositories via new CKANProvider base class

    • Add BGR Geoportal content provider with ISO 19139 metadata extraction (#61)

  • New Features

    • Add support for world files (.wld, .jgw, .pgw, .tfw, etc.) (#36)

    • Add --keep-files option to preserve downloaded files for debugging

    • Dynamically generated supported formats list

    • Explicitly remove temp files on cleanup (#82)

  • Bug Fixes

    • Skip layers with degenerate extent [0,0,0,0] and emit a user-visible warning instead of silently including invalid coordinates

    • Fix resource leak for GeoPackage/SQLite-backed files (unclosed database connections)

  • Improvements

    • Enable parallel test execution by default using pytest-xdist (#38)

    • Refactor CRS extraction into shared utility function

    • Harden CSV handler: force CSV GDAL driver to prevent misidentification, add extension-based pre-filtering, improve geometry column detection

    • Optimize gazetteer queries to avoid duplicate API calls for closed polygon points

    • Suppress noisy pandas date-parsing warnings during temporal extent detection

0.11.0

  • Add --ext-metadata option to retrieve bibliographic metadata for DOIs from CrossRef and DataCite APIs

  • Add --ext-metadata-method option to control metadata source (auto, all, crossref, datacite)

  • Add display names to file format handlers

  • Fix unwanted coordinate flipping for GML bounding boxes with GDAL >= 3.2

0.10.0

  • fromRemote() now accepts both single identifiers (string) and multiple identifiers (list)

  • Add --list-features CLI option and get_supported_features() Python API for discovering supported file formats and content providers

  • Add validate_remote_identifier() and validate_file_format() validation functions

0.9.0

  • Content Providers

    • Add TU Dresden Opara content provider supporting DSpace 7.x repositories (#77)

    • Add GFZ Data Services as content provider (#17)

    • Add Dataverse repository support (#57)

    • Add support for OSF (Open Science Framework) (#19)

    • Add support for Dryad and Figshare repositories

    • Add support for Pensoft journals (#64)

    • Enhance PANGAEA provider with non-tabular data support and ZIP archive handling

    • Add download size limiting for repositories (#70)

    • Add --no-data-download option for metadata-only extraction

    • Add --download-skip-nogeo option to skip non-geospatial files

    • Add --max-download-size, --max-download-workers, and --max-download-method options

    • Restructure regex patterns for better repository candidate detection

  • Output and Visualization

    • Add extraction metadata to GeoJSON output (geoextent_extraction field)

    • Add --no-metadata option to exclude extraction metadata from GeoJSON output

    • Add geojson.io URL generation (--geojsonio)

    • Add --browse option to open visualizations in default web browser

    • Add WKT and WKB output format support (#46)

    • Add convex hull extraction (--convex-hull) (#37)

  • CLI and Processing

    • Add --no-subdirs option to control recursive directory processing (#55)

    • Add support for processing multiple files with automatic extent merging

    • Add progress bars with --no-progress option to disable (#32)

    • Add --quiet option to suppress all console messages

    • Add --placename option for geographic placename lookup via GeoNames, Nominatim, and Photon (#74)

    • Add file filtering and parallel downloads (#75)

    • Add FlatGeobuf format support (#43)

  • Infrastructure

    • Refactor CI workflows to use custom GDAL installation script instead of pygdal

    • Run code formatter (#54)

    • Skip GDAL auxiliary files during directory processing

  • Fix Figshare URL validation and download handling

0.8.0

  • Move configuration from setup.py to pyproject.toml

0.7.1

  • Add DOI-based retrieval functions for Zenodo (#100)

  • Add export function --output for folders, ZIP files and repositories (#124)

0.6.0

  • Add details option --details for folders and ZIP files (#116)

0.5.0

  • Add support for spatial extent for osgeo files (via OGR/GDAL) with generic vector (GeoPackage, Shapefile, GeoJSON, GML, GPX, KML) and raster handling (GeoTIFF) (#87, #99)

0.4.0

  • Add support for ZIP files and folders (#79)

0.3.0

  • Add debug option --debug and environment variable GEOEXTENT_DEBUG (#73)

  • Remove need for -input= for passing input paths in CLI (#73)

  • Switch to pygdal and update docs (#73)

0.2.0

0.1.0

  • Initial release with core functionality