Package 'taxifydb'

Description

Apply a binary delta to produce a new .vtr

Usage

apply_delta(old_path, delta_path, new_path)

Arguments

Value

The new path (invisibly), or NULL on failure.

Description

TAXIFYDB_PUBLISH_RESUME=1 reuses whatever .vtr sits in an enrichment's output directory instead of rebuilding it, which is what lets an interrupted publish finish cheaply. It also means a .vtr left from an earlier source release looks exactly like a current one: publishing it uploads stale data and writes a manifest entry naming a source that build never read. The meta.json sidecar records the URL and version the build actually used, so that is what the registry is compared against.

Usage

assert_built_matches_registry(name, vtr_path)

Arguments

Value

TRUE, invisibly. Stops when the build and the registry disagree.

Description

Marks df as carrying per-value provenance: its ⁠<col>_source⁠ columns hold reference ids that resolve against references. build_enrichment() carries the table through name resolution and build_enrichment_vtr() publishes it.

Usage

attach_references(df, references)

Arguments

Value

df, with the table in its references attribute.

Description

Build the AlgaeBase backbone .vtr

Usage

build_algaebase(
  output_dir = "output/algaebase",
  version = NULL,
  verbose = TRUE
)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Locates each backbone in the user's taxify data dir and writes a ⁠{backend}_name_lookup.vtr⁠ alongside it.

Usage

build_all_name_lookups(backends = list_backends(), overwrite = FALSE)

Arguments

Value

Character vector of paths to the built lookups.

Description

Build the AviList backbone .vtr from source

Usage

build_avilist(output_dir = "output/avilist", version = NULL, verbose = TRUE)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Single entry point for backbone builds. Downloads the raw source, normalizes to the unified schema, precomputes matching keys, and writes the final .vtr with indexes and metadata sidecar.

Usage

build_backend(name, output_dir = NULL, version = NULL, verbose = TRUE)

Arguments

Value

Path to the built .vtr file (invisibly).

Description

For each backbone in register_backbones() (resolved via resolve_register_backbone_paths()), extracts the genus list and writes a long-format backend_coverage.vtr recording which genera are covered by which backend and at what version/date.

Usage

build_backend_coverage(
  backbone_paths = NULL,
  output_dir = NULL,
  version = NULL,
  manifest_path = "manifest/manifest.json",
  verbose = TRUE
)

Arguments

Value

Path to backend_coverage.vtr (invisibly).

Description

Convenience wrapper over build_enrichment() restricted to the per-taxon assets rebuilt from published BETSI-derived matrices (see list_betsi_recovery()). The heavy lifting – staging the frozen matrices, the parse, cross-backbone name resolution, T-SITA + provenance metadata, the .vtr write – is the shared enrichment pipeline; this only guards the name.

Usage

build_betsi_recovery(name, ...)

Arguments

Value

Path to the built .vtr (invisibly).

See Also

list_betsi_recovery()

Description

Also writes the SpeciesProfile.tsv (extinct/marine flags) as a separate col_species_profile.vtr next to the main backbone, if present.

Usage

build_col(output_dir = "output/col", version = NULL, verbose = TRUE)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Build the COL XR backbone .vtr

Usage

build_colxr(
  output_dir = "output/colxr",
  version = NULL,
  gbif_path = NULL,
  col_path = NULL,
  manifest_path = "manifest/manifest.json",
  verbose = TRUE
)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Downloads the raw source for an enrichment, parses it via the registered parser, expands names across all 7 taxify backbones via resolve_enrichment_names(), and writes a .vtr plus meta.json sidecar via build_enrichment_vtr().

Usage

build_enrichment(
  name,
  output_dir = NULL,
  version = NULL,
  url = NULL,
  resolve_names = TRUE,
  strict_names = FALSE,
  verbose = TRUE
)

Arguments

Value

Path to the built .vtr file (invisibly).

Description

Sorts by canonical_name, writes the .vtr, creates hash indexes on canonical_name (and optionally a group column), and writes a meta.json sidecar.

Usage

build_enrichment_vtr(
  df,
  vtr_path,
  name,
  version,
  source_url,
  source_doi = NULL,
  source_record = NULL,
  upstream_id = NULL,
  license = "unknown",
  attribution = NULL,
  group_col = NULL,
  species_col = NULL,
  static = TRUE,
  source_format = NULL,
  provenance = NULL,
  references = NULL,
  resolved_backbones = NULL,
  batch_size = 50000L
)

Arguments

Value

The path to the .vtr file (invisibly).

Description

Build the Euro+Med backbone .vtr

Usage

build_euromed(
  output_dir = "output/euromed",
  version = NULL,
  local_dir = NULL,
  verbose = TRUE
)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Build the EuroVegChecklist syntaxon backbone .vtr

Usage

build_evc(
  output_dir = "output/evc",
  version = NULL,
  local_dir = NULL,
  verbose = TRUE
)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Build the FishBase backbone .vtr

Usage

build_fishbase(output_dir = "output/fishbase", version = NULL, verbose = TRUE)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Build the Species Fungorum Plus backbone .vtr

Usage

build_fungorum(output_dir = "output/fungorum", version = NULL, verbose = TRUE)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Build the GBIF backbone .vtr

Usage

build_gbif(output_dir = "output/gbif", version = NULL, verbose = TRUE)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Collects the keys gbif_count_keys() selects from a built GBIF backbone, counts their occurrence records, and writes them as a gzipped TSV for release under ⁠gbif-occurrence-counts-<version>⁠.

Usage

build_gbif_occurrence_counts(vtr_path, out_dir = "output/gbif", verbose = TRUE)

Arguments

Value

Path to the written .tsv.gz (invisibly).

Description

Reads genus-rank rows from each of register_backbones() (resolved via resolve_register_backbone_paths()), unions them, resolves classification conflicts, normalizes non-standard kingdom names, assigns kingdom_group / taxon_group / life_form (via assign_life_form()), and writes genus_register.vtr.

Usage

build_genus_register(
  backbone_paths = NULL,
  output_dir = NULL,
  version = NULL,
  manifest_path = "manifest/manifest.json",
  verbose = TRUE
)

Arguments

Details

Every call unions the same fixed backbone set, so the built register is reproducible: it does not depend on which backbones happen to be installed on the machine running the build (contrast taxify's runtime taxify_build_register(), which unions whichever backbones the caller has installed).

Value

Path to genus_register.vtr (invisibly).

Description

Downloads the ITIS SQLite dump, normalizes it, precomputes matching keys, embeds synonym info, and writes the final .vtr with indexes.

Usage

build_itis(output_dir = "output/itis", version = NULL, verbose = TRUE)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Build the LCVP backbone .vtr from source

Usage

build_lcvp(output_dir = "output/lcvp", version = NULL, verbose = TRUE)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Build the LPSN backbone .vtr from source

Usage

build_lpsn(output_dir = "output/lpsn", version = NULL, verbose = TRUE)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Build the Mammal Diversity Database backbone .vtr from source

Usage

build_mdd(output_dir = "output/mdd", version = NULL, verbose = TRUE)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Downloads the MEOW ecoregion GeoJSON snapshot (or reads a local copy), parses it to a long-format vertex table, and writes meow.vtr with an index on code and a metadata sidecar. taxify downloads this pre-built .vtr and rebuilds the boundary geometry offline, so the runtime never fetches the GeoJSON. The marine analogue of build_wgsrpd().

Usage

build_meow(
  output_dir = NULL,
  version = NULL,
  source_path = NULL,
  verbose = TRUE
)

Arguments

Value

Path to the built .vtr (invisibly).

Description

Build a name-lookup .vtr from a backbone .vtr

Usage

build_name_lookup(bb_path, out_path, verbose = TRUE)

Arguments

Value

out_path (invisibly).

Description

Build the NCBI backbone .vtr

Usage

build_ncbi(output_dir = "output/ncbi", version = NULL, verbose = TRUE)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Build the OTT backbone .vtr

Usage

build_ott(output_dir = "output/ott", version = NULL, verbose = TRUE)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Convenience wrapper that calls build_genus_register() and build_backend_coverage() in sequence. Resolves every backbone's .vtr exactly once (via resolve_register_backbone_paths(), downloading into ⁠output/genus_register/_cache⁠ when needed) and passes the fully resolved paths to both builders as backbone_paths, so a manifest-downloaded backbone (several hundred MB each, ~5 GB across the set) is fetched once for both artifacts rather than once per builder.

Usage

build_register(
  backbone_paths = NULL,
  version = NULL,
  manifest_path = "manifest/manifest.json",
  verbose = TRUE
)

Arguments

Value

Named list with paths to genus_register.vtr and backend_coverage.vtr (invisibly).

Description

Build the Reptile Database backbone .vtr

Usage

build_reptiledb(
  output_dir = "output/reptiledb",
  version = NULL,
  verbose = TRUE
)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Build the SeaLifeBase backbone .vtr

Usage

build_sealifebase(
  output_dir = "output/sealifebase",
  version = NULL,
  verbose = TRUE
)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Build the Czech vegetation classification syntaxon backbone .vtr

Usage

build_vegcz(
  output_dir = "output/vegcz",
  version = NULL,
  local_dir = NULL,
  verbose = TRUE
)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Sorts by genus (for zone-map pruning), writes the .vtr, creates hash indexes, and writes a metadata sidecar.

Usage

build_vtr(
  df,
  vtr_path,
  backend_name,
  version,
  source_url,
  source_date,
  batch_size = 50000L,
  block_col = NULL,
  extra = NULL
)

Arguments

Value

The path to the .vtr file (invisibly).

Description

The streaming counterpart to build_vtr(), for a source too large to assemble in memory. Rows arrive a chunk at a time from feed; the finished store is identical to what build_vtr() would have written from the concatenation of those chunks.

Usage

build_vtr_streamed(
  feed,
  vtr_path,
  backend_name,
  version,
  source_url,
  source_date,
  synonym_pattern = "SYNONYM",
  meta_extra = NULL,
  batch_size = 50000L,
  work_dir = NULL,
  verbose = TRUE
)

Arguments

Value

The path to the .vtr file (invisibly).

Description

Build the WCVP backbone .vtr from source

Usage

build_wcvp(output_dir = "output/wcvp", version = NULL, verbose = TRUE)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Build the WFO backbone .vtr from source

Usage

build_wfo(output_dir = "output/wfo", version = NULL, verbose = TRUE)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Downloads the WGSRPD Level 3 GeoJSON (or reads a local copy), parses it to a long-format vertex table, and writes wgsrpd.vtr with an index on code and a metadata sidecar. taxify downloads this pre-built .vtr and rebuilds the boundary geometry offline, so the runtime never fetches the GeoJSON.

Usage

build_wgsrpd(
  output_dir = NULL,
  version = NULL,
  source_path = NULL,
  verbose = TRUE
)

Arguments

Value

Path to the built .vtr (invisibly).

Description

Also writes the SpeciesProfile.tsv (habitat flags) as a separate worms_species_profile.vtr next to the main backbone.

Usage

build_worms(output_dir = "output/worms", version = NULL, verbose = TRUE)

Arguments

Value

Path to the .vtr file (invisibly).

Description

Check all non-static enrichments in a manifest for version freshness

Usage

check_all_enrichment_versions(manifest_path = "manifest/manifest.json")

Arguments

Value

Data.frame with columns: name, source_version, built_id, upstream_id, upstream_version, outdated, check_url, note.

Description

The resource is named by ⁠/resource/<id>⁠ in the URL, else in the source's host record (a portal whose downloads go through its datastore API).

Usage

check_ckan_version(source_url, record = NULL)

Arguments

Value

Named list with id (the resource's last_modified stamp), version, date, and url; NULL if the resource cannot be read.

Description

For a host that serves a generated export with no version and no Last-Modified header (a PHP endpoint, a database export). The bytes are the only identity there is, so the probe downloads them; it is listed only for sources small enough to fetch weekly, and only where two downloads hash alike.

Usage

check_content_md5(source_url, record = NULL)

Arguments

Value

Named list with id (the md5), version (its first 8 characters), and url; NULL if the file cannot be read.

Description

A published article does not get new versions, it gets corrections and retractions, which Crossref records against it as updated-by. A source whose values come from an article is therefore checked by that list.

Usage

check_crossref_version(source_url, record = NULL)

Arguments

Value

Named list with id (the updating DOIs, or "none"), version, date (the latest of publication and update), and url; NULL if Crossref does not know the DOI.

Description

For a host whose own pages give no version but whose DataCite record carries one. The record's updated stamp moves with any metadata edit, so as a date it errs late, never early.

Usage

check_datacite_version(source_url, record = NULL)

Arguments

Value

Named list with id and version (DataCite's version), date, and url; NULL if the record has no version.

Description

DataONE system metadata links each object to the one that replaced it (obsoletedBy), so following the chain from the object the URL names reaches the newest version.

Usage

check_dataone_version(source_url, record = NULL)

Arguments

Value

Named list with id (the newest identifier), version, date, pinned (the identifier the URL names), and url; NULL if the system metadata cannot be read.

Description

The URL either is a Dataverse API call naming the dataset by persistentId=doi:... or is the dataset DOI itself, in which case the installation serving it is the host DataCite resolves the DOI to.

Usage

check_dataverse_version(source_url, record = NULL)

Arguments

Value

Named list with id and version (major.minor), date, and url; NULL if the versions cannot be read.

Description

Check a Dryad dataset for the latest version

Usage

check_dryad_version(source_url, record = NULL)

Arguments

Value

Named list with id and version (both the newest version number), date, and url; NULL if the dataset cannot be read.

Description

An EDI (pasta.lternet.edu) data URL names a package revision (⁠<scope>/<identifier>/<revision>⁠) and usually one data entity in it. An entity keeps its id across revisions, so an entity URL is compared by the entity's checksum in the revision it names and in the newest one: a revision that changes only metadata or other tables leaves it current. A URL naming the revision alone is compared by revision number.

Usage

check_edi_version(source_url, record = NULL)

Arguments

Details

PASTA has required an API key for every call since 2026-07-30, read here from EDI_API_KEY as download_edi_file() reads it.

Value

Named list with id and pinned (entity checksums, or revision numbers), version (the newest revision), date (its publication date), and url; NULL without a key or when PASTA cannot be read.

Description

Freshness is decided by comparing the upstream identity the build read against the one upstream carries now. Where the source URL pins a version, that pinned identity is what the build read, whatever was recorded. Where it floats, the build's identity comes from upstream_id, written into the entry when the .vtr was built.

Usage

check_enrichment_source_version(entry, probe = probe_upstream_identity)

Arguments

Details

A floating source with no recorded identity is still answerable when the newest upstream version is older than the build: the entry's release tag (latest, YYYY.MM) bounds the build date from below, and a build reads the newest version there is, so the build read that version.

Where none of these holds the answer is NA, not a guess: the recorded source_version is this package's own release string and the upstream version is whatever the host counts in, so comparing the two answers a different question than the one asked.

Value

Named list with source_version, built_id, upstream_id, upstream_version, outdated, check_url, and optionally note.

Description

An article URL is compared by the article's version number. A file download URL (⁠ndownloader.figshare.com/files/<id>⁠) names one immutable file, so it is compared by content: the file's md5 is the pinned identity, and the newest version's file of the same name is the upstream one (⁠missing:<name>⁠ when the newest version dropped it).

Usage

check_figshare_version(source_url, record = NULL)

Arguments

Details

Nothing in a file URL names its article, so the article comes from the source's host record: a Figshare DOI (⁠10.6084/m9.figshare.<id>⁠), a collection DOI (⁠10.6084/m9.figshare.c.<id>⁠, whose articles are searched for the file), or a Figshare article URL.

Value

Named list with id, version, date, pinned (file URLs only), and url; NULL if the article cannot be read.

Description

Check the GBIF backbone API for the latest update date

Usage

check_gbif_api_version(source_url, record = NULL)

Arguments

Value

Named list with id (the dataset's modified timestamp), version (that timestamp as YYYY.MM), date, and url; NULL if the dataset cannot be read.

Description

The GIFT API lists every database version; each description opens with the date the version was released.

Usage

check_gift_version(source_url, record = NULL)

Arguments

Value

Named list with id and version (the newest version), date, and url; NULL if the versions cannot be read.

Description

A raw.githubusercontent.com URL serves the file at the head of a branch, so it pins nothing. The identity is the newest commit touching that path.

Usage

check_github_version(source_url, record = NULL)

Arguments

Value

Named list with id (the commit SHA), version (its short form), date, and url; NULL if the commit cannot be read.

Description

An IPT archive URL names its version (⁠v=⁠); the same URL without it serves the newest, under a file name carrying that version.

Usage

check_ipt_version(source_url, record = NULL)

Arguments

Value

Named list with id and version (the newest version), date, pinned (the version the URL names), and url; NULL if the archive cannot be read.

Description

Check the Kew Plant DNA C-values database for its newest release

Usage

check_kew_cvalues_version(source_url, record = NULL)

Arguments

Value

Named list with id and version (the release number), date (the release month), and url; NULL if the release history cannot be read.

Description

For a host that serves a static file and says when it last changed. The header stamp is kept whole as the identity and shortened to YYYY.MM only for display: two files a fortnight apart share a month, and an identity that cannot tell them apart reports a moved file as unchanged.

Usage

check_last_modified(source_url, record = NULL)

Arguments

Value

Named list with id (the Last-Modified stamp), version (that stamp as YYYY.MM), date, and url; NULL if the header is absent.

Description

LCVP is distributed as an R data package; its data version is the Version field of the package DESCRIPTION on GitHub.

Usage

check_lcvp_version(source_url)

Arguments

Value

Named list with version and url.

Description

NCBI rewrites its rolling new_taxdump daily, and keeps one dump a month in ⁠taxdump_archive/⁠ as ⁠new_taxdump_<YYYY-MM-DD>.zip⁠. A source pinned to an archived dump compares its date against the newest one there.

Usage

check_ncbi_archive_version(source_url, record = NULL)

Arguments

Value

Named list with id, version and date (the newest dump's date), pinned (the date the URL names), and url; NULL if the archive listing cannot be read.

Description

Check the Open Tree Taxonomy for its newest release

Usage

check_ott_version(source_url, record = NULL)

Arguments

Value

Named list with id and version (the newest release), pinned (the release the URL names), and url; NULL if the listing cannot be read.

Description

A PANGAEA DOI names one dataset, which is replaced by a new DOI rather than edited; the metadata record states the data's last modification, its status, and any DataCite relation to a successor.

Usage

check_pangaea_version(source_url, record = NULL)

Arguments

Value

Named list with id (last modification, status and any successor), version, date, and url; NULL if the record cannot be read.

Description

Phaidra (University of Vienna) objects are immutable; a new version is a new object listed under the old one's versions. The portal sits behind an Anubis proof-of-work page, which .anubis_fetch() clears.

Usage

check_phaidra_version(source_url, record = NULL)

Arguments

Value

Named list with id (the newest object), version, date, pinned (the object the URL names), and url; NULL if the object cannot be read.

Description

rfishbase reads FishBase and SeaLifeBase snapshots published as release folders of one Hugging Face dataset, and takes the newest by default.

Usage

check_rfishbase_version(source_url, record = NULL)

Arguments

Value

Named list with id and version (the release label), date (the dataset's last change, an upper bound on the release date), and url; NULL if the releases cannot be read.

Description

SEANOE updates a dataset in place under one DOI, adding the new files beside the old ones. The set of file names is therefore the dataset's identity: a new version shows up as a file the set did not have.

Usage

check_seanoe_version(source_url, record = NULL)

Arguments

Value

Named list with id (the sorted file names), version (the file count), and url; NULL if the dataset cannot be read.

Description

A snapshot URL (⁠web.archive.org/web/<timestamp>id_/<original>⁠) fixes one capture of a page that may since have disappeared. Each capture carries a digest of the bytes it holds, so the capture the URL names and the newest capture of the same original compare by content: a re-capture of an unchanged file is not a new version.

Usage

check_wayback_version(source_url, record = NULL)

Arguments

Value

Named list with id (the newest capture's digest), version (its timestamp), date, pinned (the digest of the capture the URL names), and url; NULL if the capture index cannot be read.

Description

The database grows by datasets, each uploaded with a date and a record count, so their number, their records and the newest upload together say whether the export would read differently.

Usage

check_wst_version(source_url, record = NULL)

Arguments

Value

Named list with id, version, date (the newest upload), and url; NULL if the dataset list cannot be read.

Description

A Zenodo record is immutable, so querying the pinned record reports the day that record was published however many newer versions its concept has since gained. The versions/latest endpoint resolves the pinned record to the newest one in the same concept. A URL naming the concept record itself floats: Zenodo serves its newest version, so it pins nothing.

Usage

check_zenodo_version(source_url, record = NULL)

Arguments

Details

Where the URL names one file of the record, the identity is that file's checksum, read from the pinned and the newest record alike: a new version that only adds a metadata file leaves the data the build reads unchanged. A file the newest record no longer carries reads as ⁠missing:<key>⁠.

Value

Named list with id, version, date, pinned (absent for a concept record), and url; NULL if the record cannot be read.

Description

COL XR keeps the Base Release's identifiers, so a usage's publication reference can be read from the COL build by taxon_id. Only rows carrying a reference are kept.

Usage

colxr_col_publications(col_path)

Arguments

Value

A data.frame of taxon_id, canonical_name, authorship, name_published_in and year.

Description

COL XR replaced the taxonomy behind GBIF.org, but GBIF serves occurrence records only for the numeric keys of its legacy backbone. This reads the gbif backbone and groups its usages by canonical name, rank and authorship (see crosswalk_authorship()). GBIF files the records of a synonym under its accepted taxon, so a synonym usage contributes its accepted taxon's key. A name several GBIF usages share (a homonym, or one name GBIF split itself) maps to the set of their keys, |-delimited in ascending order.

Usage

colxr_gbif_crosswalk(gbif_path)

Arguments

Value

A colxr_gbif_crosswalk: the key sets grouped by full authorship and by outer authorship, for colxr_gbif_lookup().

Description

In order: the usage's full authorship against GBIF's, its outer authorship against GBIF's outer authorship, and, where GBIF records no authorship for the name, the name and rank alone. A GBIF usage that names a different author is a different name and is never matched.

Usage

colxr_gbif_lookup(crosswalk, canonical_name, authorship, taxon_rank)

Arguments

Value

Character vector of |-delimited GBIF keys, NA where none.

Description

Queries ChecklistBank for the most recently issued COL XR dataset whose Darwin Core export can be downloaded. Each monthly release carries its own dataset key, so the key is looked up rather than hard-coded. ChecklistBank generates an export only when a logged-in account asks for one, so when the newest release has none, one is requested with colxr_request_export(). When that cannot run (no GBIF_USER / GBIF_PWD) or does not finish, the newest release with an export is built instead.

Usage

colxr_latest_release(
  verbose = TRUE,
  export_ready = colxr_export_ready,
  request_export = function(key) {
     colxr_request_export(key, verbose = verbose)
 }
)

Arguments

Value

A list with key, alias, version and issued.

Description

ChecklistBank generates a dataset's export only when a logged-in GBIF account asks for one, so a release can be issued and never become downloadable. This starts a DwCA export job with the account in GBIF_USER / GBIF_PWD and polls it until it finishes.

Usage

colxr_request_export(key, poll = 30, timeout = 3600, verbose = TRUE)

Arguments

Value

TRUE when the export finished, FALSE when no account is configured or the job did not finish.

Description

Count GBIF occurrence records per taxon key

Usage

count_gbif_occurrences(
  keys,
  batch_size = 200L,
  facet_limit = 50000L,
  verbose = TRUE
)

Arguments

Value

A data.frame with taxon_id and n_occurrences.

Description

Count rows in a .vtr file

Usage

count_vtr_rows(vtr_path)

Arguments

Value

Integer.

Description

Besides the patch, writes ⁠<delta_path>.base⁠: one line holding the content id (md5) of old_path. A patch applies only to the exact bytes it was cut against, and a release tag does not identify those bytes (a re-cut reuses it), so the manifest records this id as delta_from_content_id and the runtime patches only a local build carrying it.

Usage

create_delta(old_path, new_path, delta_path)

Arguments

Value

The delta path (invisibly), or NULL if xdelta3 is unavailable.

Feed a delimited file to build_vtr_streamed() a chunk at a time

Description

Reads a header once, then hands back successive blocks of rows, so a file larger than memory can be normalized and staged block by block.

Usage

delim_chunk_feed(
  path,
  normalize,
  chunk_rows = 500000L,
  sep = "\t",
  quote = "",
  encoding = "UTF-8",
  select = NULL,
  na_strings = c("", "NA"),
  col_names = NULL,
  file_encoding = NULL,
  verbose = TRUE
)

Arguments

Details

quote, encoding and file_encoding must be set to whatever the backbone's whole-file reader uses. They decide how the text is parsed, not just how fast it is read: WoRMS ships its TSV with every field wrapped in double quotes, so reading it with quoting disabled leaves the quote characters inside the names and breaks every downstream join, while a source carrying genuine embedded quotes in informal names needs quoting disabled to keep them.

quote also decides how the file is cut. With quoting disabled a newline always ends a record, so blocks are taken by row offset and data.table::fread() reads each directly, which is the faster path and the one the unquoted sources use.

With quoting enabled that stops holding: a field may contain a newline, and then a line offset is not a record offset. WoRMS carries 4,626 newlines inside quoted namePublishedIn fields, and every reader's row argument counts lines – fread()'s skip does, and so does read.delim(nrows=) through scan()'s nlines – so a line-offset read is 445 records adrift by record 200,000 and silently repeats rows. A quoted source is therefore cut only where the running count of quote characters is even, which is the one place a record can end, and the number of records that implies is checked against the number the parser returns.

Lines are taken at the byte level, split on LF alone. R's own line reader recognizes a lone CR as a terminator as well, and WoRMS holds 59 of them inside fields: read as text it manufactures 57 lines the file does not have and drops 49 quote characters, which flips the parity and is why utils::read.delim() gives up on that file partway through.

A field may also contain "", which RFC 4180 defines as one literal quote. utils::read.delim() collapses it and fread() returns it doubled (Rdatatable/data.table#1109); WFO has 4,117 such fields, and leaving them doubled would put stray quote characters in the published values, which is the defect worms-2026.07 was cut to fix. Rather than let the choice of reader decide that, the blocks are parsed with fread() and the unescaping is done here, so it holds whichever source is being read.

Value

A function of no arguments suitable as the feed of build_vtr_streamed().

Description

A backbone that denormalizes its higher classification stores foreign keys rather than names, so normalizing a row needs a name that lives in some other row. GBIF is the case: kingdom, phylum, class, order and family all arrive as ⁠*_key⁠ columns pointing at rows that the build then drops.

Usage

delim_fk_lookup(
  path,
  id_col,
  value_col,
  key_cols,
  chunk_rows = 1000000L,
  sep = "\t",
  quote = "",
  encoding = "UTF-8",
  na_strings = c("", "NA"),
  col_names = NULL,
  verbose = TRUE
)

Arguments

Details

The whole id -> name map would be as large as the backbone, but only the ids actually referenced are ever looked up, and those are the few thousand rows at the ranks the keys point to. Two scans find them without holding the map: the first collects the distinct keys, the second keeps the id/value pair of each row a key names. What comes back is small enough to sit beside a chunk, which makes the rest of the normalization row-local.

The result is exactly the lookup a whole-table self-join would produce; it is restricted by which ids are referenced, never by an assumption about the rank a key points at.

Value

A named character vector of value, named by id.

Description

Fetch all AlgaeBase records via /nameusage/search

Usage

download_algaebase(verbose = TRUE)

Arguments

Value

A list of record objects (raw JSON list-of-lists).

Description

Downloads url to dest_dir/source.zip (if not cached), extracts into ⁠dest_dir/extracted/⁠, and returns the first file matching pattern. If pattern is NULL, returns the extraction directory itself.

Usage

download_and_unzip(url, dest_dir, pattern = NULL)

Arguments

Value

Path to the matched file, or the extraction directory if pattern is NULL.

Description

Download the AviList extended checklist workbook

Usage

download_avilist(dest = tempdir(), verbose = TRUE)

Arguments

Value

Path to the downloaded .xlsx.

Description

Downloads url into dest_dir/filename with a browser TLS impersonation that passes Cloudflare's "Just a moment" challenge. A host running a JavaScript challenge, which no impersonation can satisfy, is instead cleared once in a visible Chrome window and the resulting cookies reused for 12h; that rung needs the Python package nodriver. Cached: skips if the destination exists and is non-empty.

Usage

download_cf_file(url, dest_dir, filename)

Arguments

Value

Path to the downloaded file.

Description

Download and extract the COL DwC-A

Usage

download_col(dest = tempdir(), verbose = TRUE)

Arguments

Value

Path to the extracted COL directory.

Description

Download and extract the COL XR Darwin Core Archive

Usage

download_colxr(dest = tempdir(), key = NULL, verbose = TRUE)

Arguments

Value

Path to the extracted TSV.

Description

Downloads url into dest_dir/filename if the destination does not already exist (or is empty). Follows redirects, sets a User-Agent header, and optionally a Referer.

Usage

download_curl_file(
  url,
  dest_dir,
  filename,
  referer = NULL,
  user_agent = NULL,
  extra_headers = NULL,
  max_tries = 4L
)

Arguments

Value

Path to the downloaded file.

Description

Resolves the newest version of a Dryad dataset DOI to a data file matching file_pattern, then downloads it through .anubis_fetch(). Cached: skips the download if dest_dir/filename already exists and is non-empty.

Usage

download_dryad_file(doi, dest_dir, filename, file_pattern = "\\.csv$")

Arguments

Value

Path to the downloaded file.

Description

EDI has required authentication for all REST API access since 2026-07-30 (an anti-DoS/scraping measure; see edirepository.org/news/news-20260727.00). Anonymous requests to pasta.lternet.edu now get a 403 ("Public Access is not authorized"). This wraps download_curl_file() with EDI's documented auth: an API Access Key appended as a ⁠?key=⁠ query parameter (the supported mechanism for "coded workflows, automated scripts, and other integrations", per https://edirepository.org/resources/iam), read from the EDI_API_KEY env var. EDI_EDI_TOKEN (a raw edi-token session JWT, from a logged-in browser) is an interactive fallback – it expires in days, so it is not suitable for a scheduled/CI rebuild.

Usage

download_edi_file(url, dest_dir, filename, max_tries = 4L)

Arguments

Value

Path to the downloaded file.

Description

Fetches the crawler's NDJSON snapshot and node-path table from the release. If a local crawl directory already holds both files (the machine that ran the crawl), they are used directly.

Usage

download_euromed(dest = tempdir(), local_dir = NULL, verbose = TRUE)

Arguments

Value

Named list with jsonl and nodes paths.

Description

Fetches the unit list and the harvested synonymy from the ⁠evc-snapshot-<version>⁠ release. A local crawl directory holding both files (the machine that ran the crawl) is used directly.

Usage

download_evc(dest = tempdir(), local_dir = NULL, verbose = TRUE)

Arguments

Value

Named list with synonyms, units and synonyms_2016 paths.

Description

Download and extract the Species Fungorum Plus ColDP archive

Usage

download_fungorum(dest = tempdir(), verbose = TRUE)

Arguments

Value

Path to the directory containing the extracted ColDP files.

Description

Download the GBIF backbone simple.txt.gz

Usage

download_gbif(dest = tempdir(), verbose = TRUE)

Arguments

Value

Path to the .gz file.

Description

Pages through a GBIF API endpoint that uses offset + limit and an endOfRecords flag. Returns a combined data.frame of ⁠$results⁠ payloads.

Usage

download_gbif_api_pages(base_url, params, limit = 1000L, max_pages = 100L)

Arguments

Value

A data.frame of combined results. Empty data.frame if the endpoint returned no rows.

Description

Download the published GBIF occurrence-count snapshot

Usage

download_gbif_occurrence_counts(dest = tempdir(), verbose = TRUE)

Arguments

Value

Path to the downloaded .tsv.gz.

Description

The aggregate drives every trait; the full version is needed to repair specific_root_area (see parse_groot()). The full download is best-effort: if it fails, the aggregate SRA is used as-is.

Usage

download_groot(url, dest)

Arguments

Value

Named character vector ⁠c(aggregate = <csv>, full = <csv or NA>)⁠.

Description

Download and extract the ITIS SQLite database

Usage

download_itis(dest = tempdir(), verbose = TRUE)

Arguments

Value

Path to the extracted SQLite database file.

Description

Download the LCVP tab_lcvp data file

Usage

download_lcvp(dest = tempdir(), verbose = TRUE)

Arguments

Value

Path to the downloaded tab_lcvp.rda.

Description

Download the LPSN ColDP archive from ChecklistBank

Usage

download_lpsn(dest = tempdir(), verbose = TRUE)

Arguments

Value

Path to the directory holding the extracted NameUsage.tsv.

Description

Download and unpack the Mammal Diversity Database archive

Usage

download_mdd(dest = tempdir(), verbose = TRUE)

Arguments

Value

Path to the directory holding the extracted CSVs.

Description

Download the MEOW ecoregion GeoJSON snapshot

Usage

download_meow(dest = tempfile(fileext = ".geojson"), verbose = TRUE)

Arguments

Value

The path to the downloaded GeoJSON (invisibly).

Description

Download and extract NCBI taxdump

Usage

download_ncbi(dest = tempdir(), verbose = TRUE)

Arguments

Value

Path to the extraction directory.

Description

Download and extract OTT taxonomy

Usage

download_ott(dest = tempdir(), verbose = TRUE)

Arguments

Value

Path to the extraction directory containing taxonomy.tsv.

Description

Download the Reptile Database bulk files

Usage

download_reptiledb(dest = tempdir(), verbose = TRUE)

Arguments

Value

Named list with paths: taxa, synonyms, checklist.

Description

Download the frozen Pladias vegetation snapshot

Usage

download_vegcz(dest = tempdir(), local_dir = NULL, verbose = TRUE)

Arguments

Value

Named list with synonyms and units paths.

Description

Download and extract the WCVP names file

Usage

download_wcvp(dest = tempdir(), verbose = TRUE)

Arguments

Value

Path to the extracted wcvp_names.csv.

Description

Download and extract the WFO classification file

Usage

download_wfo(dest = tempdir(), verbose = TRUE, url = NULL)

Arguments

Value

Path to the extracted classification file.

Description

Download the WGSRPD Level 3 GeoJSON

Usage

download_wgsrpd(dest = tempfile(fileext = ".geojson"), verbose = TRUE)

Arguments

Value

The path to the downloaded GeoJSON (invisibly).

Description

Wiley's supplement endpoint (⁠/action/downloadSupplement⁠) returns 403 to a cold request: it requires session cookies set by first loading the article page and an article Referer. The bundled cf_fetch.py wiley mode primes the session and sends the correct headers through a browser TLS impersonation, clearing both the Cloudflare tier and the cookie/Referer check. Cached: skips if the destination exists and is non-empty.

Usage

download_wiley_supplement(article_url, doi, sup_file, dest_dir, filename)

Arguments

Value

Path to the downloaded file.

Description

Download and extract WoRMS DwC-A

Usage

download_worms(dest = tempdir(), verbose = TRUE)

Arguments

Value

Path to the extracted DwC-A directory.

Description

Same as build_enrichment() but returns the parsed data.frame instead of writing a .vtr file. Useful when the runtime needs the raw data in memory (e.g. emergency fallback wired into taxify).

Usage

enrichment_emergency_fallback(name, verbose = TRUE)

Arguments

Value

data.frame with canonical_name + trait columns.

Description

The columns taxify's runtime exposes from an enrichment .vtr: every column except the join keys (canonical_name, accepted_name, genus) and the group column, which is surfaced through available_groups and the door's group argument rather than as a trait. This is what a manifest entry's trait_cols records.

Usage

enrichment_trait_cols(columns, group_col = NULL)

Arguments

Value

Character vector of trait column names, in .vtr order.

Description

Taxon keys whose occurrence count can decide a GBIF pick

Usage

gbif_count_keys(vtr_path)

Arguments

Value

Character vector of taxon keys: every record sharing a matching key with a record of another accepted taxon, plus the accepted targets of those records.

Description

The BETSI-derived code-keyed matrices (INRAE UU2FQT / UCYSLH, Bonfanti 2022) encode a species as the first three letters of the genus and the first three of the specific epithet, uppercased and joined by an underscore: "Brachystomella parvula" becomes "BRA_PAR". Hyphens in the epithet are dropped before the first three letters are taken. Returns NA for anything that is not a capitalised ⁠Genus species⁠ binomial.

Usage

gen_spe(binomial)

Arguments

Value

Character vector of GEN_SPE codes, NA where the input is not a capitalised ⁠Genus species⁠ binomial.

Examples

gen_spe(c("Brachystomella parvula", "Allacma gallica", "Onychiurus"))

Description

A chunked reader seeks by row offset, and every seek into a .gz restarts the decompression, so a gzipped source is expanded once up front instead.

Usage

gunzip_file(gz_path, out_path, verbose = TRUE)

Arguments

Value

out_path, invisibly.

Description

CKAN's elasticsearch-backed datastore caps offset at max_result_window (10000); the helper pages the full table via the after (search_after) cursor, through Cloudflare, writing one JSON record per line. Cached.

Usage

harvest_ckan_datastore(api_base, resource_id, dest_dir, filename)

Arguments

Value

Path to the downloaded JSONL file.

Description

Plazi republishes each marked-up paper as its own GBIF checklist dataset, so a taxon's treatments are spread over hundreds of small archives rather than one bulk file. Discovery is the union of two GBIF dataset-search filters because neither is complete alone: taxonKey depends on GBIF having indexed the dataset's taxonomic coverage, and the free-text query depends on the paper's title or abstract naming the taxon. For class Collembola the two return 210 and 489 datasets and their union is 508.

Usage

harvest_plazi_dwca(taxon_key, dest_dir, query = NULL, max_datasets = 20000L)

Arguments

Details

Each dataset is resolved to its DWC_ARCHIVE endpoint and cached as ⁠<dataset key>.zip⁠, so a re-run only fetches what is new.

Value

Path to the directory holding the cached archives.

Description

Check if xdelta3 is available on PATH

Usage

has_xdelta3()

Value

Logical.

Description

The verified TOKEN -> genus mapping for the bespoke genus abbreviations in the INRAE UU2FQT / UCYSLH code-keyed matrices, read from inst/extdata/betsi/inrae_genus_abbr.csv. Each token is confirmed by a named Collembola species in COLTRAIT (Salmon et al. 2014) or the built reference pool; unconfirmable tokens are omitted so resolve_betsi_codes() leaves them unresolved. Suitable to pass as the genus_dict argument.

Usage

inrae_genus_dict()

Value

Named character vector mapping token to genus.

See Also

resolve_betsi_codes()

Description

The taxonomic backbones only. This is the set enrichment name resolution unions over, so reference geometry (list_geometry()) is not part of it even though build_backend() builds both.

Usage

list_backends()

Value

Character vector of backend identifiers.

Description

The per-taxon assets rebuilt from published BETSI-derived matrices. Each is a normal entry in the enrichment build registry; this lists the recovery subset build_betsi_recovery() accepts.

Usage

list_betsi_recovery()

Value

Character vector of recovery enrichment names.

See Also

build_betsi_recovery(), parse_betsi_recovery()

Description

Candidate enrichment sources that taxifydb deliberately does not build into a .vtr, record in manifest.json, or place in the cross-source trait registry, because each source's own current licence permits citation and personal/scientific use only, not third-party redistribution. Unlike the scrape-only sources in list_scrape_only_enrichments() (which taxify still reaches at runtime on the user's own machine), these have no runtime accessor. The catalog records the verified licence decision so a source that resurfaces is not re-evaluated from scratch; a source moves into the build registry only if its live licence is relicensed to permit redistribution.

Usage

list_build_only_enrichments()

Value

A data.frame with columns name, source, organism_groups, candidate_trait, source_url, license, access, reason, attribution, and issue.

See Also

list_scrape_only_enrichments(), list_enrichments()

Description

List available enrichment builders

Usage

list_enrichments()

Value

Character vector of enrichment identifiers.

Description

Boundary polygons taxify reads for the ⁠region=⁠ / ⁠coords=⁠ constraint. Published and versioned like a backbone, but they carry no taxonomic names.

Usage

list_geometry()

Value

Character vector of geometry identifiers.

Description

taxifydb builds no .vtr for these sources because they cannot be redistributed. Pignatti's values originate in a copyrighted publication and are accessed on demand by taxify's add_pignatti() through the TR8 package (which ships a copy under its own licence). This returns the catalog of those sources.

Usage

list_scrape_only_enrichments()

Value

A data.frame with columns name, tr8_db, taxify_fn, license, and reason.

Description

Renames source-specific columns to canonical names and ensures consistent types and formatting (uppercase ranks/statuses, trimmed whitespace, etc.).

Usage

normalize_backbone(df, col_map, extra_cols = NULL)

Arguments

Value

A data.frame with standardized column names and formatting.

Description

Split out of read_colxr() so the streaming build can apply it to a chunk at a time. Nothing here depends on rows outside the block.

Usage

normalize_colxr(df, gbif_crosswalk = NULL, publications = NULL, verbose = TRUE)

Arguments

Value

A normalized data.frame.

Description

Split out of read_gbif() so the streaming build can apply it to a chunk at a time. Everything here depends only on the row in hand once higher is supplied, which is the whole reason that lookup is a separate argument: the ⁠*_key⁠ columns point at rows anywhere in the file.

Usage

normalize_gbif(df, higher, counts = NULL, verbose = TRUE)

Arguments

Value

A normalized data.frame ready for precompute_backbone().

Description

Split out of read_wcvp() so the streaming build can apply it to a chunk at a time. Acceptance is read from a comparison of two columns of the same row, so nothing here depends on rows outside the block.

Usage

normalize_wcvp(df, verbose = TRUE)

Arguments

Value

A normalized data.frame ready for precompute_backbone().

Description

Split out of read_wfo() so the streaming build can apply it to a chunk at a time. Nothing here depends on rows outside the block.

Usage

normalize_wfo(df, verbose = TRUE)

Arguments

Value

A normalized data.frame.

Description

Split out of read_worms() so the streaming build can apply it to a chunk at a time. Nothing here depends on rows outside the block.

Usage

normalize_worms(df, verbose = TRUE)

Arguments

Value

A normalized data.frame.

Description

Parse AlgaeTraits macroalgal traits (WoRMS ZIP export)

Usage

parse_algae_traits(path)

Arguments

Value

data.frame with canonical_name + algae trait columns.

Description

Reads the public dataset table and reduces it to one row per species x location. The release's locations are non-overlapping: countries, and islands or other parts of a country recorded as regions of their own. A location that is a whole country is keyed on its ISO 3166-1 alpha-2 code and a sub-national one on its own name (.seebens_location_key()), in the location column; country_code carries the ISO code of the country every location lies in, so a roll-up to countries is an explicit choice made on that column. Folding the parts into their country would date the country's first record from an island's: in v4.0 that moved 665 species x country years earlier (Hawaii alone 337, by a median of 25 years) and credited 3,541 pairs to a country whose own location has no record of the species.

Usage

parse_alien_first_records(path)

Arguments

Details

Every location the source names must map. The map is keyed on wording, so a release that renames a place would otherwise drop its records and return a smaller table rather than an error: v4.0 alone renamed 24 locations, among them "United States" to "United States of America", which is 8,050 records. A name the map does not hold is therefore a hard error, and the entries that resolve to no single country (a record spanning several, "USACanada") are recorded in the map as NA so they read as known rather than missing.

Where a species has several records for one location the earliest year wins, but a record asserting the species is present is preferred to one recording it as absent, uncertain or captive whatever the years are, so the retained row does not date a country's occurrence from a record denying it. That record's own status is published beside its year as alien_first_record_status, since the remaining columns are reduced over every record for the pair and a status reduced that way would describe some record other than the one the year came from. The same earliest-wins reduction (.keep_earliest_first_record()) runs again after cross-backbone name resolution, at the accepted-species grain, so synonyms collapsing onto one concept take its earliest observed year rather than an arbitrary one. The two convention markers (.alien_first_record_sentinels) are dropped to NA here so neither can win that reduction or be served as a year.

Value

data.frame with canonical_name + location + country_code + first-record cols.

Description

Traits.csv from the Dryad deposit behind Chamberlain et al.'s study of carabid occurrence along Alpine elevation gradients: 185 species recorded at 416 sites between 697 and 2840 m in the Italian Alps, semicolon-delimited.

Usage

parse_alpine_carabids(path)

Arguments

Details

The widest European carabid trait block after the German table, and the one that reaches the Alpine fauna the lowland compilations miss. Its body sizes share the carabids.org ancestry every European carabid source has (23.9% exactly equal to the German table over 92 shared species), so agreement with those is not corroboration; it runs 1.03 against them, close enough to be the same quantity and far enough from 1.000 not to be a verbatim copy in the way the Eberswalde table is.

The wing column ships as bare letters with no legend, so the codes were read off the data rather than assumed: crossed against the German table's words, b is short-winged (16 of 18), m long-winged (50 of 55) and d dimorphic (18 of 19), which is brachypterous, macropterous and dimorphic.

Value

data.frame with canonical_name, body_length_mm, wing_morph.

Description

One row per species across birds, mammals and reptiles. The dataset codes missing values as -999; these are converted to NA on every numeric trait.

Usage

parse_amniote(path)

Arguments

Value

data.frame with canonical_name + life-history traits.

Description

Parse AmphiBIO amphibian traits (CSV from ZIP)

Usage

parse_amphibio(path)

Arguments

Value

data.frame with canonical_name + trait columns.

Description

Parse AnAge longevity and life-history traits (TSV from ZIP)

Usage

parse_anage(path)

Arguments

Value

data.frame with canonical_name + AnAge trait columns.

Description

Parse AnimalTraits observations CSV (aggregate to species medians)

Usage

parse_animaltraits(path)

Arguments

Value

data.frame with canonical_name + body_mass_kg + metabolic_rate_w.

Description

Long-format fuzzy-coded traits (taxon x trait x category affinity); each trait is reduced to its dominant category per species, using the database's own category labels.

Usage

parse_arctic(path)

Arguments

Value

data.frame with canonical_name + benthic traits.

Description

Parse NW European Arthropod DwC-A (taxon + measurement + description)

Usage

parse_arthropod_traits(dir_path)

Arguments

Value

data.frame with canonical_name + arthropod trait columns.

Description

Reads the long traits.csv from the plain-text release and pivots a set of high-coverage functional traits to one row per taxon, aggregating across all records (numeric by median, categorical by mode). Numeric traits such as leaf mass per area are recorded mostly at population/individual level, so records of every entity type are aggregated up to the taxon.

Usage

parse_austraits(path)

Arguments

Details

Every trait column ⁠<col>⁠ has a ⁠<col>_source⁠ column naming the AusTraits datasets (dataset_id) behind the value: for a categorical trait the datasets whose records state the reported value, for a numeric trait every dataset whose records enter the median. The ids resolve against the reference table attached with attach_references(), built from the release's methods.csv (each dataset's primary source citation) and sources.bib (its DOI).

Value

data.frame with canonical_name + plant traits.

Description

Parse AVONET bird morphology XLSX

Usage

parse_avonet(path)

Arguments

Value

data.frame with canonical_name + morphology columns.

Description

Reads the gzipped JSONL of full strain records, extracts phenotypic and growth-condition traits per strain, and aggregates to one row per species (categorical by mode, numeric by median). Temperature and pH prefer the optimum measurement, falling back to the growth measurement; ranges are reduced to their midpoint.

Usage

parse_bacdive(path)

Arguments

Value

data.frame with canonical_name + microbial trait columns.

Description

Reads the baseflor sheet of Julve's baseflor.xlsx, builds a clean binomial canonical_name from the nomH/nomB/nomA columns, splits the floraison flowering-period field into begin/end months, and recodes the French categorical traits (pollination vector, dispersal mode, breeding system, flower colour, fruit type, woody growth form) to English. Also keeps the two Ellenberg-style axes absent from EIVE: continentality and salinity.

Usage

parse_baseflor(path)

Arguments

Details

Indicator values L/T/M/R/N are intentionally not emitted here (the European-calibration EIVE enrichment already covers them); Raunkiaer life form is left to the leda enrichment for the same flora; maximum vegetative height is omitted because the source column is empty.

Value

data.frame with canonical_name + 10 trait columns.

Description

Long-format Darwin Core measurement table; reduced to one row per species. verbatimAcceptedNameUsage carries authorship and subgenera, so the join key is reduced to a clean binomial.

Usage

parse_bee_ostwald(path)

Arguments

Value

data.frame with canonical_name + bee morphometrics.

Description

One row per species. The source is R write.table output (space-separated, quoted strings, leading row-id column). Ellenberg-style indicator values use "x" for indifferent, which is mapped to NA. The six one-hot substrate class flags are also collapsed into a pipe-delimited substrate set (e.g. "soil|rock|bark") over soil, rock, bark, wood, living_plants and dung_carcass.

Usage

parse_bet(path)

Arguments

Value

data.frame with canonical_name + bryophyte traits.

Description

Reads the Zenodo BETSI body-length export and reduces its per-source measurements to one row per Collembola species. betsi_body_length_mm is the median of the recorded measurements (mm); betsi_body_length_min_mm / betsi_body_length_max_mm bound them; betsi_body_length_n and betsi_body_length_sources are the number of measurements and of distinct literature sources behind the species' value.

Usage

parse_betsi_collembola_body_length(path)

Arguments

Value

data.frame with canonical_name and the ⁠betsi_body_length_*⁠ columns, one row per species.

Description

Reads the frozen matrices for the sources feeding enrichment and returns one row per species. Fuzzy sources (⁠species,trait,class,pct⁠) are checked for the fuzzy-coding invariant (each species-by-trait affinity block sums to 100, tolerating +-1 from integer rounding) and pivoted to one numeric column per ⁠<trait>__<modality>⁠ bin (0-100). Hard sources (one value per trait) are read into one column per trait. Provenance is not a data column here; it is written to meta.json from the descriptor by the build.

Usage

parse_betsi_recovery(enrichment, path)

Arguments

Value

data.frame with canonical_name and the trait columns, one row per species.

See Also

build_betsi_recovery()

Description

Per-region records (one row per species x marine region) reduced to species-level values (numeric by median, categorical by mode).

Usage

parse_beukhof(path)

Arguments

Value

data.frame with canonical_name + marine-fish traits.

Description

Like fishbase via rfishbase, BIEN is queried directly: a single bulk pull of the selected traits across all species (one query per trait list), filtered to public-access records, then pivoted to one row per species (numeric by median, categorical by mode). The path argument is unused (kept for a uniform parser interface). Curated numeric columns are checked against the unit each BIEN record carries and converted to the unit in the column name (leaf dry mass is recorded in g and written in mg).

Usage

parse_bien(path)

Arguments

Value

data.frame with canonical_name + plant traits.

Description

Reads the Data sheet (real header is the second physical row). Presence-coded columns (island, restricted-range) use a blank for absent, mapped to 0. Traits redundant with AVONET morphology are not carried. One row per species.

Usage

parse_birdbase(path)

Arguments

Value

data.frame with canonical_name + bird traits.

Description

Genus-level ant traits with codes defined in the source headers; the clearly defined ones are mapped to their labels. Columns are read by position because the source header cells embed long code legends.

Usage

parse_blanchard(path)

Arguments

Value

data.frame with canonical_name (genus) + ant traits.

Description

Long-format trait records reduced to one row per species. Numeric units follow the standardised BROT 2.0 conventions (seed mass mg, SLA mm2/mg, height m, leaf area mm2).

Usage

parse_brot(path, sources = file.path(dirname(path), "brot_sources.csv"))

Arguments

Details

Every trait column ⁠<col>⁠ has a ⁠<col>_source⁠ column naming the BROT sources (SourceID) behind the value: for a categorical trait the sources whose records state the reported value, for a numeric trait every source whose records enter the median. The ids resolve against BROT's own sources file (BROT2_sou.csv: ID, FullSource), attached with attach_references().

Value

data.frame with canonical_name + plant traits.

Description

One row per moss, liverwort or hornwort taxon. Retained traits are the five Ellenberg indicator values (light, moisture, reaction, nitrogen, salt), and the source's life-form, plant-group (moss / liverwort / hornwort) and status codes (kept verbatim).

Usage

parse_bryoatt(path)

Arguments

Value

data.frame with canonical_name + British/Irish bryophyte attributes.

Description

Somatic chromosome numbers for plants, taken from the CCDB statistics service (one server-side per-taxon reduction per major plant group). Names are collapsed to the binomial (infraspecific and authorship qualifiers are dropped) and reduced to one row per species: the median chromosome number (2n) with its minimum and maximum across the aggregated records. The service reports gametic numbers, which are doubled to the somatic 2n the columns name; the spread between minimum and maximum is real ploidy variation among a species' cytotypes.

Usage

parse_ccdb(path)

Arguments

Value

data.frame with canonical_name + chromosome-number traits.

Description

Fuzzy-coded biological traits of North-West European benthic invertebrates, distributed as a flat one-row-per-taxon matrix at genus rank and above. Each of the ten biological traits is split across several modality columns holding a 0-3 fuzzy score (0 = no evidence ... 3 = strong evidence); each trait is collapsed to the single modality with the highest score. The traits are body size (mm class), morphology, lifespan (year class), egg development, larval development, living habit, sediment position, feeding mode, mobility and bioturbation mode. The WoRMS AphiaID is present in the source but the Genus-rank name is used as the resolution key.

Usage

parse_cefas_btrait(path)

Arguments

Value

data.frame with canonical_name + benthic biological traits.

Description

Reads the data sheet (real header is the second physical row), treats the literal "NA" as missing, and parses "min-max" text ranges to their midpoint. One row per species.

Usage

parse_chelonians(path)

Arguments

Value

data.frame with canonical_name + turtle traits.

Description

Supporting Information Data S1 of "Widespread Decline of Ground Beetles in Germany" (Diversity and Distributions, doi:10.1111/ddi.70112). One row per species: 383 rows, of which 382 carry trait values (Ocys tachysoides has none), which makes this by a wide margin the largest lawfully reachable carabid trait block.

Usage

parse_chowdhury(path)

Arguments

Details

Trait values are kept verbatim rather than normalized here. The source vocabularies are small and self-describing, and taxify's cross-source trait registry is where a vocabulary is mapped, so the .vtr keeps what the authors wrote and the registry decides what it means.

Read off the file rather than the paper: trophicLevel has three values (Predator 274, Herbivore 80, Omnivore 28) and no mycetophage category, and habitatPref has seven classes, not eight.

Alongside the traits the table carries the paper's own result, a modelled two-year occupancy trend per species. Nothing else in taxify holds a national population trend, so those columns are kept.

Value

data.frame with canonical_name plus trait and trend columns.

Description

One row per species for 29 traits of clonal growth, bud banks and lifespan. Names are collapsed to the binomial (authorship and infraspecific rank are dropped, hybrid markers kept), so infraspecific records are aggregated to the species. Numeric traits (growth-form flags, bud-bank sizes/depths, persistence, offspring, spread, clonal index) are reduced by median; the nominal coded traits (branching, cyclicity, clonal-growth-organ type, regenerative-bud position and role) by mode. Coded fields are kept verbatim as in the source; consult the CLO-PLA metadata to decode them.

Usage

parse_clopla(path)

Arguments

Value

data.frame with canonical_name + CLO-PLA clonal / bud-bank traits.

Description

Uses the reported (observed/compiled) trait table, not the phylogenetically imputed one, so gap-filled numbers are never shipped as observed. Keyed on the IUCN 2020 binomial.

Usage

parse_combine(path)

Arguments

Value

data.frame with canonical_name + mammal traits.

Description

Merges vernacular names from three sources:

• GBIF: VernacularName.tsv (has ISO 639-1 language codes)

• NCBI: names.dmp where name_class == "common name" (no language)

• OTT: synonyms.tsv where type == "common name" (no language)

Usage

parse_common_names(dir_path)

Arguments

Details

NCBI and OTT common names have no language tag, so lang is set to NA.

Value

data.frame with canonical_name + lang + common_name.

Description

Species-level functional traits of marine copepods, distributed as a multi-sheet PANGAEA workbook (one sheet per trait family). Each trait sheet is melted to (taxon, trait, value); the binary feeding-mode and spawning-strategy flags are first collapsed to a single categorical label per record. Records are then reduced to one row per taxon (numeric traits by median, categorical by mode). Subgenus qualifiers in parentheses (⁠Acartia (Acartiura) clausi⁠) are stripped so the name matches a backbone binomial or genus. Retained traits: body length (mm), egg outer diameter (micrometres), clutch size (egg count), feeding mode, feeder type, spawning strategy, presence of myelination and of resting eggs.

Usage

parse_copepod_traits(path)

Arguments

Value

data.frame with canonical_name + copepod traits.

Description

The release is long format (one row per observation). This pivots a set of high-coverage functional traits to one row per species, aggregating numeric traits by median and categorical traits by mode.

Usage

parse_coral_traits(path)

Arguments

Value

data.frame with canonical_name + coral trait columns.

Description

Parse Diaz et al. 2022 supplementary traits (XLSX)

Usage

parse_diaz_traits(path)

Arguments

Value

data.frame with canonical_name + seed_mass_mg + plant_height_m.

Description

Genus-level fuzzy-coded traits. Each fuzzy trait group (body size, life cycle, reproductive cycles, dispersal strategy, adult life span, female wing length, wing-pair type, fecundity) is reduced to its dominant modality, labelled with the database's own modality descriptions. Modality labels are ASCII-normalized (<=, >=, -), and the three binned physical-magnitude ordinals (disperse_body_size_cm, disperse_female_wing_mm, disperse_fecundity) also get a coarse ⁠_mid⁠ numeric-midpoint column for numeric joins; the count/time ordinals stay categorical.

Usage

parse_disperse(path)

Arguments

Value

data.frame with canonical_name (genus) + dominant-modality traits + the three ⁠_mid⁠ midpoint columns.

Description

⁠EWcarabids1999-2022_species_trends_traits.csv⁠ from the Leuphana deposit behind the 24-year drought study of Weiss, von Wehrden and Linde: 27 species from 13 forest plots near Eberswalde, semicolon-delimited with German decimal commas.

Usage

parse_eberswalde(path)

Arguments

Details

Two of its columns are carabids.org verbatim, which the deposit's own README states outright ("size = mean size (body length, mm) ... Source: Homburg et al. (2014) / carabids.org", likewise wings and latitude). The data confirm it: over the 19 species shared with the Chowdhury table every single size is exactly equal and every wing class agrees. They are kept because a door should surface what its source carries, and they are excluded from the cross-source trait registry, where they would double-count one lineage.

What belongs to this deposit alone: a 24-year local abundance trend, a sensitivity to the 72-month SPEI drought index, and a feeding guild refined by the authors' field observations to name the prey (Gastropoda, Collembola). The humidity preference is Sustek's (2004) 1-8 scale, an independent source.

Value

data.frame with canonical_name plus trait columns.

Description

Reads the per-species Ecoflora scrape (results.csv, one row per species, trait columns named by their Ecoflora short codes) and returns a clean wide data.frame keyed on canonical_name. Numeric fields (heights, seed weight, flowering months) take the median of any multiple scraped values; categorical fields keep the unique values joined with "; ". Ecoflora records height in cm ("Height: 3. typical maximum (cm)", "Height: 4. typical minimum (cm)") and seed weight in mg ("Seed weight: 3. mean (mg)"); heights are written in mm, seed weight as recorded. Every trait column carries a ⁠_uk⁠ suffix to mark the British-flora calibration and to avoid collisions when chained with other plant-trait enrichments.

Usage

parse_ecoflora(path)

Arguments

Details

Ecoflora (Fitter & Peat 1994) has no bulk download or API; the snapshot was collected one species at a time and is redistributed under the source licence (CC BY-NC-SA 4.0).

Value

data.frame with canonical_name + 18 ⁠_uk⁠ trait columns.

Description

Reads the Zenodo ecomorphosis species list and returns one row per species known to display ecomorphosis.

Usage

parse_ecomorphosis(path)

Arguments

Details

Columns:

• ecomorphosis – TRUE for every listed species (a presence list).

• ecomorphosis_area – geographical area the record applies to.

• ecomorphosis_reference – the literature record establishing it.

Value

data.frame with canonical_name and the ⁠ecomorphosis*⁠ columns, one row per species.

Description

Per-culture nitrogen (nitrate, ammonium) and phosphorus utilization traits (maximum growth rate, half-saturation constants for growth and uptake, minimum subsistence quota, maximum quota, maximum uptake rate) from Edwards et al. (2015, Ecological Archives E096-202). Each source row is one culture measurement; rows are reduced to one per species (numeric traits by median, taxon group and freshwater/marine system by mode). Experimental conditions (temperature, irradiance, daylength) and per-measurement citations are not carried. Nutrient-trait column names are kept verbatim from the source: units are documented in the source metadata, and the source is internally inconsistent on the V_max time unit, so no unit is asserted in the name.

Usage

parse_edwards_phyto(path)

Arguments

Value

data.frame with canonical_name + phytoplankton nutrient traits.

Description

Parse EIVE 1.0 ecological indicator values (XLSX)

Usage

parse_eive(path)

Arguments

Value

data.frame with canonical_name + indicator columns.

Description

Reads the CC0 Collembola_trait_data.csv (Ellers et al. 2018) and returns one row per species. Ordinal codes whose levels are named in the source are decoded (reproduction, moisture preference); the biogeographic temperature-zone class and the thermal-niche-breadth count are kept as their source ordinal codes.

Usage

parse_ellers_collembola(path)

Arguments

Details

Columns:

• ellers_vertical_distribution – soil, subsurface or surface.

• ellers_body_size_mm – body size (mm).

• ellers_reproduction – asexual or sexual.

• ellers_moisture_pref – xerophilous, xero-mesophilous, mesophilous, hygro-mesophilous or hygrophilous.

• ellers_temperature_pref – biogeographic temperature-zone class, ordinal 1-5 (1 = boreal zone only ... 5 = Mediterranean zone only) per the source codebook.

• ellers_thermal_niche_breadth – number of biogeographic zones occupied, ordinal 1-3.

Value

data.frame with canonical_name and the ⁠ellers_*⁠ trait columns, one row per species.

Description

Parse EltonTraits 1.0 birds + mammals TSVs

Usage

parse_elton_traits(birds_path, mammals_path)

Arguments

Value

data.frame with canonical_name + diet/foraging/body mass columns.

Description

Ecological and life-history traits of North-American freshwater macroinvertebrates, read from the transposed distribution (one row per taxon and source citation). A curated set of interpretable primary trait columns is melted to (taxon, trait, value) and reduced to one row per taxon by mode, so the multiple per-taxon citation records collapse to the dominant state. The retained traits are primary feeding mode, habit, voltinism, thermal preference, maximum body-size class, body shape, rheophily, primary oviposition behaviour and diapause. ⁠Other (specify in comments)⁠ placeholder entries are dropped. The taxon grain is mixed (species and genus); the ITIS TSN is present in the source. This dataset shares its lineage (USGS Data Series 187, Vieira et al. 2006) with freshwater_insects_conus, which is at genus grain; the species-grain records here are the net-new contribution.

Usage

parse_epa_freshwater(path)

Arguments

Value

data.frame with canonical_name + freshwater macroinvertebrate traits.

Description

Long-format Darwin Core measurement-or-fact table joined to the taxon core (genus + specific epithet) for clean binomials; reduced to one row per species for a curated set of morphological, biogeographic and ecological traits.

Usage

parse_eupolltrait(path)

Arguments

Value

data.frame with canonical_name + pollinator traits.

Description

Species-level traits of European bats spread across thematic Darwin Core measurement-or-fact tables (morphology, life history, diet, foraging habitat, roost type). Each table is melted to (species, trait, value), stacked, and reduced to one row per species (numeric traits by median, categorical by mode). Underscored source names (Genus_species) are normalised to a binomial.

Usage

parse_eurobat(path)

Arguments

Value

data.frame with canonical_name + European bat traits.

Description

Reads euromed_distribution.jsonl and emits one row per (accepted taxon, Euro+Med area): the status class, the term as the portal states it, the area with its ISO 3166-1 alpha-2 country code where the area is one country, the taxon's CDM UUID and the ids of the references behind the reported status. A retracted report (⁠reported in error⁠) is dropped; a former presence is extinct. Where records of one taxon-area disagree the status follows .euromed_status_order and the references are those stating it.

Usage

parse_euromed_distribution(path)

Arguments

Details

Areas whose name is not in .euromed_area_iso and is not a combined area or part of a country are reported so the table can be extended.

Value

data.frame with canonical_name, area_code, area_name, area_level, iso2, euromed_status, euromed_status_detail, euromed_status_source and taxon_id, with a reference table (ref_id, citation, doi) attached by attach_references().

Description

FAPROTAX.txt from the release zip (Louca et al. 2016, Science 353:1272-1277): 92 metabolic and ecological function groups, each listing the prokaryotic taxa known to perform it, compiled from IJSEM and Bergey's Manual.

Usage

parse_faprotax(path)

Arguments

Details

The file is a grouped list rather than a table. A group begins with its name plus a tab-separated attribute string; the lines under it are member taxa written as taxonomic path suffixes (⁠*Proteobacteria*Nitrosomonas*⁠, ⁠*Escherichia*coli*⁠). A group may also inherit or exclude another group's members through ⁠add_group:⁠ / ⁠subtract_group:⁠, which are resolved here in file order – a referenced group is always defined earlier.

Members are annotated at whatever rank the evidence supports, so the file mixes species, genera and higher taxa in one list. taxify joins an enrichment on a species name or a genus, so each member is reduced to whichever of those it carries: the last genus-plus-epithet pair in the path if there is one, else the last plain capitalised name. Strain designations (107, ⁠4_3_47FAA⁠, ⁠2.4.3⁠) carry neither and are dropped. Rows keyed on a rank above genus are kept as written – they simply never match a genus, at the cost of a few hundred rows.

The result is one row per taxon with the function groups it belongs to as a pipe-delimited set, the same shape the NestTrait modality columns use: a prokaryote genuinely performs several of these, and collapsing to one would be a choice the source does not make.

Value

data.frame with canonical_name, faprotax_functions, faprotax_n_functions.

Description

Species_traits_zenodo.xlsx from the Zenodo deposit behind Finand & Kotze's urban forest fragmentation study: 34 species from 25 remnant forests in Helsinki, with body length measured from the specimens the authors caught rather than compiled from a database.

Usage

parse_finand(path)

Arguments

Details

That independence is the reason to carry a table this small. Every other reachable carabid trait source traces back to carabids.org (Homburg et al. 2014), so none of them can check another. This one can: its body lengths run at a median ratio of 0.9946 against arthropod_traits over 28 shared species (IQR 0.985-1.008) with only 7.1 percent exactly equal, and at ~1.00 against the Chowdhury table over 14. Values such as 6.066667 mm are means over individuals, not a rounded lookup.

Body_length is a whole-body length, not the elytra length the name might suggest: elytra would sit near 0.6 of these values, and the published body lengths agree directly (Cychrus caraboides 16.9 mm against 14-19 in the literature, Carabus glabratus 28.6 against 22-34).

The wing column is the same three-state trait the German table carries. On the 25 species they share there is no long-versus-short disagreement at all, but seven species the German table calls dimorphic get a definite morph here, because this records the morph of the beetles actually caught in Helsinki while the German table states the species' capacity.

Value

data.frame with canonical_name plus trait columns.

Description

Parse FishBase species traits (via rfishbase)

Usage

parse_fishbase(path)

Arguments

Value

data.frame with canonical_name + body/depth/diet columns.

Description

Parse FISHMORPH freshwater fish morphological traits (CSV)

Usage

parse_fishmorph(path)

Arguments

Value

data.frame with canonical_name + morphology columns.

Description

Species-level traits of United States freshwater fishes, one row per species in a legacy Excel workbook. The table is already at the species grain, so it is read wide and a curated, documented subset of columns is retained: the binary trophic-guild indicators, life-history quantities (maximum total length in cm, age at maturity and longevity in years, maximum fecundity, spawning-season length), salinity and temperature tolerances (30-year mean January minimum and July maximum at the range centroid, in degrees C), and conservation flags. The 25 Balon reproductive-guild indicator columns are collapsed to a single repro_guild code (the flagged guild). The ITIS TSN is carried as a cross-reference. Field meanings follow the dataset's AttributeDefinitions.

Usage

parse_fishtraits(path)

Arguments

Value

data.frame with canonical_name + US freshwater-fish traits.

Description

Reads the long-format FloraWeb scrape (results_long.csv, one row per species/page/label/value) and pivots it to a clean wide data.frame keyed on canonical_name. German "no data" placeholders are dropped, multiple values for one trait are collapsed to unique values joined with "; ", and the German trait labels are mapped to English column names (each with a ⁠_de⁠ suffix). Covers morphology, reproductive biology, the nine Ellenberg indicator values, ploidy and chromosome number, and chorological distribution.

Usage

parse_floraweb(path)

Arguments

Details

FloraWeb (Bundesamt fuer Naturschutz) is the live portal carrying the BiolFlor trait data (Klotz, Kuehn & Durka 2002) plus Rothmaler morphology and Ellenberg indicator values. It has no bulk export or API, so the snapshot was scraped per species; the access date is its version.

Value

data.frame with canonical_name + the mapped ⁠_de⁠ trait columns.

Description

Combines the "simple" mammal and bird tables on a shared core (diet, body size/mass, longevity, generation time) with a taxon_group discriminator.

Usage

parse_frugivoria(path)

Arguments

Value

data.frame with canonical_name + traits.

Description

Reads the Polme et al. (2020) supplementary XLSX, selects the most informative trait columns, and returns a data.frame keyed on genus.

Usage

parse_fungal_traits(path)

Arguments

Value

data.frame with genus + 9 trait columns.

Description

Reads the FungalRoot occurrence core (occurrences.csv) and its MeasurementOrFact extension (measurements.csv), keeps the per-observation ⁠Mycorrhiza type⁠ measurements, standardizes them to a small set of type tokens, and reduces them to one row per plant genus by majority consensus. Mycorrhizal type is phylogenetically conserved at the genus level, which is the resolution FungalRoot itself recommends for inference; the per-genus value here is the most frequent standardized type across that genus's observations (taxifydb's own aggregation, not FungalRoot's published per-genus assignment).

Usage

parse_fungalroot(path)

Arguments

Value

data.frame keyed on genus (also copied to canonical_name) with mycorrhizal_type, mycorrhizal_status, and mycorrhizal_records.

Description

Reads the JSON array returned by the FUNGuild API, filters to genus and species-level entries, and returns a clean data.frame with canonical_name and trait columns.

Usage

parse_funguild(path)

Arguments

Value

data.frame with canonical_name + trophic_mode + guild + growth_morphology + confidence_ranking.

Description

Genus-level ecological and life-history traits of North American freshwater insects, distributed in long form (one row per genus and trait group). Pivoted to one row per genus; every trait group is categorical (the source ships the abbreviated modality codes, kept verbatim).

Usage

parse_fw_insects_conus(path)

Arguments

Value

data.frame with canonical_name (genus) + freshwater-insect traits.

Description

Parse GBIF vernacular names (VernacularName.tsv + Taxon.tsv)

Usage

parse_gbif_common_names(dir_path)

Arguments

Value

data.frame with canonical_name + lang + common_name.

Description

Reads the GIDIAS machine-readable impact records and reduces them to per-species environmental- and socio-economic-impact indicators. The .vtr carries these derived aggregates; for the individual impact records, use the GIDIAS figshare download.

Usage

parse_gidias(path)

Arguments

Details

gidias_eicat_category is the species' EICAT category: the most severe magnitude among its negative (harmful) environmental-impact records, mapped from the GIDIAS 0-3 magnitude scale to the IUCN levels MC/MN/MO/MR/MV (the per-record global-extinction flag splits magnitude 3 into local MR and global MV); a species with negative records but no scored magnitude is "DD". gidias_seicat_category applies the same rule to the constituents-of- well-being block (SEICAT, MC/MN/MO/MR). The remaining columns summarise the evidence: the driving mechanism(s), affected well-being constituents, the species' functional group and realms, and record/source counts.

Nature's Contributions to People (NCP) is reported as a direction rather than a category: gidias_ncp_direction gives the direction(s) recorded ("Negative", "Positive", "Neutral", or a combination for the 16% of species with a mixed record) and gidias_ncp_affected the contributions affected. GIDIAS scores no magnitude scale for this block, so no IUCN-style category is derived; the two columns are not paired, so use the source records where the direction of a specific contribution matters.

Each species is aggregated twice. affected_taxon = "Any" summarises every record, and is the grain to use unless you want one affected group. The other rows summarise the species' impacts on one affected native taxon ("Plant", "Invertebrate", "Vertebrate", "Microbe", "Fungi"), which 28% of species split across two or more. Only the "Any" row carries SEICAT and the negative records with no affected taxon recorded: the affected-taxon axis slices the environmental-impact block, so the SEICAT columns are NA on a per-taxon row.

Value

data.frame with canonical_name, affected_taxon, and the ⁠gidias_*⁠ aggregate columns.

Description

Downloads GIFT's full catalogue of species-level traits from the live API (batched over all trait IDs) and pivots them to one row per species. GIFT aggregates records to a single value per species (mean for numeric traits, most frequent entry for categorical ones). Only the redistributable subset the API returns is kept; restricted references are excluded by the default call. The species name is GIFT's work_species.

Usage

parse_gift(path = NULL, agreement = 0.66, batch_size = 12L, verbose = TRUE)

Arguments

Details

Every trait column ⁠<col>⁠ has a ⁠<col>_source⁠ column holding the GIFT reference ids (ref_ID) GIFT lists for the aggregated value (the references field GIFT::GIFT_traits() returns beside it). For a categorical trait these are the references that state the reported value; for a numeric trait the references entering the mean. GIFT prefixes some ids with - or +, a marking its documentation does not define. The sign is dropped from the cell and kept per reference in the table as listed_negative / listed_positive. A negatively listed reference never appears among the records GIFT_traits_raw(bias_ref = FALSE) returns (the tree-only GlobalTreeSearch is one), which fits GIFT's bias_ref flag. The reference table comes from GIFT::GIFT_references(); an id GIFT lists but does not serve there has NA citation and a note saying so. It is attached with attach_references().

Value

data.frame keyed on canonical_name with one ⁠gift_<trait>⁠ column per GIFT trait. Numeric traits are doubles; the rest are character.

Description

Reads the crawled distinct (source, type, target) edges, resolves both endpoints to their cross-backbone accepted names, and counts, per accepted species, the number of distinct interaction partners (undirected), distinct interaction types, and total interaction records. Degree is a distinct-count, so it is aggregated at the accepted grain (both endpoints resolved) rather than per raw name; the registry entry therefore sets resolve_names = FALSE.

Usage

parse_globi(path)

Arguments

Value

data.frame with canonical_name, interaction_degree, n_interaction_types, n_interaction_records.

Description

One row per species, wide format with two measurement blocks (upper/heat and lower/cold limit), each carrying its own metric definition. The metric columns are kept because a temperature value is ambiguous without them (CTmax, LT50 and upper thermoneutral zone are different quantities).

Usage

parse_globtherm(path)

Arguments

Value

data.frame with canonical_name + thermal-limit traits.

Description

Parse GloNAF taxon-region data (multiple CSVs/XLSXs from ZIP)

Usage

parse_glonaf(dir_path)

Arguments

Value

data.frame with canonical_name + region_id + naturalized.

Description

Host-parasite association records for wild ungulates, carnivores and primates, aggregated to one row per host species. Reported traits are the parasite richness (distinct parasite species), the distinct-parasite count by parasite type (helminth, virus, bacteria, protozoan, arthropod, fungus, prion), the mean reported prevalence and the host group.

Usage

parse_gmpd(path)

Arguments