Release v5.0.0-alpha.2 · fluree/db

What's Changed

RDF 1.2 triple terms: rdf:reifies links, triple terms as values, faster annotation queries by @bplatz in #2017
feat(indexer): one leaf size for import, rebuild and incremental builds by @bplatz in #2051
Lock file storage by default; route CLI commands before opening storage by @bplatz in #2132
Iceberg on Google: a Google identity for REST catalogs, a per-source storage identity with a GCS reader, and one codec for every surface by @aaj3f in #2196
doc ingest: --ontology replaces --model, --catalog names --entities (and takes a mapped table) by @aaj3f in #2206
fix(query): a graph source that can't serve a query says why, and registration refuses one that never could by @aaj3f in #2190
feat(lists): each list on a property keeps its own id; lists read and edited as RDF on request by @bplatz in #2229
fix(config): a branch or a copy reads its source's ledger config by @bplatz in #2249
feat: DROP takes a graph and its history out of the ledger by @bplatz in #2251

RDF 1.2 in Python, and RDF documents without a ledger (fluree.parse / serialize, fluree_db_api::rdf) by @bplatz in #2042
feat(index): small named graphs share one pooled index tree by @bplatz in #2140
fluree ontology from tables (CSV, XLSX, SQL DDL) and fluree catalog from rows by @aaj3f in #2210
fluree ontology from a semantic layer: Snowflake semantic views and Genie exports by @aaj3f in #2213
fluree ontology from dbt projects and Apache Ossie documents by @aaj3f in #2215
fluree ontology from a Neo4j schema export by @aaj3f in #2216
doc ingest: read with an HTTP reader ([doc.reader], --model, --cache) by @aaj3f in #2207
doc read: read one file with an HTTP reader and print the reading by @aaj3f in #2209
python: fluree.doc.read reads a document with an HTTP reader (the CLI's own read) by @aaj3f in #2219
fluree login / logout: sign a machine in to a fluree.ai account by @aaj3f in #2205
feat: bulk import into a new named graph of an existing ledger by @bplatz in #2232

Graph sources: a crawl keeps its ORDER BY, and a fused scan stays above an OPTIONAL by @aaj3f in #2199
fix(query): every R2RML lane answers an un-annotated column the way the generic pipeline does by @aaj3f in #2193
fix(query): a planner stats view is keyed by the index root it came from by @aaj3f in #2200
fix(core): disk cache byte accounting is exact around its scans by @aaj3f in #2183
fix(query): a subject probe over a graph with no branch finds nothing by @aaj3f in #2187
Iceberg: a table's row count comes from its manifest list when the summary has none by @aaj3f in #2198
fix(cypher): a statement's clauses see one another's writes, and the last holds by @aaj3f in #2184
fix(iceberg): a pushed FILTER literal keeps only rows it can compare exactly by @aaj3f in #2186
fix(server): a peer forwards a write under the path it was sent to by @bplatz in #2245
fix(search): a full-text index whose indexing query reasons rebuilds on sync by @bplatz in #2248
fix(api): reindex reads a branch's history through its source by @bplatz in #2246
fix(api): a push of a config write reaches a ledger of the same name by @bplatz in #2247

refactor(storage): shared byte handles from get; the disk cache as a store wrapper by @bplatz in #2035
perf(query): batched probes skip empty leaflets and gallop subject runs by @bplatz in #2050
perf: read encrypted storage by range within the cache budget by @bplatz in #2130
perf: sustained writes on large ledgers — staging, indexing, publish stalls by @bplatz in #2230

feat: inline order-preserving xsd:decimal encoding + numeric range pushdown by @bplatz in #1335
feat: ledger lifecycle — name bindings, restorable drops, fenced writers by @bplatz in #1980
style(doc): rustfmt reader/session.rs (main's fmt check) by @aaj3f in #2255
ci(release): build Apple silicon on macOS 26, for macOS 14 and later by @bplatz in #2231

Full Changelog: v4.2.5...v5.0.0-alpha.2