Fix model pages, redirects, developer pages, and charts for folded model ids

refactor(model-resolution-rework): drop client-side identity logic, use view-layer

Resilience + linux gate: connection-failure reset + pre-push DuckDB read-path gate

Cache + gzip the /models and /evals index payloads; remove temp diag route

Load parquet snapshots into in-memory DuckDB tables instead of /data cache

diag: add file-integrity hash + /tmp-vs-/data read probes

Split model and eval DuckDB projections

Log failing model and eval context

Trim DuckDB detail projection

Serialize DuckDB detail queries

Warm sidecar cache and show loading progress

Add eval query diagnostics

Fix DuckDB generation config projection

Wrap derived tags for DuckDB space

Fix DuckDB HF Space eval queries

view-data: split runAndRead/readAll and log column types on failure

Revert duckdb INSTALL/LOAD json — getRowObjectsJson() is the lib's documented path

duckdb: explicit INSTALL/LOAD json so linux-x64 binding can materialise JSON columns

Adapt upstream category-based code to v2 tag taxonomy

Merge remote-tracking branch 'origin/main' into merge/main-into-v2-cleanup

WIP: v2 cleanup checkpoint before merging origin/main

view-data: coerce BIGINT-encoded numeric strings; restore distribution embed route

view-data: switch reader to getRowObjectsJson(); drop CAST/to_json hacks

view-data: paranoid CAST every output column to a primitive type

view-data: explicit r.* enumeration + parse the wrapped columns in JS

view-data: lengthen readRows error log to 1200 chars so the full SELECT is visible

view-data: also wrap VARCHAR[] and TIMESTAMP columns; add app/error.tsx boundary

view-data: cast every to_json wrap to VARCHAR

view-data: wrap every complex column with to_json to bypass binding crashes

view-data: wrap JSON-typed parquet columns in to_json() to dodge binding crash

view-data: log SQL + error context when readRows throws

Embeds: histogram route, leaderboard slices + sort, brand mark; cross-source row dedup

Internal-feedback pass: rename to "Evaluation Cards", rework Summary view, simplify §4 metrics

Add Official/Community/All scope filter for developers; drop bar

Add rule-based policy-mode summaries for model & eval views

Cross-source dedup, plotbox polish, pretty URLs, eval page fallbacks

Auto-purge sidecar bucket when Next.js BUILD_ID changes

Bump clean-hierarchy cache version to v13 to drop stale blob

Merge cross-source benchmark families; tidy leaderboard panel + table chrome

Drop alias-only single-bench families without merging them

Restore curated benchmark families; polish frontier panel UX

Precompute eval matrices for multi-metric + per-slice leaderboards

Restore HF Open LLM v2 composite and dedup vals.ai aliases

Add local parquet read support

Sort evals list by family name; add sortable columns; use cleaned display names

Dedup logic to counts

Compute and apply cleaned benchmark counts per model

Remove raw-hierarchy fallback — only ever serve cleaned hierarchy

Harden cleanHierarchy fallback and add family-name filter chips

Bump clean-hierarchy cache version to v10 to bust stale HF Space cache