Commit graph

16 commits

Author SHA1 Message Date
2b1bba0703 feat(rag): show cited sources and exit cleanly on interrupt/errors
- After a synthesized answer, print a legend mapping the [n] tags the
  model actually cited to their source files (on stderr, so piping the
  answer stays clean). Only cited tags are listed, not every retrieved
  chunk.
- Catch KeyboardInterrupt/BrokenPipeError for a quiet abort and collapse
  other exceptions to a one-line message; set RAG_DEBUG=1 for the full
  traceback.
2026-05-22 09:12:22 +02:00
89c82cf804 feat(rag): install on m4 and amd, default endpoints to sgx
Add the rag CLI to the m4 and amd hosts and point its default API_BASE
and QDRANT_URL at sgx (where the gateway and Qdrant run) instead of
localhost. The services live on sgx, so a localhost default only worked
there; sgx resolves to itself on sgx, so this default is correct on every
host and leaves only RAG_API_KEY to set.
2026-05-22 09:12:22 +02:00
6fd6060dd7 fix(rag): send explicit encoding_format to avoid llama.cpp null error
When encoding_format is unset, LiteLLM forwards it to the backend as JSON
null, and llama.cpp's embeddings endpoint rejects it with a 500
("type must be string, but is null"). Pin encoding_format="float" so the
gateway always relays a string.
2026-05-22 08:34:42 +02:00
116d04665d fix(sgx): move LiteLLM off port 4000 to avoid uptime-kuma clash
Uptime Kuma already binds 4000, so the gateway never got the port and
requests hit the wrong service. Move LiteLLM to 4001 and update the rag
CLI default endpoint to match.
2026-05-22 07:08:26 +02:00
95668b71a7 feat(sgx): add CLI RAG stack (Qdrant + embeddings gateway + rag tool)
Stand up document retrieval as shared, client-agnostic primitives rather
than locking it inside Open WebUI:

- Qdrant as the LAN-reachable vector store
- LiteLLM gains a bge-m3 route so sgx:4000 also serves /v1/embeddings
- a thin `rag` CLI (ingest/query, optional coder synthesis) usable from
  any machine and from scripts

Embeddings and synthesis run on halo via the gateway; the CLI is
configured entirely through RAG_* env vars.
2026-05-22 00:35:54 +02:00
a2b7dc510b fix(pi): PI_OFFLINE 1 2026-05-16 08:51:46 +02:00
aa3bc3c457 feat(pi): package @earendil-works/pi-coding-agent as pi
Vendors the npm tarball + lockfile and wraps the `pi` binary with `fd` and
`ripgrep` on PATH. Also installs it on the m4 darwin host.

`buildNpmPackage` is pulled from `inputs.unstable` because nixos-25.11's
`prefetch-npm-deps-0.1.0` panics on cacache index entries that contain
either multiple lines or JSON values with embedded spaces (npm's
`accept: application/...; q=1.0, ...` headers). For this lockfile,
`@esbuild/netbsd-arm64` and `@rollup/rollup-linux-x64-musl` trigger
both conditions and `--map-cache` fails with `EOF while parsing a
string at line 1 column 369`. Fixed upstream in nixos-unstable, which
now uses `lines()` + `split_once('\t')`.
2026-05-13 16:34:38 +02:00
b3e5a80936 feat(packages): expose geekbench_6 as flake package
Allows running the overlay-pinned 6.7.0 build via `nix run .#geekbench_6`
without needing a host to install it.
2026-04-27 15:46:35 +02:00
04150f10d4 nix fmt 2026-02-24 13:25:42 +01:00
a8b2b36c64 feat: add zeroclaw package 2026-02-20 11:51:59 +01:00
22cc56926b refactor(nix): update system attribute usage in deploy modules
- Replaced `host.pkgs.system` with `host.pkgs.stdenv.hostPlatform.system` in `default.nix` for consistent attribute referencing.
- Aligns with recent changes in Nixpkgs attribute structure and improves compatibility.
2026-01-19 15:53:21 +01:00
549b3633d1 nixos-25.11 2025-12-02 18:58:52 +01:00
ba867efe01 fix(nixos-hosts): use arithmetic comparison in argument loop
Replaces `[[ $# > 0 ]]` with `(( $# > 0 ))` for clarity and correctness. This ensures improved readability and adherence to best practices in Bash scripting.
2025-02-26 15:42:15 +01:00
900f95169f nix fmt
Signed-off-by: Harald Hoyer <harald@hoyer.xyz>
2024-11-19 10:31:29 +01:00
90ee556acd feat: Add Brother DCP-L2530DW printer driver support
This commit adds support for the Brother DCP-L2530DW printer driver. The new file, dcpl2530dw-cups/default.nix, includes necessary build inputs and sets up installation paths. Currently, this build only supports the x86_64 platform.
2024-07-11 12:14:51 +02:00
f4e2368893 A new start 2024-03-21 15:00:36 +01:00