open/portfolio-optimiser
1
0
Fork 0
Code Issues Releases 4 Activity

docs(readme): name the okf 1.1 server by full path -- a bare okf finds the 0.8.5 dependency

Browse source 
The example config now has the placeholder `/path/to/okf` for `command`, and the README explains
how to find the real path (`command -v okf` outside this project's environment). It also explains
why the bare name fails under `uv run`: it resolves to this framework's own okf 0.8.5, which has
no `mcp` command. A new pin asserts that the example's command is an absolute path; it was red at
daccdbe. The xfail smoke test stays as the measurement behind the rule.

Chose this over lifting the okf dependency to 1.1: it fixes the startup failure without a version
bump and without touching --prepass-payload.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This commit is contained in:
Kjell Tore Guttormsen 2026-09-21 15:28:53 +02:00
commit 459b00f512
Signed by: ktg
SSH key fingerprint: SHA256:JakMjO6FTBBzN0Bhfj9saOoEjaFxlSdYuZQQpM/lF9Q
4 changed files with 23 additions and 11 deletions
Download patch file Download diff file Expand all files Collapse all files

21
README.md
View file

@ -360,10 +360,17 @@ bases exist), `okf_describe` (a one-line-per-document map of one base), `okf_ask
for several sub-questions in one call, with a warning when the base does not seem to cover the
question) and `okf_fetch` (one concept by name).
**What you need.** `okf` 1.1.0 or newer installed and on your `PATH`. That is separate from the
library pin in `pyproject.toml`, which does not change. Copy
[`examples/okf-server.mcp.json`](examples/okf-server.mcp.json), replace the placeholder path in
`args` with your folder of knowledge bases, and leave the rest.
**What you need.** `okf` 1.1.0 or newer, installed as its own tool (for example
`uv tool install`). That is separate from the library pin in `pyproject.toml`, which does not
change. Copy [`examples/okf-server.mcp.json`](examples/okf-server.mcp.json) and replace its two
placeholders:
- `command`: the **full path** to the `okf` 1.1 executable. Find it with `command -v okf` in a
shell where this project's environment is not active (after `uv tool install` it is usually
`~/.local/bin/okf`, written out in full). Do not write just `okf`. This framework depends on an
older `okf` library (0.8.5) whose own `okf` command has no `mcp`. Under `uv run` that one is
found first, and the server fails at startup with `invalid choice: 'mcp'`.
- `args`: your folder of knowledge bases.
**Before** - an ordinary run:
@ -393,10 +400,8 @@ With it, the server and each allowed tool are named in the announcement before t
- **Measured without a model** (`tests/test_okf_server_smoke.py`, against a small invented
collection): through the same loader and tool builder a run uses, `okf` 1.1 offers exactly the
four tools, `allowed_tools` narrows what is offered, `okf_list` names the collection and one
`okf_ask` returns the excerpt. **This works only when `command` is the full path to the `okf`
1.1 executable.** With the bare `okf` shown above, a run started with `uv run` or from an
activated environment finds this framework's own `okf` dependency (0.8.5) first. That version
has no `mcp` command, and the server fails at startup (`invalid choice: 'mcp'`).
`okf_ask` returns the excerpt. The same test shows that a bare `okf` in `command` fails under
`uv run`, which is why the example asks for the full path.
- It is **not measured** whether the agents follow the way of working the server describes.
- `--prepass-payload` does not accept the v1.1 format (`okf-consumption/2`). That is a known gap.